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;
302 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
313 Architectures: all except ARM, arm64
315 Parameters: struct kvm_regs (in)
316 Returns: 0 on success, -1 on error
318 Writes the general purpose registers into the vcpu.
320 See KVM_GET_REGS for the data structure.
326 Architectures: x86, ppc
328 Parameters: struct kvm_sregs (out)
329 Returns: 0 on success, -1 on error
331 Reads special registers from the vcpu.
335 struct kvm_segment cs, ds, es, fs, gs, ss;
336 struct kvm_segment tr, ldt;
337 struct kvm_dtable gdt, idt;
338 __u64 cr0, cr2, cr3, cr4, cr8;
341 __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64];
344 /* ppc -- see arch/powerpc/include/uapi/asm/kvm.h */
346 interrupt_bitmap is a bitmap of pending external interrupts. At most
347 one bit may be set. This interrupt has been acknowledged by the APIC
348 but not yet injected into the cpu core.
354 Architectures: x86, ppc
356 Parameters: struct kvm_sregs (in)
357 Returns: 0 on success, -1 on error
359 Writes special registers into the vcpu. See KVM_GET_SREGS for the
368 Parameters: struct kvm_translation (in/out)
369 Returns: 0 on success, -1 on error
371 Translates a virtual address according to the vcpu's current address
374 struct kvm_translation {
376 __u64 linear_address;
379 __u64 physical_address;
390 Architectures: x86, ppc, mips
392 Parameters: struct kvm_interrupt (in)
393 Returns: 0 on success, -1 on error
395 Queues a hardware interrupt vector to be injected. This is only
396 useful if in-kernel local APIC or equivalent is not used.
398 /* for KVM_INTERRUPT */
399 struct kvm_interrupt {
406 Note 'irq' is an interrupt vector, not an interrupt pin or line.
410 Queues an external interrupt to be injected. This ioctl is overleaded
411 with 3 different irq values:
415 This injects an edge type external interrupt into the guest once it's ready
416 to receive interrupts. When injected, the interrupt is done.
418 b) KVM_INTERRUPT_UNSET
420 This unsets any pending interrupt.
422 Only available with KVM_CAP_PPC_UNSET_IRQ.
424 c) KVM_INTERRUPT_SET_LEVEL
426 This injects a level type external interrupt into the guest context. The
427 interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET
430 Only available with KVM_CAP_PPC_IRQ_LEVEL.
432 Note that any value for 'irq' other than the ones stated above is invalid
433 and incurs unexpected behavior.
437 Queues an external interrupt to be injected into the virtual CPU. A negative
438 interrupt number dequeues the interrupt.
449 Support for this has been removed. Use KVM_SET_GUEST_DEBUG instead.
457 Parameters: struct kvm_msrs (in/out)
458 Returns: 0 on success, -1 on error
460 Reads model-specific registers from the vcpu. Supported msr indices can
461 be obtained using KVM_GET_MSR_INDEX_LIST.
464 __u32 nmsrs; /* number of msrs in entries */
467 struct kvm_msr_entry entries[0];
470 struct kvm_msr_entry {
476 Application code should set the 'nmsrs' member (which indicates the
477 size of the entries array) and the 'index' member of each array entry.
478 kvm will fill in the 'data' member.
486 Parameters: struct kvm_msrs (in)
487 Returns: 0 on success, -1 on error
489 Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the
492 Application code should set the 'nmsrs' member (which indicates the
493 size of the entries array), and the 'index' and 'data' members of each
502 Parameters: struct kvm_cpuid (in)
503 Returns: 0 on success, -1 on error
505 Defines the vcpu responses to the cpuid instruction. Applications
506 should use the KVM_SET_CPUID2 ioctl if available.
509 struct kvm_cpuid_entry {
518 /* for KVM_SET_CPUID */
522 struct kvm_cpuid_entry entries[0];
526 4.21 KVM_SET_SIGNAL_MASK
531 Parameters: struct kvm_signal_mask (in)
532 Returns: 0 on success, -1 on error
534 Defines which signals are blocked during execution of KVM_RUN. This
535 signal mask temporarily overrides the threads signal mask. Any
536 unblocked signal received (except SIGKILL and SIGSTOP, which retain
537 their traditional behaviour) will cause KVM_RUN to return with -EINTR.
539 Note the signal will only be delivered if not blocked by the original
542 /* for KVM_SET_SIGNAL_MASK */
543 struct kvm_signal_mask {
554 Parameters: struct kvm_fpu (out)
555 Returns: 0 on success, -1 on error
557 Reads the floating point state from the vcpu.
559 /* for KVM_GET_FPU and KVM_SET_FPU */
564 __u8 ftwx; /* in fxsave format */
580 Parameters: struct kvm_fpu (in)
581 Returns: 0 on success, -1 on error
583 Writes the floating point state to the vcpu.
585 /* for KVM_GET_FPU and KVM_SET_FPU */
590 __u8 ftwx; /* in fxsave format */
601 4.24 KVM_CREATE_IRQCHIP
603 Capability: KVM_CAP_IRQCHIP, KVM_CAP_S390_IRQCHIP (s390)
604 Architectures: x86, ia64, ARM, arm64, s390
607 Returns: 0 on success, -1 on error
609 Creates an interrupt controller model in the kernel. On x86, creates a virtual
610 ioapic, a virtual PIC (two PICs, nested), and sets up future vcpus to have a
611 local APIC. IRQ routing for GSIs 0-15 is set to both PIC and IOAPIC; GSI 16-23
612 only go to the IOAPIC. On ia64, a IOSAPIC is created. On ARM/arm64, a GIC is
613 created. On s390, a dummy irq routing table is created.
615 Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled
616 before KVM_CREATE_IRQCHIP can be used.
621 Capability: KVM_CAP_IRQCHIP
622 Architectures: x86, ia64, arm, arm64
624 Parameters: struct kvm_irq_level
625 Returns: 0 on success, -1 on error
627 Sets the level of a GSI input to the interrupt controller model in the kernel.
628 On some architectures it is required that an interrupt controller model has
629 been previously created with KVM_CREATE_IRQCHIP. Note that edge-triggered
630 interrupts require the level to be set to 1 and then back to 0.
632 On real hardware, interrupt pins can be active-low or active-high. This
633 does not matter for the level field of struct kvm_irq_level: 1 always
634 means active (asserted), 0 means inactive (deasserted).
636 x86 allows the operating system to program the interrupt polarity
637 (active-low/active-high) for level-triggered interrupts, and KVM used
638 to consider the polarity. However, due to bitrot in the handling of
639 active-low interrupts, the above convention is now valid on x86 too.
640 This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED. Userspace
641 should not present interrupts to the guest as active-low unless this
642 capability is present (or unless it is not using the in-kernel irqchip,
646 ARM/arm64 can signal an interrupt either at the CPU level, or at the
647 in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to
648 use PPIs designated for specific cpus. The irq field is interpreted
651 Â bits: | 31 ... 24 | 23 ... 16 | 15 ... 0 |
652 field: | irq_type | vcpu_index | irq_id |
654 The irq_type field has the following values:
655 - irq_type[0]: out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ
656 - irq_type[1]: in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.)
657 (the vcpu_index field is ignored)
658 - irq_type[2]: in-kernel GIC: PPI, irq_id between 16 and 31 (incl.)
660 (The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs)
662 In both cases, level is used to assert/deassert the line.
664 struct kvm_irq_level {
667 __s32 status; /* not used for KVM_IRQ_LEVEL */
669 __u32 level; /* 0 or 1 */
675 Capability: KVM_CAP_IRQCHIP
676 Architectures: x86, ia64
678 Parameters: struct kvm_irqchip (in/out)
679 Returns: 0 on success, -1 on error
681 Reads the state of a kernel interrupt controller created with
682 KVM_CREATE_IRQCHIP into a buffer provided by the caller.
685 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
688 char dummy[512]; /* reserving space */
689 struct kvm_pic_state pic;
690 struct kvm_ioapic_state ioapic;
697 Capability: KVM_CAP_IRQCHIP
698 Architectures: x86, ia64
700 Parameters: struct kvm_irqchip (in)
701 Returns: 0 on success, -1 on error
703 Sets the state of a kernel interrupt controller created with
704 KVM_CREATE_IRQCHIP from a buffer provided by the caller.
707 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
710 char dummy[512]; /* reserving space */
711 struct kvm_pic_state pic;
712 struct kvm_ioapic_state ioapic;
717 4.28 KVM_XEN_HVM_CONFIG
719 Capability: KVM_CAP_XEN_HVM
722 Parameters: struct kvm_xen_hvm_config (in)
723 Returns: 0 on success, -1 on error
725 Sets the MSR that the Xen HVM guest uses to initialize its hypercall
726 page, and provides the starting address and size of the hypercall
727 blobs in userspace. When the guest writes the MSR, kvm copies one
728 page of a blob (32- or 64-bit, depending on the vcpu mode) to guest
731 struct kvm_xen_hvm_config {
744 Capability: KVM_CAP_ADJUST_CLOCK
747 Parameters: struct kvm_clock_data (out)
748 Returns: 0 on success, -1 on error
750 Gets the current timestamp of kvmclock as seen by the current guest. In
751 conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios
754 struct kvm_clock_data {
755 __u64 clock; /* kvmclock current value */
763 Capability: KVM_CAP_ADJUST_CLOCK
766 Parameters: struct kvm_clock_data (in)
767 Returns: 0 on success, -1 on error
769 Sets the current timestamp of kvmclock to the value specified in its parameter.
770 In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios
773 struct kvm_clock_data {
774 __u64 clock; /* kvmclock current value */
780 4.31 KVM_GET_VCPU_EVENTS
782 Capability: KVM_CAP_VCPU_EVENTS
783 Extended by: KVM_CAP_INTR_SHADOW
786 Parameters: struct kvm_vcpu_event (out)
787 Returns: 0 on success, -1 on error
789 Gets currently pending exceptions, interrupts, and NMIs as well as related
792 struct kvm_vcpu_events {
816 KVM_VCPUEVENT_VALID_SHADOW may be set in the flags field to signal that
817 interrupt.shadow contains a valid state. Otherwise, this field is undefined.
820 4.32 KVM_SET_VCPU_EVENTS
822 Capability: KVM_CAP_VCPU_EVENTS
823 Extended by: KVM_CAP_INTR_SHADOW
826 Parameters: struct kvm_vcpu_event (in)
827 Returns: 0 on success, -1 on error
829 Set pending exceptions, interrupts, and NMIs as well as related states of the
832 See KVM_GET_VCPU_EVENTS for the data structure.
834 Fields that may be modified asynchronously by running VCPUs can be excluded
835 from the update. These fields are nmi.pending and sipi_vector. Keep the
836 corresponding bits in the flags field cleared to suppress overwriting the
837 current in-kernel state. The bits are:
839 KVM_VCPUEVENT_VALID_NMI_PENDING - transfer nmi.pending to the kernel
840 KVM_VCPUEVENT_VALID_SIPI_VECTOR - transfer sipi_vector
842 If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in
843 the flags field to signal that interrupt.shadow contains a valid state and
844 shall be written into the VCPU.
847 4.33 KVM_GET_DEBUGREGS
849 Capability: KVM_CAP_DEBUGREGS
852 Parameters: struct kvm_debugregs (out)
853 Returns: 0 on success, -1 on error
855 Reads debug registers from the vcpu.
857 struct kvm_debugregs {
866 4.34 KVM_SET_DEBUGREGS
868 Capability: KVM_CAP_DEBUGREGS
871 Parameters: struct kvm_debugregs (in)
872 Returns: 0 on success, -1 on error
874 Writes debug registers into the vcpu.
876 See KVM_GET_DEBUGREGS for the data structure. The flags field is unused
877 yet and must be cleared on entry.
880 4.35 KVM_SET_USER_MEMORY_REGION
882 Capability: KVM_CAP_USER_MEM
885 Parameters: struct kvm_userspace_memory_region (in)
886 Returns: 0 on success, -1 on error
888 struct kvm_userspace_memory_region {
891 __u64 guest_phys_addr;
892 __u64 memory_size; /* bytes */
893 __u64 userspace_addr; /* start of the userspace allocated memory */
896 /* for kvm_memory_region::flags */
897 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0)
898 #define KVM_MEM_READONLY (1UL << 1)
900 This ioctl allows the user to create or modify a guest physical memory
901 slot. When changing an existing slot, it may be moved in the guest
902 physical memory space, or its flags may be modified. It may not be
903 resized. Slots may not overlap in guest physical address space.
905 Memory for the region is taken starting at the address denoted by the
906 field userspace_addr, which must point at user addressable memory for
907 the entire memory slot size. Any object may back this memory, including
908 anonymous memory, ordinary files, and hugetlbfs.
910 It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr
911 be identical. This allows large pages in the guest to be backed by large
914 The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and
915 KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of
916 writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to
917 use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it,
918 to make a new slot read-only. In this case, writes to this memory will be
919 posted to userspace as KVM_EXIT_MMIO exits.
921 When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of
922 the memory region are automatically reflected into the guest. For example, an
923 mmap() that affects the region will be made visible immediately. Another
924 example is madvise(MADV_DROP).
926 It is recommended to use this API instead of the KVM_SET_MEMORY_REGION ioctl.
927 The KVM_SET_MEMORY_REGION does not allow fine grained control over memory
928 allocation and is deprecated.
931 4.36 KVM_SET_TSS_ADDR
933 Capability: KVM_CAP_SET_TSS_ADDR
936 Parameters: unsigned long tss_address (in)
937 Returns: 0 on success, -1 on error
939 This ioctl defines the physical address of a three-page region in the guest
940 physical address space. The region must be within the first 4GB of the
941 guest physical address space and must not conflict with any memory slot
942 or any mmio address. The guest may malfunction if it accesses this memory
945 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
946 because of a quirk in the virtualization implementation (see the internals
947 documentation when it pops into existence).
952 Capability: KVM_CAP_ENABLE_CAP, KVM_CAP_ENABLE_CAP_VM
953 Architectures: ppc, s390
954 Type: vcpu ioctl, vm ioctl (with KVM_CAP_ENABLE_CAP_VM)
955 Parameters: struct kvm_enable_cap (in)
956 Returns: 0 on success; -1 on error
958 +Not all extensions are enabled by default. Using this ioctl the application
959 can enable an extension, making it available to the guest.
961 On systems that do not support this ioctl, it always fails. On systems that
962 do support it, it only works for extensions that are supported for enablement.
964 To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should
967 struct kvm_enable_cap {
971 The capability that is supposed to get enabled.
975 A bitfield indicating future enhancements. Has to be 0 for now.
979 Arguments for enabling a feature. If a feature needs initial values to
980 function properly, this is the place to put them.
985 The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl
986 for vm-wide capabilities.
988 4.38 KVM_GET_MP_STATE
990 Capability: KVM_CAP_MP_STATE
991 Architectures: x86, ia64, s390
993 Parameters: struct kvm_mp_state (out)
994 Returns: 0 on success; -1 on error
996 struct kvm_mp_state {
1000 Returns the vcpu's current "multiprocessing state" (though also valid on
1001 uniprocessor guests).
1003 Possible values are:
1005 - KVM_MP_STATE_RUNNABLE: the vcpu is currently running [x86, ia64]
1006 - KVM_MP_STATE_UNINITIALIZED: the vcpu is an application processor (AP)
1007 which has not yet received an INIT signal [x86,
1009 - KVM_MP_STATE_INIT_RECEIVED: the vcpu has received an INIT signal, and is
1010 now ready for a SIPI [x86, ia64]
1011 - KVM_MP_STATE_HALTED: the vcpu has executed a HLT instruction and
1012 is waiting for an interrupt [x86, ia64]
1013 - KVM_MP_STATE_SIPI_RECEIVED: the vcpu has just received a SIPI (vector
1014 accessible via KVM_GET_VCPU_EVENTS) [x86, ia64]
1015 - KVM_MP_STATE_STOPPED: the vcpu is stopped [s390]
1016 - KVM_MP_STATE_CHECK_STOP: the vcpu is in a special error state [s390]
1017 - KVM_MP_STATE_OPERATING: the vcpu is operating (running or halted)
1019 - KVM_MP_STATE_LOAD: the vcpu is in a special load/startup state
1022 On x86 and ia64, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an
1023 in-kernel irqchip, the multiprocessing state must be maintained by userspace on
1024 these architectures.
1027 4.39 KVM_SET_MP_STATE
1029 Capability: KVM_CAP_MP_STATE
1030 Architectures: x86, ia64, s390
1032 Parameters: struct kvm_mp_state (in)
1033 Returns: 0 on success; -1 on error
1035 Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for
1038 On x86 and ia64, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an
1039 in-kernel irqchip, the multiprocessing state must be maintained by userspace on
1040 these architectures.
1043 4.40 KVM_SET_IDENTITY_MAP_ADDR
1045 Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR
1048 Parameters: unsigned long identity (in)
1049 Returns: 0 on success, -1 on error
1051 This ioctl defines the physical address of a one-page region in the guest
1052 physical address space. The region must be within the first 4GB of the
1053 guest physical address space and must not conflict with any memory slot
1054 or any mmio address. The guest may malfunction if it accesses this memory
1057 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
1058 because of a quirk in the virtualization implementation (see the internals
1059 documentation when it pops into existence).
1062 4.41 KVM_SET_BOOT_CPU_ID
1064 Capability: KVM_CAP_SET_BOOT_CPU_ID
1065 Architectures: x86, ia64
1067 Parameters: unsigned long vcpu_id
1068 Returns: 0 on success, -1 on error
1070 Define which vcpu is the Bootstrap Processor (BSP). Values are the same
1071 as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default
1077 Capability: KVM_CAP_XSAVE
1080 Parameters: struct kvm_xsave (out)
1081 Returns: 0 on success, -1 on error
1087 This ioctl would copy current vcpu's xsave struct to the userspace.
1092 Capability: KVM_CAP_XSAVE
1095 Parameters: struct kvm_xsave (in)
1096 Returns: 0 on success, -1 on error
1102 This ioctl would copy userspace's xsave struct to the kernel.
1107 Capability: KVM_CAP_XCRS
1110 Parameters: struct kvm_xcrs (out)
1111 Returns: 0 on success, -1 on error
1122 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1126 This ioctl would copy current vcpu's xcrs to the userspace.
1131 Capability: KVM_CAP_XCRS
1134 Parameters: struct kvm_xcrs (in)
1135 Returns: 0 on success, -1 on error
1146 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1150 This ioctl would set vcpu's xcr to the value userspace specified.
1153 4.46 KVM_GET_SUPPORTED_CPUID
1155 Capability: KVM_CAP_EXT_CPUID
1158 Parameters: struct kvm_cpuid2 (in/out)
1159 Returns: 0 on success, -1 on error
1164 struct kvm_cpuid_entry2 entries[0];
1167 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
1168 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
1169 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
1171 struct kvm_cpuid_entry2 {
1182 This ioctl returns x86 cpuid features which are supported by both the hardware
1183 and kvm. Userspace can use the information returned by this ioctl to
1184 construct cpuid information (for KVM_SET_CPUID2) that is consistent with
1185 hardware, kernel, and userspace capabilities, and with user requirements (for
1186 example, the user may wish to constrain cpuid to emulate older hardware,
1187 or for feature consistency across a cluster).
1189 Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure
1190 with the 'nent' field indicating the number of entries in the variable-size
1191 array 'entries'. If the number of entries is too low to describe the cpu
1192 capabilities, an error (E2BIG) is returned. If the number is too high,
1193 the 'nent' field is adjusted and an error (ENOMEM) is returned. If the
1194 number is just right, the 'nent' field is adjusted to the number of valid
1195 entries in the 'entries' array, which is then filled.
1197 The entries returned are the host cpuid as returned by the cpuid instruction,
1198 with unknown or unsupported features masked out. Some features (for example,
1199 x2apic), may not be present in the host cpu, but are exposed by kvm if it can
1200 emulate them efficiently. The fields in each entry are defined as follows:
1202 function: the eax value used to obtain the entry
1203 index: the ecx value used to obtain the entry (for entries that are
1205 flags: an OR of zero or more of the following:
1206 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
1207 if the index field is valid
1208 KVM_CPUID_FLAG_STATEFUL_FUNC:
1209 if cpuid for this function returns different values for successive
1210 invocations; there will be several entries with the same function,
1211 all with this flag set
1212 KVM_CPUID_FLAG_STATE_READ_NEXT:
1213 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
1214 the first entry to be read by a cpu
1215 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
1216 this function/index combination
1218 The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned
1219 as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC
1220 support. Instead it is reported via
1222 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER)
1224 if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the
1225 feature in userspace, then you can enable the feature for KVM_SET_CPUID2.
1228 4.47 KVM_PPC_GET_PVINFO
1230 Capability: KVM_CAP_PPC_GET_PVINFO
1233 Parameters: struct kvm_ppc_pvinfo (out)
1234 Returns: 0 on success, !0 on error
1236 struct kvm_ppc_pvinfo {
1242 This ioctl fetches PV specific information that need to be passed to the guest
1243 using the device tree or other means from vm context.
1245 The hcall array defines 4 instructions that make up a hypercall.
1247 If any additional field gets added to this structure later on, a bit for that
1248 additional piece of information will be set in the flags bitmap.
1250 The flags bitmap is defined as:
1252 /* the host supports the ePAPR idle hcall
1253 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0)
1255 4.48 KVM_ASSIGN_PCI_DEVICE
1257 Capability: KVM_CAP_DEVICE_ASSIGNMENT
1258 Architectures: x86 ia64
1260 Parameters: struct kvm_assigned_pci_dev (in)
1261 Returns: 0 on success, -1 on error
1263 Assigns a host PCI device to the VM.
1265 struct kvm_assigned_pci_dev {
1266 __u32 assigned_dev_id;
1276 The PCI device is specified by the triple segnr, busnr, and devfn.
1277 Identification in succeeding service requests is done via assigned_dev_id. The
1278 following flags are specified:
1280 /* Depends on KVM_CAP_IOMMU */
1281 #define KVM_DEV_ASSIGN_ENABLE_IOMMU (1 << 0)
1282 /* The following two depend on KVM_CAP_PCI_2_3 */
1283 #define KVM_DEV_ASSIGN_PCI_2_3 (1 << 1)
1284 #define KVM_DEV_ASSIGN_MASK_INTX (1 << 2)
1286 If KVM_DEV_ASSIGN_PCI_2_3 is set, the kernel will manage legacy INTx interrupts
1287 via the PCI-2.3-compliant device-level mask, thus enable IRQ sharing with other
1288 assigned devices or host devices. KVM_DEV_ASSIGN_MASK_INTX specifies the
1289 guest's view on the INTx mask, see KVM_ASSIGN_SET_INTX_MASK for details.
1291 The KVM_DEV_ASSIGN_ENABLE_IOMMU flag is a mandatory option to ensure
1292 isolation of the device. Usages not specifying this flag are deprecated.
1294 Only PCI header type 0 devices with PCI BAR resources are supported by
1295 device assignment. The user requesting this ioctl must have read/write
1296 access to the PCI sysfs resource files associated with the device.
1299 4.49 KVM_DEASSIGN_PCI_DEVICE
1301 Capability: KVM_CAP_DEVICE_DEASSIGNMENT
1302 Architectures: x86 ia64
1304 Parameters: struct kvm_assigned_pci_dev (in)
1305 Returns: 0 on success, -1 on error
1307 Ends PCI device assignment, releasing all associated resources.
1309 See KVM_CAP_DEVICE_ASSIGNMENT for the data structure. Only assigned_dev_id is
1310 used in kvm_assigned_pci_dev to identify the device.
1313 4.50 KVM_ASSIGN_DEV_IRQ
1315 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1316 Architectures: x86 ia64
1318 Parameters: struct kvm_assigned_irq (in)
1319 Returns: 0 on success, -1 on error
1321 Assigns an IRQ to a passed-through device.
1323 struct kvm_assigned_irq {
1324 __u32 assigned_dev_id;
1325 __u32 host_irq; /* ignored (legacy field) */
1333 The following flags are defined:
1335 #define KVM_DEV_IRQ_HOST_INTX (1 << 0)
1336 #define KVM_DEV_IRQ_HOST_MSI (1 << 1)
1337 #define KVM_DEV_IRQ_HOST_MSIX (1 << 2)
1339 #define KVM_DEV_IRQ_GUEST_INTX (1 << 8)
1340 #define KVM_DEV_IRQ_GUEST_MSI (1 << 9)
1341 #define KVM_DEV_IRQ_GUEST_MSIX (1 << 10)
1343 It is not valid to specify multiple types per host or guest IRQ. However, the
1344 IRQ type of host and guest can differ or can even be null.
1347 4.51 KVM_DEASSIGN_DEV_IRQ
1349 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1350 Architectures: x86 ia64
1352 Parameters: struct kvm_assigned_irq (in)
1353 Returns: 0 on success, -1 on error
1355 Ends an IRQ assignment to a passed-through device.
1357 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1358 by assigned_dev_id, flags must correspond to the IRQ type specified on
1359 KVM_ASSIGN_DEV_IRQ. Partial deassignment of host or guest IRQ is allowed.
1362 4.52 KVM_SET_GSI_ROUTING
1364 Capability: KVM_CAP_IRQ_ROUTING
1365 Architectures: x86 ia64 s390
1367 Parameters: struct kvm_irq_routing (in)
1368 Returns: 0 on success, -1 on error
1370 Sets the GSI routing table entries, overwriting any previously set entries.
1372 struct kvm_irq_routing {
1375 struct kvm_irq_routing_entry entries[0];
1378 No flags are specified so far, the corresponding field must be set to zero.
1380 struct kvm_irq_routing_entry {
1386 struct kvm_irq_routing_irqchip irqchip;
1387 struct kvm_irq_routing_msi msi;
1388 struct kvm_irq_routing_s390_adapter adapter;
1393 /* gsi routing entry types */
1394 #define KVM_IRQ_ROUTING_IRQCHIP 1
1395 #define KVM_IRQ_ROUTING_MSI 2
1396 #define KVM_IRQ_ROUTING_S390_ADAPTER 3
1398 No flags are specified so far, the corresponding field must be set to zero.
1400 struct kvm_irq_routing_irqchip {
1405 struct kvm_irq_routing_msi {
1412 struct kvm_irq_routing_s390_adapter {
1416 __u32 summary_offset;
1421 4.53 KVM_ASSIGN_SET_MSIX_NR
1423 Capability: KVM_CAP_DEVICE_MSIX
1424 Architectures: x86 ia64
1426 Parameters: struct kvm_assigned_msix_nr (in)
1427 Returns: 0 on success, -1 on error
1429 Set the number of MSI-X interrupts for an assigned device. The number is
1430 reset again by terminating the MSI-X assignment of the device via
1431 KVM_DEASSIGN_DEV_IRQ. Calling this service more than once at any earlier
1434 struct kvm_assigned_msix_nr {
1435 __u32 assigned_dev_id;
1440 #define KVM_MAX_MSIX_PER_DEV 256
1443 4.54 KVM_ASSIGN_SET_MSIX_ENTRY
1445 Capability: KVM_CAP_DEVICE_MSIX
1446 Architectures: x86 ia64
1448 Parameters: struct kvm_assigned_msix_entry (in)
1449 Returns: 0 on success, -1 on error
1451 Specifies the routing of an MSI-X assigned device interrupt to a GSI. Setting
1452 the GSI vector to zero means disabling the interrupt.
1454 struct kvm_assigned_msix_entry {
1455 __u32 assigned_dev_id;
1457 __u16 entry; /* The index of entry in the MSI-X table */
1462 4.55 KVM_SET_TSC_KHZ
1464 Capability: KVM_CAP_TSC_CONTROL
1467 Parameters: virtual tsc_khz
1468 Returns: 0 on success, -1 on error
1470 Specifies the tsc frequency for the virtual machine. The unit of the
1474 4.56 KVM_GET_TSC_KHZ
1476 Capability: KVM_CAP_GET_TSC_KHZ
1480 Returns: virtual tsc-khz on success, negative value on error
1482 Returns the tsc frequency of the guest. The unit of the return value is
1483 KHz. If the host has unstable tsc this ioctl returns -EIO instead as an
1489 Capability: KVM_CAP_IRQCHIP
1492 Parameters: struct kvm_lapic_state (out)
1493 Returns: 0 on success, -1 on error
1495 #define KVM_APIC_REG_SIZE 0x400
1496 struct kvm_lapic_state {
1497 char regs[KVM_APIC_REG_SIZE];
1500 Reads the Local APIC registers and copies them into the input argument. The
1501 data format and layout are the same as documented in the architecture manual.
1506 Capability: KVM_CAP_IRQCHIP
1509 Parameters: struct kvm_lapic_state (in)
1510 Returns: 0 on success, -1 on error
1512 #define KVM_APIC_REG_SIZE 0x400
1513 struct kvm_lapic_state {
1514 char regs[KVM_APIC_REG_SIZE];
1517 Copies the input argument into the Local APIC registers. The data format
1518 and layout are the same as documented in the architecture manual.
1523 Capability: KVM_CAP_IOEVENTFD
1526 Parameters: struct kvm_ioeventfd (in)
1527 Returns: 0 on success, !0 on error
1529 This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address
1530 within the guest. A guest write in the registered address will signal the
1531 provided event instead of triggering an exit.
1533 struct kvm_ioeventfd {
1535 __u64 addr; /* legal pio/mmio address */
1536 __u32 len; /* 1, 2, 4, or 8 bytes */
1542 For the special case of virtio-ccw devices on s390, the ioevent is matched
1543 to a subchannel/virtqueue tuple instead.
1545 The following flags are defined:
1547 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch)
1548 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio)
1549 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign)
1550 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \
1551 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify)
1553 If datamatch flag is set, the event will be signaled only if the written value
1554 to the registered address is equal to datamatch in struct kvm_ioeventfd.
1556 For virtio-ccw devices, addr contains the subchannel id and datamatch the
1562 Capability: KVM_CAP_SW_TLB
1565 Parameters: struct kvm_dirty_tlb (in)
1566 Returns: 0 on success, -1 on error
1568 struct kvm_dirty_tlb {
1573 This must be called whenever userspace has changed an entry in the shared
1574 TLB, prior to calling KVM_RUN on the associated vcpu.
1576 The "bitmap" field is the userspace address of an array. This array
1577 consists of a number of bits, equal to the total number of TLB entries as
1578 determined by the last successful call to KVM_CONFIG_TLB, rounded up to the
1579 nearest multiple of 64.
1581 Each bit corresponds to one TLB entry, ordered the same as in the shared TLB
1584 The array is little-endian: the bit 0 is the least significant bit of the
1585 first byte, bit 8 is the least significant bit of the second byte, etc.
1586 This avoids any complications with differing word sizes.
1588 The "num_dirty" field is a performance hint for KVM to determine whether it
1589 should skip processing the bitmap and just invalidate everything. It must
1590 be set to the number of set bits in the bitmap.
1593 4.61 KVM_ASSIGN_SET_INTX_MASK
1595 Capability: KVM_CAP_PCI_2_3
1598 Parameters: struct kvm_assigned_pci_dev (in)
1599 Returns: 0 on success, -1 on error
1601 Allows userspace to mask PCI INTx interrupts from the assigned device. The
1602 kernel will not deliver INTx interrupts to the guest between setting and
1603 clearing of KVM_ASSIGN_SET_INTX_MASK via this interface. This enables use of
1604 and emulation of PCI 2.3 INTx disable command register behavior.
1606 This may be used for both PCI 2.3 devices supporting INTx disable natively and
1607 older devices lacking this support. Userspace is responsible for emulating the
1608 read value of the INTx disable bit in the guest visible PCI command register.
1609 When modifying the INTx disable state, userspace should precede updating the
1610 physical device command register by calling this ioctl to inform the kernel of
1611 the new intended INTx mask state.
1613 Note that the kernel uses the device INTx disable bit to internally manage the
1614 device interrupt state for PCI 2.3 devices. Reads of this register may
1615 therefore not match the expected value. Writes should always use the guest
1616 intended INTx disable value rather than attempting to read-copy-update the
1617 current physical device state. Races between user and kernel updates to the
1618 INTx disable bit are handled lazily in the kernel. It's possible the device
1619 may generate unintended interrupts, but they will not be injected into the
1622 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1623 by assigned_dev_id. In the flags field, only KVM_DEV_ASSIGN_MASK_INTX is
1627 4.62 KVM_CREATE_SPAPR_TCE
1629 Capability: KVM_CAP_SPAPR_TCE
1630 Architectures: powerpc
1632 Parameters: struct kvm_create_spapr_tce (in)
1633 Returns: file descriptor for manipulating the created TCE table
1635 This creates a virtual TCE (translation control entry) table, which
1636 is an IOMMU for PAPR-style virtual I/O. It is used to translate
1637 logical addresses used in virtual I/O into guest physical addresses,
1638 and provides a scatter/gather capability for PAPR virtual I/O.
1640 /* for KVM_CAP_SPAPR_TCE */
1641 struct kvm_create_spapr_tce {
1646 The liobn field gives the logical IO bus number for which to create a
1647 TCE table. The window_size field specifies the size of the DMA window
1648 which this TCE table will translate - the table will contain one 64
1649 bit TCE entry for every 4kiB of the DMA window.
1651 When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE
1652 table has been created using this ioctl(), the kernel will handle it
1653 in real mode, updating the TCE table. H_PUT_TCE calls for other
1654 liobns will cause a vm exit and must be handled by userspace.
1656 The return value is a file descriptor which can be passed to mmap(2)
1657 to map the created TCE table into userspace. This lets userspace read
1658 the entries written by kernel-handled H_PUT_TCE calls, and also lets
1659 userspace update the TCE table directly which is useful in some
1663 4.63 KVM_ALLOCATE_RMA
1665 Capability: KVM_CAP_PPC_RMA
1666 Architectures: powerpc
1668 Parameters: struct kvm_allocate_rma (out)
1669 Returns: file descriptor for mapping the allocated RMA
1671 This allocates a Real Mode Area (RMA) from the pool allocated at boot
1672 time by the kernel. An RMA is a physically-contiguous, aligned region
1673 of memory used on older POWER processors to provide the memory which
1674 will be accessed by real-mode (MMU off) accesses in a KVM guest.
1675 POWER processors support a set of sizes for the RMA that usually
1676 includes 64MB, 128MB, 256MB and some larger powers of two.
1678 /* for KVM_ALLOCATE_RMA */
1679 struct kvm_allocate_rma {
1683 The return value is a file descriptor which can be passed to mmap(2)
1684 to map the allocated RMA into userspace. The mapped area can then be
1685 passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the
1686 RMA for a virtual machine. The size of the RMA in bytes (which is
1687 fixed at host kernel boot time) is returned in the rma_size field of
1688 the argument structure.
1690 The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl
1691 is supported; 2 if the processor requires all virtual machines to have
1692 an RMA, or 1 if the processor can use an RMA but doesn't require it,
1693 because it supports the Virtual RMA (VRMA) facility.
1698 Capability: KVM_CAP_USER_NMI
1702 Returns: 0 on success, -1 on error
1704 Queues an NMI on the thread's vcpu. Note this is well defined only
1705 when KVM_CREATE_IRQCHIP has not been called, since this is an interface
1706 between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP
1707 has been called, this interface is completely emulated within the kernel.
1709 To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the
1710 following algorithm:
1713 - read the local APIC's state (KVM_GET_LAPIC)
1714 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1)
1715 - if so, issue KVM_NMI
1718 Some guests configure the LINT1 NMI input to cause a panic, aiding in
1722 4.65 KVM_S390_UCAS_MAP
1724 Capability: KVM_CAP_S390_UCONTROL
1727 Parameters: struct kvm_s390_ucas_mapping (in)
1728 Returns: 0 in case of success
1730 The parameter is defined like this:
1731 struct kvm_s390_ucas_mapping {
1737 This ioctl maps the memory at "user_addr" with the length "length" to
1738 the vcpu's address space starting at "vcpu_addr". All parameters need to
1739 be aligned by 1 megabyte.
1742 4.66 KVM_S390_UCAS_UNMAP
1744 Capability: KVM_CAP_S390_UCONTROL
1747 Parameters: struct kvm_s390_ucas_mapping (in)
1748 Returns: 0 in case of success
1750 The parameter is defined like this:
1751 struct kvm_s390_ucas_mapping {
1757 This ioctl unmaps the memory in the vcpu's address space starting at
1758 "vcpu_addr" with the length "length". The field "user_addr" is ignored.
1759 All parameters need to be aligned by 1 megabyte.
1762 4.67 KVM_S390_VCPU_FAULT
1764 Capability: KVM_CAP_S390_UCONTROL
1767 Parameters: vcpu absolute address (in)
1768 Returns: 0 in case of success
1770 This call creates a page table entry on the virtual cpu's address space
1771 (for user controlled virtual machines) or the virtual machine's address
1772 space (for regular virtual machines). This only works for minor faults,
1773 thus it's recommended to access subject memory page via the user page
1774 table upfront. This is useful to handle validity intercepts for user
1775 controlled virtual machines to fault in the virtual cpu's lowcore pages
1776 prior to calling the KVM_RUN ioctl.
1779 4.68 KVM_SET_ONE_REG
1781 Capability: KVM_CAP_ONE_REG
1784 Parameters: struct kvm_one_reg (in)
1785 Returns: 0 on success, negative value on failure
1787 struct kvm_one_reg {
1792 Using this ioctl, a single vcpu register can be set to a specific value
1793 defined by user space with the passed in struct kvm_one_reg, where id
1794 refers to the register identifier as described below and addr is a pointer
1795 to a variable with the respective size. There can be architecture agnostic
1796 and architecture specific registers. Each have their own range of operation
1797 and their own constants and width. To keep track of the implemented
1798 registers, find a list below:
1800 Arch | Register | Width (bits)
1802 PPC | KVM_REG_PPC_HIOR | 64
1803 PPC | KVM_REG_PPC_IAC1 | 64
1804 PPC | KVM_REG_PPC_IAC2 | 64
1805 PPC | KVM_REG_PPC_IAC3 | 64
1806 PPC | KVM_REG_PPC_IAC4 | 64
1807 PPC | KVM_REG_PPC_DAC1 | 64
1808 PPC | KVM_REG_PPC_DAC2 | 64
1809 PPC | KVM_REG_PPC_DABR | 64
1810 PPC | KVM_REG_PPC_DSCR | 64
1811 PPC | KVM_REG_PPC_PURR | 64
1812 PPC | KVM_REG_PPC_SPURR | 64
1813 PPC | KVM_REG_PPC_DAR | 64
1814 PPC | KVM_REG_PPC_DSISR | 32
1815 PPC | KVM_REG_PPC_AMR | 64
1816 PPC | KVM_REG_PPC_UAMOR | 64
1817 PPC | KVM_REG_PPC_MMCR0 | 64
1818 PPC | KVM_REG_PPC_MMCR1 | 64
1819 PPC | KVM_REG_PPC_MMCRA | 64
1820 PPC | KVM_REG_PPC_MMCR2 | 64
1821 PPC | KVM_REG_PPC_MMCRS | 64
1822 PPC | KVM_REG_PPC_SIAR | 64
1823 PPC | KVM_REG_PPC_SDAR | 64
1824 PPC | KVM_REG_PPC_SIER | 64
1825 PPC | KVM_REG_PPC_PMC1 | 32
1826 PPC | KVM_REG_PPC_PMC2 | 32
1827 PPC | KVM_REG_PPC_PMC3 | 32
1828 PPC | KVM_REG_PPC_PMC4 | 32
1829 PPC | KVM_REG_PPC_PMC5 | 32
1830 PPC | KVM_REG_PPC_PMC6 | 32
1831 PPC | KVM_REG_PPC_PMC7 | 32
1832 PPC | KVM_REG_PPC_PMC8 | 32
1833 PPC | KVM_REG_PPC_FPR0 | 64
1835 PPC | KVM_REG_PPC_FPR31 | 64
1836 PPC | KVM_REG_PPC_VR0 | 128
1838 PPC | KVM_REG_PPC_VR31 | 128
1839 PPC | KVM_REG_PPC_VSR0 | 128
1841 PPC | KVM_REG_PPC_VSR31 | 128
1842 PPC | KVM_REG_PPC_FPSCR | 64
1843 PPC | KVM_REG_PPC_VSCR | 32
1844 PPC | KVM_REG_PPC_VPA_ADDR | 64
1845 PPC | KVM_REG_PPC_VPA_SLB | 128
1846 PPC | KVM_REG_PPC_VPA_DTL | 128
1847 PPC | KVM_REG_PPC_EPCR | 32
1848 PPC | KVM_REG_PPC_EPR | 32
1849 PPC | KVM_REG_PPC_TCR | 32
1850 PPC | KVM_REG_PPC_TSR | 32
1851 PPC | KVM_REG_PPC_OR_TSR | 32
1852 PPC | KVM_REG_PPC_CLEAR_TSR | 32
1853 PPC | KVM_REG_PPC_MAS0 | 32
1854 PPC | KVM_REG_PPC_MAS1 | 32
1855 PPC | KVM_REG_PPC_MAS2 | 64
1856 PPC | KVM_REG_PPC_MAS7_3 | 64
1857 PPC | KVM_REG_PPC_MAS4 | 32
1858 PPC | KVM_REG_PPC_MAS6 | 32
1859 PPC | KVM_REG_PPC_MMUCFG | 32
1860 PPC | KVM_REG_PPC_TLB0CFG | 32
1861 PPC | KVM_REG_PPC_TLB1CFG | 32
1862 PPC | KVM_REG_PPC_TLB2CFG | 32
1863 PPC | KVM_REG_PPC_TLB3CFG | 32
1864 PPC | KVM_REG_PPC_TLB0PS | 32
1865 PPC | KVM_REG_PPC_TLB1PS | 32
1866 PPC | KVM_REG_PPC_TLB2PS | 32
1867 PPC | KVM_REG_PPC_TLB3PS | 32
1868 PPC | KVM_REG_PPC_EPTCFG | 32
1869 PPC | KVM_REG_PPC_ICP_STATE | 64
1870 PPC | KVM_REG_PPC_TB_OFFSET | 64
1871 PPC | KVM_REG_PPC_SPMC1 | 32
1872 PPC | KVM_REG_PPC_SPMC2 | 32
1873 PPC | KVM_REG_PPC_IAMR | 64
1874 PPC | KVM_REG_PPC_TFHAR | 64
1875 PPC | KVM_REG_PPC_TFIAR | 64
1876 PPC | KVM_REG_PPC_TEXASR | 64
1877 PPC | KVM_REG_PPC_FSCR | 64
1878 PPC | KVM_REG_PPC_PSPB | 32
1879 PPC | KVM_REG_PPC_EBBHR | 64
1880 PPC | KVM_REG_PPC_EBBRR | 64
1881 PPC | KVM_REG_PPC_BESCR | 64
1882 PPC | KVM_REG_PPC_TAR | 64
1883 PPC | KVM_REG_PPC_DPDES | 64
1884 PPC | KVM_REG_PPC_DAWR | 64
1885 PPC | KVM_REG_PPC_DAWRX | 64
1886 PPC | KVM_REG_PPC_CIABR | 64
1887 PPC | KVM_REG_PPC_IC | 64
1888 PPC | KVM_REG_PPC_VTB | 64
1889 PPC | KVM_REG_PPC_CSIGR | 64
1890 PPC | KVM_REG_PPC_TACR | 64
1891 PPC | KVM_REG_PPC_TCSCR | 64
1892 PPC | KVM_REG_PPC_PID | 64
1893 PPC | KVM_REG_PPC_ACOP | 64
1894 PPC | KVM_REG_PPC_VRSAVE | 32
1895 PPC | KVM_REG_PPC_LPCR | 64
1896 PPC | KVM_REG_PPC_PPR | 64
1897 PPC | KVM_REG_PPC_ARCH_COMPAT | 32
1898 PPC | KVM_REG_PPC_DABRX | 32
1899 PPC | KVM_REG_PPC_WORT | 64
1900 PPC | KVM_REG_PPC_TM_GPR0 | 64
1902 PPC | KVM_REG_PPC_TM_GPR31 | 64
1903 PPC | KVM_REG_PPC_TM_VSR0 | 128
1905 PPC | KVM_REG_PPC_TM_VSR63 | 128
1906 PPC | KVM_REG_PPC_TM_CR | 64
1907 PPC | KVM_REG_PPC_TM_LR | 64
1908 PPC | KVM_REG_PPC_TM_CTR | 64
1909 PPC | KVM_REG_PPC_TM_FPSCR | 64
1910 PPC | KVM_REG_PPC_TM_AMR | 64
1911 PPC | KVM_REG_PPC_TM_PPR | 64
1912 PPC | KVM_REG_PPC_TM_VRSAVE | 64
1913 PPC | KVM_REG_PPC_TM_VSCR | 32
1914 PPC | KVM_REG_PPC_TM_DSCR | 64
1915 PPC | KVM_REG_PPC_TM_TAR | 64
1917 MIPS | KVM_REG_MIPS_R0 | 64
1919 MIPS | KVM_REG_MIPS_R31 | 64
1920 MIPS | KVM_REG_MIPS_HI | 64
1921 MIPS | KVM_REG_MIPS_LO | 64
1922 MIPS | KVM_REG_MIPS_PC | 64
1923 MIPS | KVM_REG_MIPS_CP0_INDEX | 32
1924 MIPS | KVM_REG_MIPS_CP0_CONTEXT | 64
1925 MIPS | KVM_REG_MIPS_CP0_USERLOCAL | 64
1926 MIPS | KVM_REG_MIPS_CP0_PAGEMASK | 32
1927 MIPS | KVM_REG_MIPS_CP0_WIRED | 32
1928 MIPS | KVM_REG_MIPS_CP0_HWRENA | 32
1929 MIPS | KVM_REG_MIPS_CP0_BADVADDR | 64
1930 MIPS | KVM_REG_MIPS_CP0_COUNT | 32
1931 MIPS | KVM_REG_MIPS_CP0_ENTRYHI | 64
1932 MIPS | KVM_REG_MIPS_CP0_COMPARE | 32
1933 MIPS | KVM_REG_MIPS_CP0_STATUS | 32
1934 MIPS | KVM_REG_MIPS_CP0_CAUSE | 32
1935 MIPS | KVM_REG_MIPS_CP0_EPC | 64
1936 MIPS | KVM_REG_MIPS_CP0_CONFIG | 32
1937 MIPS | KVM_REG_MIPS_CP0_CONFIG1 | 32
1938 MIPS | KVM_REG_MIPS_CP0_CONFIG2 | 32
1939 MIPS | KVM_REG_MIPS_CP0_CONFIG3 | 32
1940 MIPS | KVM_REG_MIPS_CP0_CONFIG7 | 32
1941 MIPS | KVM_REG_MIPS_CP0_ERROREPC | 64
1942 MIPS | KVM_REG_MIPS_COUNT_CTL | 64
1943 MIPS | KVM_REG_MIPS_COUNT_RESUME | 64
1944 MIPS | KVM_REG_MIPS_COUNT_HZ | 64
1946 ARM registers are mapped using the lower 32 bits. The upper 16 of that
1947 is the register group type, or coprocessor number:
1949 ARM core registers have the following id bit patterns:
1950 0x4020 0000 0010 <index into the kvm_regs struct:16>
1952 ARM 32-bit CP15 registers have the following id bit patterns:
1953 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3>
1955 ARM 64-bit CP15 registers have the following id bit patterns:
1956 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3>
1958 ARM CCSIDR registers are demultiplexed by CSSELR value:
1959 0x4020 0000 0011 00 <csselr:8>
1961 ARM 32-bit VFP control registers have the following id bit patterns:
1962 0x4020 0000 0012 1 <regno:12>
1964 ARM 64-bit FP registers have the following id bit patterns:
1965 0x4030 0000 0012 0 <regno:12>
1968 arm64 registers are mapped using the lower 32 bits. The upper 16 of
1969 that is the register group type, or coprocessor number:
1971 arm64 core/FP-SIMD registers have the following id bit patterns. Note
1972 that the size of the access is variable, as the kvm_regs structure
1973 contains elements ranging from 32 to 128 bits. The index is a 32bit
1974 value in the kvm_regs structure seen as a 32bit array.
1975 0x60x0 0000 0010 <index into the kvm_regs struct:16>
1977 arm64 CCSIDR registers are demultiplexed by CSSELR value:
1978 0x6020 0000 0011 00 <csselr:8>
1980 arm64 system registers have the following id bit patterns:
1981 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3>
1984 MIPS registers are mapped using the lower 32 bits. The upper 16 of that is
1985 the register group type:
1987 MIPS core registers (see above) have the following id bit patterns:
1988 0x7030 0000 0000 <reg:16>
1990 MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit
1991 patterns depending on whether they're 32-bit or 64-bit registers:
1992 0x7020 0000 0001 00 <reg:5> <sel:3> (32-bit)
1993 0x7030 0000 0001 00 <reg:5> <sel:3> (64-bit)
1995 MIPS KVM control registers (see above) have the following id bit patterns:
1996 0x7030 0000 0002 <reg:16>
1999 4.69 KVM_GET_ONE_REG
2001 Capability: KVM_CAP_ONE_REG
2004 Parameters: struct kvm_one_reg (in and out)
2005 Returns: 0 on success, negative value on failure
2007 This ioctl allows to receive the value of a single register implemented
2008 in a vcpu. The register to read is indicated by the "id" field of the
2009 kvm_one_reg struct passed in. On success, the register value can be found
2010 at the memory location pointed to by "addr".
2012 The list of registers accessible using this interface is identical to the
2016 4.70 KVM_KVMCLOCK_CTRL
2018 Capability: KVM_CAP_KVMCLOCK_CTRL
2019 Architectures: Any that implement pvclocks (currently x86 only)
2022 Returns: 0 on success, -1 on error
2024 This signals to the host kernel that the specified guest is being paused by
2025 userspace. The host will set a flag in the pvclock structure that is checked
2026 from the soft lockup watchdog. The flag is part of the pvclock structure that
2027 is shared between guest and host, specifically the second bit of the flags
2028 field of the pvclock_vcpu_time_info structure. It will be set exclusively by
2029 the host and read/cleared exclusively by the guest. The guest operation of
2030 checking and clearing the flag must an atomic operation so
2031 load-link/store-conditional, or equivalent must be used. There are two cases
2032 where the guest will clear the flag: when the soft lockup watchdog timer resets
2033 itself or when a soft lockup is detected. This ioctl can be called any time
2034 after pausing the vcpu, but before it is resumed.
2039 Capability: KVM_CAP_SIGNAL_MSI
2042 Parameters: struct kvm_msi (in)
2043 Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error
2045 Directly inject a MSI message. Only valid with in-kernel irqchip that handles
2056 No flags are defined so far. The corresponding field must be 0.
2059 4.71 KVM_CREATE_PIT2
2061 Capability: KVM_CAP_PIT2
2064 Parameters: struct kvm_pit_config (in)
2065 Returns: 0 on success, -1 on error
2067 Creates an in-kernel device model for the i8254 PIT. This call is only valid
2068 after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following
2069 parameters have to be passed:
2071 struct kvm_pit_config {
2078 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */
2080 PIT timer interrupts may use a per-VM kernel thread for injection. If it
2081 exists, this thread will have a name of the following pattern:
2083 kvm-pit/<owner-process-pid>
2085 When running a guest with elevated priorities, the scheduling parameters of
2086 this thread may have to be adjusted accordingly.
2088 This IOCTL replaces the obsolete KVM_CREATE_PIT.
2093 Capability: KVM_CAP_PIT_STATE2
2096 Parameters: struct kvm_pit_state2 (out)
2097 Returns: 0 on success, -1 on error
2099 Retrieves the state of the in-kernel PIT model. Only valid after
2100 KVM_CREATE_PIT2. The state is returned in the following structure:
2102 struct kvm_pit_state2 {
2103 struct kvm_pit_channel_state channels[3];
2110 /* disable PIT in HPET legacy mode */
2111 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001
2113 This IOCTL replaces the obsolete KVM_GET_PIT.
2118 Capability: KVM_CAP_PIT_STATE2
2121 Parameters: struct kvm_pit_state2 (in)
2122 Returns: 0 on success, -1 on error
2124 Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2.
2125 See KVM_GET_PIT2 for details on struct kvm_pit_state2.
2127 This IOCTL replaces the obsolete KVM_SET_PIT.
2130 4.74 KVM_PPC_GET_SMMU_INFO
2132 Capability: KVM_CAP_PPC_GET_SMMU_INFO
2133 Architectures: powerpc
2136 Returns: 0 on success, -1 on error
2138 This populates and returns a structure describing the features of
2139 the "Server" class MMU emulation supported by KVM.
2140 This can in turn be used by userspace to generate the appropriate
2141 device-tree properties for the guest operating system.
2143 The structure contains some global information, followed by an
2144 array of supported segment page sizes:
2146 struct kvm_ppc_smmu_info {
2150 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ];
2153 The supported flags are:
2155 - KVM_PPC_PAGE_SIZES_REAL:
2156 When that flag is set, guest page sizes must "fit" the backing
2157 store page sizes. When not set, any page size in the list can
2158 be used regardless of how they are backed by userspace.
2160 - KVM_PPC_1T_SEGMENTS
2161 The emulated MMU supports 1T segments in addition to the
2164 The "slb_size" field indicates how many SLB entries are supported
2166 The "sps" array contains 8 entries indicating the supported base
2167 page sizes for a segment in increasing order. Each entry is defined
2170 struct kvm_ppc_one_seg_page_size {
2171 __u32 page_shift; /* Base page shift of segment (or 0) */
2172 __u32 slb_enc; /* SLB encoding for BookS */
2173 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ];
2176 An entry with a "page_shift" of 0 is unused. Because the array is
2177 organized in increasing order, a lookup can stop when encoutering
2180 The "slb_enc" field provides the encoding to use in the SLB for the
2181 page size. The bits are in positions such as the value can directly
2182 be OR'ed into the "vsid" argument of the slbmte instruction.
2184 The "enc" array is a list which for each of those segment base page
2185 size provides the list of supported actual page sizes (which can be
2186 only larger or equal to the base page size), along with the
2187 corresponding encoding in the hash PTE. Similarly, the array is
2188 8 entries sorted by increasing sizes and an entry with a "0" shift
2189 is an empty entry and a terminator:
2191 struct kvm_ppc_one_page_size {
2192 __u32 page_shift; /* Page shift (or 0) */
2193 __u32 pte_enc; /* Encoding in the HPTE (>>12) */
2196 The "pte_enc" field provides a value that can OR'ed into the hash
2197 PTE's RPN field (ie, it needs to be shifted left by 12 to OR it
2198 into the hash PTE second double word).
2202 Capability: KVM_CAP_IRQFD
2203 Architectures: x86 s390
2205 Parameters: struct kvm_irqfd (in)
2206 Returns: 0 on success, -1 on error
2208 Allows setting an eventfd to directly trigger a guest interrupt.
2209 kvm_irqfd.fd specifies the file descriptor to use as the eventfd and
2210 kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When
2211 an event is triggered on the eventfd, an interrupt is injected into
2212 the guest using the specified gsi pin. The irqfd is removed using
2213 the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd
2216 With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify
2217 mechanism allowing emulation of level-triggered, irqfd-based
2218 interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an
2219 additional eventfd in the kvm_irqfd.resamplefd field. When operating
2220 in resample mode, posting of an interrupt through kvm_irq.fd asserts
2221 the specified gsi in the irqchip. When the irqchip is resampled, such
2222 as from an EOI, the gsi is de-asserted and the user is notified via
2223 kvm_irqfd.resamplefd. It is the user's responsibility to re-queue
2224 the interrupt if the device making use of it still requires service.
2225 Note that closing the resamplefd is not sufficient to disable the
2226 irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment
2227 and need not be specified with KVM_IRQFD_FLAG_DEASSIGN.
2229 4.76 KVM_PPC_ALLOCATE_HTAB
2231 Capability: KVM_CAP_PPC_ALLOC_HTAB
2232 Architectures: powerpc
2234 Parameters: Pointer to u32 containing hash table order (in/out)
2235 Returns: 0 on success, -1 on error
2237 This requests the host kernel to allocate an MMU hash table for a
2238 guest using the PAPR paravirtualization interface. This only does
2239 anything if the kernel is configured to use the Book 3S HV style of
2240 virtualization. Otherwise the capability doesn't exist and the ioctl
2241 returns an ENOTTY error. The rest of this description assumes Book 3S
2244 There must be no vcpus running when this ioctl is called; if there
2245 are, it will do nothing and return an EBUSY error.
2247 The parameter is a pointer to a 32-bit unsigned integer variable
2248 containing the order (log base 2) of the desired size of the hash
2249 table, which must be between 18 and 46. On successful return from the
2250 ioctl, it will have been updated with the order of the hash table that
2253 If no hash table has been allocated when any vcpu is asked to run
2254 (with the KVM_RUN ioctl), the host kernel will allocate a
2255 default-sized hash table (16 MB).
2257 If this ioctl is called when a hash table has already been allocated,
2258 the kernel will clear out the existing hash table (zero all HPTEs) and
2259 return the hash table order in the parameter. (If the guest is using
2260 the virtualized real-mode area (VRMA) facility, the kernel will
2261 re-create the VMRA HPTEs on the next KVM_RUN of any vcpu.)
2263 4.77 KVM_S390_INTERRUPT
2267 Type: vm ioctl, vcpu ioctl
2268 Parameters: struct kvm_s390_interrupt (in)
2269 Returns: 0 on success, -1 on error
2271 Allows to inject an interrupt to the guest. Interrupts can be floating
2272 (vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type.
2274 Interrupt parameters are passed via kvm_s390_interrupt:
2276 struct kvm_s390_interrupt {
2282 type can be one of the following:
2284 KVM_S390_SIGP_STOP (vcpu) - sigp restart
2285 KVM_S390_PROGRAM_INT (vcpu) - program check; code in parm
2286 KVM_S390_SIGP_SET_PREFIX (vcpu) - sigp set prefix; prefix address in parm
2287 KVM_S390_RESTART (vcpu) - restart
2288 KVM_S390_INT_CLOCK_COMP (vcpu) - clock comparator interrupt
2289 KVM_S390_INT_CPU_TIMER (vcpu) - CPU timer interrupt
2290 KVM_S390_INT_VIRTIO (vm) - virtio external interrupt; external interrupt
2291 parameters in parm and parm64
2292 KVM_S390_INT_SERVICE (vm) - sclp external interrupt; sclp parameter in parm
2293 KVM_S390_INT_EMERGENCY (vcpu) - sigp emergency; source cpu in parm
2294 KVM_S390_INT_EXTERNAL_CALL (vcpu) - sigp external call; source cpu in parm
2295 KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) - compound value to indicate an
2296 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel);
2297 I/O interruption parameters in parm (subchannel) and parm64 (intparm,
2298 interruption subclass)
2299 KVM_S390_MCHK (vm, vcpu) - machine check interrupt; cr 14 bits in parm,
2300 machine check interrupt code in parm64 (note that
2301 machine checks needing further payload are not
2302 supported by this ioctl)
2304 Note that the vcpu ioctl is asynchronous to vcpu execution.
2306 4.78 KVM_PPC_GET_HTAB_FD
2308 Capability: KVM_CAP_PPC_HTAB_FD
2309 Architectures: powerpc
2311 Parameters: Pointer to struct kvm_get_htab_fd (in)
2312 Returns: file descriptor number (>= 0) on success, -1 on error
2314 This returns a file descriptor that can be used either to read out the
2315 entries in the guest's hashed page table (HPT), or to write entries to
2316 initialize the HPT. The returned fd can only be written to if the
2317 KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and
2318 can only be read if that bit is clear. The argument struct looks like
2321 /* For KVM_PPC_GET_HTAB_FD */
2322 struct kvm_get_htab_fd {
2328 /* Values for kvm_get_htab_fd.flags */
2329 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1)
2330 #define KVM_GET_HTAB_WRITE ((__u64)0x2)
2332 The `start_index' field gives the index in the HPT of the entry at
2333 which to start reading. It is ignored when writing.
2335 Reads on the fd will initially supply information about all
2336 "interesting" HPT entries. Interesting entries are those with the
2337 bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise
2338 all entries. When the end of the HPT is reached, the read() will
2339 return. If read() is called again on the fd, it will start again from
2340 the beginning of the HPT, but will only return HPT entries that have
2341 changed since they were last read.
2343 Data read or written is structured as a header (8 bytes) followed by a
2344 series of valid HPT entries (16 bytes) each. The header indicates how
2345 many valid HPT entries there are and how many invalid entries follow
2346 the valid entries. The invalid entries are not represented explicitly
2347 in the stream. The header format is:
2349 struct kvm_get_htab_header {
2355 Writes to the fd create HPT entries starting at the index given in the
2356 header; first `n_valid' valid entries with contents from the data
2357 written, then `n_invalid' invalid entries, invalidating any previously
2358 valid entries found.
2360 4.79 KVM_CREATE_DEVICE
2362 Capability: KVM_CAP_DEVICE_CTRL
2364 Parameters: struct kvm_create_device (in/out)
2365 Returns: 0 on success, -1 on error
2367 ENODEV: The device type is unknown or unsupported
2368 EEXIST: Device already created, and this type of device may not
2369 be instantiated multiple times
2371 Other error conditions may be defined by individual device types or
2372 have their standard meanings.
2374 Creates an emulated device in the kernel. The file descriptor returned
2375 in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR.
2377 If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the
2378 device type is supported (not necessarily whether it can be created
2381 Individual devices should not define flags. Attributes should be used
2382 for specifying any behavior that is not implied by the device type
2385 struct kvm_create_device {
2386 __u32 type; /* in: KVM_DEV_TYPE_xxx */
2387 __u32 fd; /* out: device handle */
2388 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */
2391 4.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR
2393 Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device
2394 Type: device ioctl, vm ioctl
2395 Parameters: struct kvm_device_attr
2396 Returns: 0 on success, -1 on error
2398 ENXIO: The group or attribute is unknown/unsupported for this device
2399 EPERM: The attribute cannot (currently) be accessed this way
2400 (e.g. read-only attribute, or attribute that only makes
2401 sense when the device is in a different state)
2403 Other error conditions may be defined by individual device types.
2405 Gets/sets a specified piece of device configuration and/or state. The
2406 semantics are device-specific. See individual device documentation in
2407 the "devices" directory. As with ONE_REG, the size of the data
2408 transferred is defined by the particular attribute.
2410 struct kvm_device_attr {
2411 __u32 flags; /* no flags currently defined */
2412 __u32 group; /* device-defined */
2413 __u64 attr; /* group-defined */
2414 __u64 addr; /* userspace address of attr data */
2417 4.81 KVM_HAS_DEVICE_ATTR
2419 Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device
2420 Type: device ioctl, vm ioctl
2421 Parameters: struct kvm_device_attr
2422 Returns: 0 on success, -1 on error
2424 ENXIO: The group or attribute is unknown/unsupported for this device
2426 Tests whether a device supports a particular attribute. A successful
2427 return indicates the attribute is implemented. It does not necessarily
2428 indicate that the attribute can be read or written in the device's
2429 current state. "addr" is ignored.
2431 4.82 KVM_ARM_VCPU_INIT
2434 Architectures: arm, arm64
2436 Parameters: struct kvm_vcpu_init (in)
2437 Returns: 0 on success; -1 on error
2439 Â EINVAL: Â Â Â the target is unknown, or the combination of features is invalid.
2440 Â ENOENT: Â Â Â a features bit specified is unknown.
2442 This tells KVM what type of CPU to present to the guest, and what
2443 optional features it should have. Â This will cause a reset of the cpu
2444 registers to their initial values. Â If this is not called, KVM_RUN will
2445 return ENOEXEC for that vcpu.
2447 Note that because some registers reflect machine topology, all vcpus
2448 should be created before this ioctl is invoked.
2451 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state.
2452 Depends on KVM_CAP_ARM_PSCI.
2453 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode.
2454 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only).
2455 - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 for the CPU.
2456 Depends on KVM_CAP_ARM_PSCI_0_2.
2459 4.83 KVM_ARM_PREFERRED_TARGET
2462 Architectures: arm, arm64
2464 Parameters: struct struct kvm_vcpu_init (out)
2465 Returns: 0 on success; -1 on error
2467 ENODEV: no preferred target available for the host
2469 This queries KVM for preferred CPU target type which can be emulated
2470 by KVM on underlying host.
2472 The ioctl returns struct kvm_vcpu_init instance containing information
2473 about preferred CPU target type and recommended features for it. The
2474 kvm_vcpu_init->features bitmap returned will have feature bits set if
2475 the preferred target recommends setting these features, but this is
2478 The information returned by this ioctl can be used to prepare an instance
2479 of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in
2480 in VCPU matching underlying host.
2483 4.84 KVM_GET_REG_LIST
2486 Architectures: arm, arm64, mips
2488 Parameters: struct kvm_reg_list (in/out)
2489 Returns: 0 on success; -1 on error
2491 Â E2BIG: Â Â Â Â the reg index list is too big to fit in the array specified by
2492 Â Â Â Â Â Â Â Â Â Â Â Â the user (the number required will be written into n).
2494 struct kvm_reg_list {
2495 __u64 n; /* number of registers in reg[] */
2499 This ioctl returns the guest registers that are supported for the
2500 KVM_GET_ONE_REG/KVM_SET_ONE_REG calls.
2503 4.85 KVM_ARM_SET_DEVICE_ADDR (deprecated)
2505 Capability: KVM_CAP_ARM_SET_DEVICE_ADDR
2506 Architectures: arm, arm64
2508 Parameters: struct kvm_arm_device_address (in)
2509 Returns: 0 on success, -1 on error
2511 ENODEV: The device id is unknown
2512 ENXIO: Device not supported on current system
2513 EEXIST: Address already set
2514 E2BIG: Address outside guest physical address space
2515 EBUSY: Address overlaps with other device range
2517 struct kvm_arm_device_addr {
2522 Specify a device address in the guest's physical address space where guests
2523 can access emulated or directly exposed devices, which the host kernel needs
2524 to know about. The id field is an architecture specific identifier for a
2527 ARM/arm64 divides the id field into two parts, a device id and an
2528 address type id specific to the individual device.
2530 Â bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 |
2531 field: | 0x00000000 | device id | addr type id |
2533 ARM/arm64 currently only require this when using the in-kernel GIC
2534 support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2
2535 as the device id. When setting the base address for the guest's
2536 mapping of the VGIC virtual CPU and distributor interface, the ioctl
2537 must be called after calling KVM_CREATE_IRQCHIP, but before calling
2538 KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the
2539 base addresses will return -EEXIST.
2541 Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API
2542 should be used instead.
2545 4.86 KVM_PPC_RTAS_DEFINE_TOKEN
2547 Capability: KVM_CAP_PPC_RTAS
2550 Parameters: struct kvm_rtas_token_args
2551 Returns: 0 on success, -1 on error
2553 Defines a token value for a RTAS (Run Time Abstraction Services)
2554 service in order to allow it to be handled in the kernel. The
2555 argument struct gives the name of the service, which must be the name
2556 of a service that has a kernel-side implementation. If the token
2557 value is non-zero, it will be associated with that service, and
2558 subsequent RTAS calls by the guest specifying that token will be
2559 handled by the kernel. If the token value is 0, then any token
2560 associated with the service will be forgotten, and subsequent RTAS
2561 calls by the guest for that service will be passed to userspace to be
2565 5. The kvm_run structure
2566 ------------------------
2568 Application code obtains a pointer to the kvm_run structure by
2569 mmap()ing a vcpu fd. From that point, application code can control
2570 execution by changing fields in kvm_run prior to calling the KVM_RUN
2571 ioctl, and obtain information about the reason KVM_RUN returned by
2572 looking up structure members.
2576 __u8 request_interrupt_window;
2578 Request that KVM_RUN return when it becomes possible to inject external
2579 interrupts into the guest. Useful in conjunction with KVM_INTERRUPT.
2586 When KVM_RUN has returned successfully (return value 0), this informs
2587 application code why KVM_RUN has returned. Allowable values for this
2588 field are detailed below.
2590 __u8 ready_for_interrupt_injection;
2592 If request_interrupt_window has been specified, this field indicates
2593 an interrupt can be injected now with KVM_INTERRUPT.
2597 The value of the current interrupt flag. Only valid if in-kernel
2598 local APIC is not used.
2602 /* in (pre_kvm_run), out (post_kvm_run) */
2605 The value of the cr8 register. Only valid if in-kernel local APIC is
2606 not used. Both input and output.
2610 The value of the APIC BASE msr. Only valid if in-kernel local
2611 APIC is not used. Both input and output.
2614 /* KVM_EXIT_UNKNOWN */
2616 __u64 hardware_exit_reason;
2619 If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown
2620 reasons. Further architecture-specific information is available in
2621 hardware_exit_reason.
2623 /* KVM_EXIT_FAIL_ENTRY */
2625 __u64 hardware_entry_failure_reason;
2628 If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due
2629 to unknown reasons. Further architecture-specific information is
2630 available in hardware_entry_failure_reason.
2632 /* KVM_EXIT_EXCEPTION */
2642 #define KVM_EXIT_IO_IN 0
2643 #define KVM_EXIT_IO_OUT 1
2645 __u8 size; /* bytes */
2648 __u64 data_offset; /* relative to kvm_run start */
2651 If exit_reason is KVM_EXIT_IO, then the vcpu has
2652 executed a port I/O instruction which could not be satisfied by kvm.
2653 data_offset describes where the data is located (KVM_EXIT_IO_OUT) or
2654 where kvm expects application code to place the data for the next
2655 KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array.
2658 struct kvm_debug_exit_arch arch;
2671 If exit_reason is KVM_EXIT_MMIO, then the vcpu has
2672 executed a memory-mapped I/O instruction which could not be satisfied
2673 by kvm. The 'data' member contains the written data if 'is_write' is
2674 true, and should be filled by application code otherwise.
2676 The 'data' member contains, in its first 'len' bytes, the value as it would
2677 appear if the VCPU performed a load or store of the appropriate width directly
2680 NOTE: For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_DCR,
2681 KVM_EXIT_PAPR and KVM_EXIT_EPR the corresponding
2682 operations are complete (and guest state is consistent) only after userspace
2683 has re-entered the kernel with KVM_RUN. The kernel side will first finish
2684 incomplete operations and then check for pending signals. Userspace
2685 can re-enter the guest with an unmasked signal pending to complete
2688 /* KVM_EXIT_HYPERCALL */
2697 Unused. This was once used for 'hypercall to userspace'. To implement
2698 such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390).
2699 Note KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO.
2701 /* KVM_EXIT_TPR_ACCESS */
2708 To be documented (KVM_TPR_ACCESS_REPORTING).
2710 /* KVM_EXIT_S390_SIEIC */
2713 __u64 mask; /* psw upper half */
2714 __u64 addr; /* psw lower half */
2721 /* KVM_EXIT_S390_RESET */
2722 #define KVM_S390_RESET_POR 1
2723 #define KVM_S390_RESET_CLEAR 2
2724 #define KVM_S390_RESET_SUBSYSTEM 4
2725 #define KVM_S390_RESET_CPU_INIT 8
2726 #define KVM_S390_RESET_IPL 16
2727 __u64 s390_reset_flags;
2731 /* KVM_EXIT_S390_UCONTROL */
2733 __u64 trans_exc_code;
2737 s390 specific. A page fault has occurred for a user controlled virtual
2738 machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be
2739 resolved by the kernel.
2740 The program code and the translation exception code that were placed
2741 in the cpu's lowcore are presented here as defined by the z Architecture
2742 Principles of Operation Book in the Chapter for Dynamic Address Translation
2759 MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch
2760 hypercalls and exit with this exit struct that contains all the guest gprs.
2762 If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall.
2763 Userspace can now handle the hypercall and when it's done modify the gprs as
2764 necessary. Upon guest entry all guest GPRs will then be replaced by the values
2767 /* KVM_EXIT_PAPR_HCALL */
2774 This is used on 64-bit PowerPC when emulating a pSeries partition,
2775 e.g. with the 'pseries' machine type in qemu. It occurs when the
2776 guest does a hypercall using the 'sc 1' instruction. The 'nr' field
2777 contains the hypercall number (from the guest R3), and 'args' contains
2778 the arguments (from the guest R4 - R12). Userspace should put the
2779 return code in 'ret' and any extra returned values in args[].
2780 The possible hypercalls are defined in the Power Architecture Platform
2781 Requirements (PAPR) document available from www.power.org (free
2782 developer registration required to access it).
2784 /* KVM_EXIT_S390_TSCH */
2786 __u16 subchannel_id;
2787 __u16 subchannel_nr;
2794 s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled
2795 and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O
2796 interrupt for the target subchannel has been dequeued and subchannel_id,
2797 subchannel_nr, io_int_parm and io_int_word contain the parameters for that
2798 interrupt. ipb is needed for instruction parameter decoding.
2805 On FSL BookE PowerPC chips, the interrupt controller has a fast patch
2806 interrupt acknowledge path to the core. When the core successfully
2807 delivers an interrupt, it automatically populates the EPR register with
2808 the interrupt vector number and acknowledges the interrupt inside
2809 the interrupt controller.
2811 In case the interrupt controller lives in user space, we need to do
2812 the interrupt acknowledge cycle through it to fetch the next to be
2813 delivered interrupt vector using this exit.
2815 It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an
2816 external interrupt has just been delivered into the guest. User space
2817 should put the acknowledged interrupt vector into the 'epr' field.
2819 /* KVM_EXIT_SYSTEM_EVENT */
2821 #define KVM_SYSTEM_EVENT_SHUTDOWN 1
2822 #define KVM_SYSTEM_EVENT_RESET 2
2827 If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered
2828 a system-level event using some architecture specific mechanism (hypercall
2829 or some special instruction). In case of ARM/ARM64, this is triggered using
2830 HVC instruction based PSCI call from the vcpu. The 'type' field describes
2831 the system-level event type. The 'flags' field describes architecture
2832 specific flags for the system-level event.
2834 /* Fix the size of the union. */
2839 * shared registers between kvm and userspace.
2840 * kvm_valid_regs specifies the register classes set by the host
2841 * kvm_dirty_regs specified the register classes dirtied by userspace
2842 * struct kvm_sync_regs is architecture specific, as well as the
2843 * bits for kvm_valid_regs and kvm_dirty_regs
2845 __u64 kvm_valid_regs;
2846 __u64 kvm_dirty_regs;
2848 struct kvm_sync_regs regs;
2852 If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access
2853 certain guest registers without having to call SET/GET_*REGS. Thus we can
2854 avoid some system call overhead if userspace has to handle the exit.
2855 Userspace can query the validity of the structure by checking
2856 kvm_valid_regs for specific bits. These bits are architecture specific
2857 and usually define the validity of a groups of registers. (e.g. one bit
2858 for general purpose registers)
2863 4.81 KVM_GET_EMULATED_CPUID
2865 Capability: KVM_CAP_EXT_EMUL_CPUID
2868 Parameters: struct kvm_cpuid2 (in/out)
2869 Returns: 0 on success, -1 on error
2874 struct kvm_cpuid_entry2 entries[0];
2877 The member 'flags' is used for passing flags from userspace.
2879 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
2880 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
2881 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
2883 struct kvm_cpuid_entry2 {
2894 This ioctl returns x86 cpuid features which are emulated by
2895 kvm.Userspace can use the information returned by this ioctl to query
2896 which features are emulated by kvm instead of being present natively.
2898 Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2
2899 structure with the 'nent' field indicating the number of entries in
2900 the variable-size array 'entries'. If the number of entries is too low
2901 to describe the cpu capabilities, an error (E2BIG) is returned. If the
2902 number is too high, the 'nent' field is adjusted and an error (ENOMEM)
2903 is returned. If the number is just right, the 'nent' field is adjusted
2904 to the number of valid entries in the 'entries' array, which is then
2907 The entries returned are the set CPUID bits of the respective features
2908 which kvm emulates, as returned by the CPUID instruction, with unknown
2909 or unsupported feature bits cleared.
2911 Features like x2apic, for example, may not be present in the host cpu
2912 but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be
2913 emulated efficiently and thus not included here.
2915 The fields in each entry are defined as follows:
2917 function: the eax value used to obtain the entry
2918 index: the ecx value used to obtain the entry (for entries that are
2920 flags: an OR of zero or more of the following:
2921 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
2922 if the index field is valid
2923 KVM_CPUID_FLAG_STATEFUL_FUNC:
2924 if cpuid for this function returns different values for successive
2925 invocations; there will be several entries with the same function,
2926 all with this flag set
2927 KVM_CPUID_FLAG_STATE_READ_NEXT:
2928 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
2929 the first entry to be read by a cpu
2930 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
2931 this function/index combination
2934 6. Capabilities that can be enabled
2935 -----------------------------------
2937 There are certain capabilities that change the behavior of the virtual CPU or
2938 the virtual machine when enabled. To enable them, please see section 4.37.
2939 Below you can find a list of capabilities and what their effect on the vCPU or
2940 the virtual machine is when enabling them.
2942 The following information is provided along with the description:
2944 Architectures: which instruction set architectures provide this ioctl.
2945 x86 includes both i386 and x86_64.
2947 Target: whether this is a per-vcpu or per-vm capability.
2949 Parameters: what parameters are accepted by the capability.
2951 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
2952 are not detailed, but errors with specific meanings are.
2960 Returns: 0 on success; -1 on error
2962 This capability enables interception of OSI hypercalls that otherwise would
2963 be treated as normal system calls to be injected into the guest. OSI hypercalls
2964 were invented by Mac-on-Linux to have a standardized communication mechanism
2965 between the guest and the host.
2967 When this capability is enabled, KVM_EXIT_OSI can occur.
2970 6.2 KVM_CAP_PPC_PAPR
2975 Returns: 0 on success; -1 on error
2977 This capability enables interception of PAPR hypercalls. PAPR hypercalls are
2978 done using the hypercall instruction "sc 1".
2980 It also sets the guest privilege level to "supervisor" mode. Usually the guest
2981 runs in "hypervisor" privilege mode with a few missing features.
2983 In addition to the above, it changes the semantics of SDR1. In this mode, the
2984 HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the
2985 HTAB invisible to the guest.
2987 When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur.
2994 Parameters: args[0] is the address of a struct kvm_config_tlb
2995 Returns: 0 on success; -1 on error
2997 struct kvm_config_tlb {
3004 Configures the virtual CPU's TLB array, establishing a shared memory area
3005 between userspace and KVM. The "params" and "array" fields are userspace
3006 addresses of mmu-type-specific data structures. The "array_len" field is an
3007 safety mechanism, and should be set to the size in bytes of the memory that
3008 userspace has reserved for the array. It must be at least the size dictated
3009 by "mmu_type" and "params".
3011 While KVM_RUN is active, the shared region is under control of KVM. Its
3012 contents are undefined, and any modification by userspace results in
3013 boundedly undefined behavior.
3015 On return from KVM_RUN, the shared region will reflect the current state of
3016 the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB
3017 to tell KVM which entries have been changed, prior to calling KVM_RUN again
3020 For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV:
3021 - The "params" field is of type "struct kvm_book3e_206_tlb_params".
3022 - The "array" field points to an array of type "struct
3023 kvm_book3e_206_tlb_entry".
3024 - The array consists of all entries in the first TLB, followed by all
3025 entries in the second TLB.
3026 - Within a TLB, entries are ordered first by increasing set number. Within a
3027 set, entries are ordered by way (increasing ESEL).
3028 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1)
3029 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value.
3030 - The tsize field of mas1 shall be set to 4K on TLB0, even though the
3031 hardware ignores this value for TLB0.
3033 6.4 KVM_CAP_S390_CSS_SUPPORT
3038 Returns: 0 on success; -1 on error
3040 This capability enables support for handling of channel I/O instructions.
3042 TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are
3043 handled in-kernel, while the other I/O instructions are passed to userspace.
3045 When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST
3046 SUBCHANNEL intercepts.
3048 Note that even though this capability is enabled per-vcpu, the complete
3049 virtual machine is affected.
3055 Parameters: args[0] defines whether the proxy facility is active
3056 Returns: 0 on success; -1 on error
3058 This capability enables or disables the delivery of interrupts through the
3059 external proxy facility.
3061 When enabled (args[0] != 0), every time the guest gets an external interrupt
3062 delivered, it automatically exits into user space with a KVM_EXIT_EPR exit
3063 to receive the topmost interrupt vector.
3065 When disabled (args[0] == 0), behavior is as if this facility is unsupported.
3067 When this capability is enabled, KVM_EXIT_EPR can occur.
3069 6.6 KVM_CAP_IRQ_MPIC
3072 Parameters: args[0] is the MPIC device fd
3073 args[1] is the MPIC CPU number for this vcpu
3075 This capability connects the vcpu to an in-kernel MPIC device.
3077 6.7 KVM_CAP_IRQ_XICS
3081 Parameters: args[0] is the XICS device fd
3082 args[1] is the XICS CPU number (server ID) for this vcpu
3084 This capability connects the vcpu to an in-kernel XICS device.
3086 6.8 KVM_CAP_S390_IRQCHIP
3092 This capability enables the in-kernel irqchip for s390. Please refer to
3093 "4.24 KVM_CREATE_IRQCHIP" for details.