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), a KVM_CAP_xyz constant, which
72 means availability needs to be checked with KVM_CHECK_EXTENSION
73 (see section 4.4), or 'none' which means that while not all kernels
74 support this ioctl, there's no capability bit to check its
75 availability: for kernels that don't support the ioctl,
76 the ioctl returns -ENOTTY.
78 Architectures: which instruction set architectures provide this ioctl.
79 x86 includes both i386 and x86_64.
81 Type: system, vm, or vcpu.
83 Parameters: what parameters are accepted by the ioctl.
85 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
86 are not detailed, but errors with specific meanings are.
89 4.1 KVM_GET_API_VERSION
95 Returns: the constant KVM_API_VERSION (=12)
97 This identifies the API version as the stable kvm API. It is not
98 expected that this number will change. However, Linux 2.6.20 and
99 2.6.21 report earlier versions; these are not documented and not
100 supported. Applications should refuse to run if KVM_GET_API_VERSION
101 returns a value other than 12. If this check passes, all ioctls
102 described as 'basic' will be available.
110 Parameters: machine type identifier (KVM_VM_*)
111 Returns: a VM fd that can be used to control the new virtual machine.
113 The new VM has no virtual cpus and no memory. An mmap() of a VM fd
114 will access the virtual machine's physical address space; offset zero
115 corresponds to guest physical address zero. Use of mmap() on a VM fd
116 is discouraged if userspace memory allocation (KVM_CAP_USER_MEMORY) is
118 You most certainly want to use 0 as machine type.
120 In order to create user controlled virtual machines on S390, check
121 KVM_CAP_S390_UCONTROL and use the flag KVM_VM_S390_UCONTROL as
122 privileged user (CAP_SYS_ADMIN).
125 4.3 KVM_GET_MSR_INDEX_LIST
130 Parameters: struct kvm_msr_list (in/out)
131 Returns: 0 on success; -1 on error
133 E2BIG: the msr index list is to be to fit in the array specified by
136 struct kvm_msr_list {
137 __u32 nmsrs; /* number of msrs in entries */
141 This ioctl returns the guest msrs that are supported. The list varies
142 by kvm version and host processor, but does not change otherwise. The
143 user fills in the size of the indices array in nmsrs, and in return
144 kvm adjusts nmsrs to reflect the actual number of msrs and fills in
145 the indices array with their numbers.
147 Note: if kvm indicates supports MCE (KVM_CAP_MCE), then the MCE bank MSRs are
148 not returned in the MSR list, as different vcpus can have a different number
149 of banks, as set via the KVM_X86_SETUP_MCE ioctl.
152 4.4 KVM_CHECK_EXTENSION
154 Capability: basic, KVM_CAP_CHECK_EXTENSION_VM for vm ioctl
156 Type: system ioctl, vm ioctl
157 Parameters: extension identifier (KVM_CAP_*)
158 Returns: 0 if unsupported; 1 (or some other positive integer) if supported
160 The API allows the application to query about extensions to the core
161 kvm API. Userspace passes an extension identifier (an integer) and
162 receives an integer that describes the extension availability.
163 Generally 0 means no and 1 means yes, but some extensions may report
164 additional information in the integer return value.
166 Based on their initialization different VMs may have different capabilities.
167 It is thus encouraged to use the vm ioctl to query for capabilities (available
168 with KVM_CAP_CHECK_EXTENSION_VM on the vm fd)
170 4.5 KVM_GET_VCPU_MMAP_SIZE
176 Returns: size of vcpu mmap area, in bytes
178 The KVM_RUN ioctl (cf.) communicates with userspace via a shared
179 memory region. This ioctl returns the size of that region. See the
180 KVM_RUN documentation for details.
183 4.6 KVM_SET_MEMORY_REGION
188 Parameters: struct kvm_memory_region (in)
189 Returns: 0 on success, -1 on error
191 This ioctl is obsolete and has been removed.
199 Parameters: vcpu id (apic id on x86)
200 Returns: vcpu fd on success, -1 on error
202 This API adds a vcpu to a virtual machine. The vcpu id is a small integer
203 in the range [0, max_vcpus).
205 The recommended max_vcpus value can be retrieved using the KVM_CAP_NR_VCPUS of
206 the KVM_CHECK_EXTENSION ioctl() at run-time.
207 The maximum possible value for max_vcpus can be retrieved using the
208 KVM_CAP_MAX_VCPUS of the KVM_CHECK_EXTENSION ioctl() at run-time.
210 If the KVM_CAP_NR_VCPUS does not exist, you should assume that max_vcpus is 4
212 If the KVM_CAP_MAX_VCPUS does not exist, you should assume that max_vcpus is
213 same as the value returned from KVM_CAP_NR_VCPUS.
215 On powerpc using book3s_hv mode, the vcpus are mapped onto virtual
216 threads in one or more virtual CPU cores. (This is because the
217 hardware requires all the hardware threads in a CPU core to be in the
218 same partition.) The KVM_CAP_PPC_SMT capability indicates the number
219 of vcpus per virtual core (vcore). The vcore id is obtained by
220 dividing the vcpu id by the number of vcpus per vcore. The vcpus in a
221 given vcore will always be in the same physical core as each other
222 (though that might be a different physical core from time to time).
223 Userspace can control the threading (SMT) mode of the guest by its
224 allocation of vcpu ids. For example, if userspace wants
225 single-threaded guest vcpus, it should make all vcpu ids be a multiple
226 of the number of vcpus per vcore.
228 For virtual cpus that have been created with S390 user controlled virtual
229 machines, the resulting vcpu fd can be memory mapped at page offset
230 KVM_S390_SIE_PAGE_OFFSET in order to obtain a memory map of the virtual
231 cpu's hardware control block.
234 4.8 KVM_GET_DIRTY_LOG (vm ioctl)
239 Parameters: struct kvm_dirty_log (in/out)
240 Returns: 0 on success, -1 on error
242 /* for KVM_GET_DIRTY_LOG */
243 struct kvm_dirty_log {
247 void __user *dirty_bitmap; /* one bit per page */
252 Given a memory slot, return a bitmap containing any pages dirtied
253 since the last call to this ioctl. Bit 0 is the first page in the
254 memory slot. Ensure the entire structure is cleared to avoid padding
257 If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 specifies
258 the address space for which you want to return the dirty bitmap.
259 They must be less than the value that KVM_CHECK_EXTENSION returns for
260 the KVM_CAP_MULTI_ADDRESS_SPACE capability.
263 4.9 KVM_SET_MEMORY_ALIAS
268 Parameters: struct kvm_memory_alias (in)
269 Returns: 0 (success), -1 (error)
271 This ioctl is obsolete and has been removed.
280 Returns: 0 on success, -1 on error
282 EINTR: an unmasked signal is pending
284 This ioctl is used to run a guest virtual cpu. While there are no
285 explicit parameters, there is an implicit parameter block that can be
286 obtained by mmap()ing the vcpu fd at offset 0, with the size given by
287 KVM_GET_VCPU_MMAP_SIZE. The parameter block is formatted as a 'struct
288 kvm_run' (see below).
294 Architectures: all except ARM, arm64
296 Parameters: struct kvm_regs (out)
297 Returns: 0 on success, -1 on error
299 Reads the general purpose registers from the vcpu.
303 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
304 __u64 rax, rbx, rcx, rdx;
305 __u64 rsi, rdi, rsp, rbp;
306 __u64 r8, r9, r10, r11;
307 __u64 r12, r13, r14, r15;
313 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
324 Architectures: all except ARM, arm64
326 Parameters: struct kvm_regs (in)
327 Returns: 0 on success, -1 on error
329 Writes the general purpose registers into the vcpu.
331 See KVM_GET_REGS for the data structure.
337 Architectures: x86, ppc
339 Parameters: struct kvm_sregs (out)
340 Returns: 0 on success, -1 on error
342 Reads special registers from the vcpu.
346 struct kvm_segment cs, ds, es, fs, gs, ss;
347 struct kvm_segment tr, ldt;
348 struct kvm_dtable gdt, idt;
349 __u64 cr0, cr2, cr3, cr4, cr8;
352 __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64];
355 /* ppc -- see arch/powerpc/include/uapi/asm/kvm.h */
357 interrupt_bitmap is a bitmap of pending external interrupts. At most
358 one bit may be set. This interrupt has been acknowledged by the APIC
359 but not yet injected into the cpu core.
365 Architectures: x86, ppc
367 Parameters: struct kvm_sregs (in)
368 Returns: 0 on success, -1 on error
370 Writes special registers into the vcpu. See KVM_GET_SREGS for the
379 Parameters: struct kvm_translation (in/out)
380 Returns: 0 on success, -1 on error
382 Translates a virtual address according to the vcpu's current address
385 struct kvm_translation {
387 __u64 linear_address;
390 __u64 physical_address;
401 Architectures: x86, ppc, mips
403 Parameters: struct kvm_interrupt (in)
404 Returns: 0 on success, -1 on error
406 Queues a hardware interrupt vector to be injected. This is only
407 useful if in-kernel local APIC or equivalent is not used.
409 /* for KVM_INTERRUPT */
410 struct kvm_interrupt {
417 Note 'irq' is an interrupt vector, not an interrupt pin or line.
421 Queues an external interrupt to be injected. This ioctl is overleaded
422 with 3 different irq values:
426 This injects an edge type external interrupt into the guest once it's ready
427 to receive interrupts. When injected, the interrupt is done.
429 b) KVM_INTERRUPT_UNSET
431 This unsets any pending interrupt.
433 Only available with KVM_CAP_PPC_UNSET_IRQ.
435 c) KVM_INTERRUPT_SET_LEVEL
437 This injects a level type external interrupt into the guest context. The
438 interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET
441 Only available with KVM_CAP_PPC_IRQ_LEVEL.
443 Note that any value for 'irq' other than the ones stated above is invalid
444 and incurs unexpected behavior.
448 Queues an external interrupt to be injected into the virtual CPU. A negative
449 interrupt number dequeues the interrupt.
460 Support for this has been removed. Use KVM_SET_GUEST_DEBUG instead.
468 Parameters: struct kvm_msrs (in/out)
469 Returns: 0 on success, -1 on error
471 Reads model-specific registers from the vcpu. Supported msr indices can
472 be obtained using KVM_GET_MSR_INDEX_LIST.
475 __u32 nmsrs; /* number of msrs in entries */
478 struct kvm_msr_entry entries[0];
481 struct kvm_msr_entry {
487 Application code should set the 'nmsrs' member (which indicates the
488 size of the entries array) and the 'index' member of each array entry.
489 kvm will fill in the 'data' member.
497 Parameters: struct kvm_msrs (in)
498 Returns: 0 on success, -1 on error
500 Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the
503 Application code should set the 'nmsrs' member (which indicates the
504 size of the entries array), and the 'index' and 'data' members of each
513 Parameters: struct kvm_cpuid (in)
514 Returns: 0 on success, -1 on error
516 Defines the vcpu responses to the cpuid instruction. Applications
517 should use the KVM_SET_CPUID2 ioctl if available.
520 struct kvm_cpuid_entry {
529 /* for KVM_SET_CPUID */
533 struct kvm_cpuid_entry entries[0];
537 4.21 KVM_SET_SIGNAL_MASK
542 Parameters: struct kvm_signal_mask (in)
543 Returns: 0 on success, -1 on error
545 Defines which signals are blocked during execution of KVM_RUN. This
546 signal mask temporarily overrides the threads signal mask. Any
547 unblocked signal received (except SIGKILL and SIGSTOP, which retain
548 their traditional behaviour) will cause KVM_RUN to return with -EINTR.
550 Note the signal will only be delivered if not blocked by the original
553 /* for KVM_SET_SIGNAL_MASK */
554 struct kvm_signal_mask {
565 Parameters: struct kvm_fpu (out)
566 Returns: 0 on success, -1 on error
568 Reads the floating point state from the vcpu.
570 /* for KVM_GET_FPU and KVM_SET_FPU */
575 __u8 ftwx; /* in fxsave format */
591 Parameters: struct kvm_fpu (in)
592 Returns: 0 on success, -1 on error
594 Writes the floating point state to the vcpu.
596 /* for KVM_GET_FPU and KVM_SET_FPU */
601 __u8 ftwx; /* in fxsave format */
612 4.24 KVM_CREATE_IRQCHIP
614 Capability: KVM_CAP_IRQCHIP, KVM_CAP_S390_IRQCHIP (s390)
615 Architectures: x86, ARM, arm64, s390
618 Returns: 0 on success, -1 on error
620 Creates an interrupt controller model in the kernel.
621 On x86, creates a virtual ioapic, a virtual PIC (two PICs, nested), and sets up
622 future vcpus to have a local APIC. IRQ routing for GSIs 0-15 is set to both
623 PIC and IOAPIC; GSI 16-23 only go to the IOAPIC.
624 On ARM/arm64, a GICv2 is created. Any other GIC versions require the usage of
625 KVM_CREATE_DEVICE, which also supports creating a GICv2. Using
626 KVM_CREATE_DEVICE is preferred over KVM_CREATE_IRQCHIP for GICv2.
627 On s390, a dummy irq routing table is created.
629 Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled
630 before KVM_CREATE_IRQCHIP can be used.
635 Capability: KVM_CAP_IRQCHIP
636 Architectures: x86, arm, arm64
638 Parameters: struct kvm_irq_level
639 Returns: 0 on success, -1 on error
641 Sets the level of a GSI input to the interrupt controller model in the kernel.
642 On some architectures it is required that an interrupt controller model has
643 been previously created with KVM_CREATE_IRQCHIP. Note that edge-triggered
644 interrupts require the level to be set to 1 and then back to 0.
646 On real hardware, interrupt pins can be active-low or active-high. This
647 does not matter for the level field of struct kvm_irq_level: 1 always
648 means active (asserted), 0 means inactive (deasserted).
650 x86 allows the operating system to program the interrupt polarity
651 (active-low/active-high) for level-triggered interrupts, and KVM used
652 to consider the polarity. However, due to bitrot in the handling of
653 active-low interrupts, the above convention is now valid on x86 too.
654 This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED. Userspace
655 should not present interrupts to the guest as active-low unless this
656 capability is present (or unless it is not using the in-kernel irqchip,
660 ARM/arm64 can signal an interrupt either at the CPU level, or at the
661 in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to
662 use PPIs designated for specific cpus. The irq field is interpreted
665 Â bits: | 31 ... 24 | 23 ... 16 | 15 ... 0 |
666 field: | irq_type | vcpu_index | irq_id |
668 The irq_type field has the following values:
669 - irq_type[0]: out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ
670 - irq_type[1]: in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.)
671 (the vcpu_index field is ignored)
672 - irq_type[2]: in-kernel GIC: PPI, irq_id between 16 and 31 (incl.)
674 (The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs)
676 In both cases, level is used to assert/deassert the line.
678 struct kvm_irq_level {
681 __s32 status; /* not used for KVM_IRQ_LEVEL */
683 __u32 level; /* 0 or 1 */
689 Capability: KVM_CAP_IRQCHIP
692 Parameters: struct kvm_irqchip (in/out)
693 Returns: 0 on success, -1 on error
695 Reads the state of a kernel interrupt controller created with
696 KVM_CREATE_IRQCHIP into a buffer provided by the caller.
699 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
702 char dummy[512]; /* reserving space */
703 struct kvm_pic_state pic;
704 struct kvm_ioapic_state ioapic;
711 Capability: KVM_CAP_IRQCHIP
714 Parameters: struct kvm_irqchip (in)
715 Returns: 0 on success, -1 on error
717 Sets the state of a kernel interrupt controller created with
718 KVM_CREATE_IRQCHIP from a buffer provided by the caller.
721 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
724 char dummy[512]; /* reserving space */
725 struct kvm_pic_state pic;
726 struct kvm_ioapic_state ioapic;
731 4.28 KVM_XEN_HVM_CONFIG
733 Capability: KVM_CAP_XEN_HVM
736 Parameters: struct kvm_xen_hvm_config (in)
737 Returns: 0 on success, -1 on error
739 Sets the MSR that the Xen HVM guest uses to initialize its hypercall
740 page, and provides the starting address and size of the hypercall
741 blobs in userspace. When the guest writes the MSR, kvm copies one
742 page of a blob (32- or 64-bit, depending on the vcpu mode) to guest
745 struct kvm_xen_hvm_config {
758 Capability: KVM_CAP_ADJUST_CLOCK
761 Parameters: struct kvm_clock_data (out)
762 Returns: 0 on success, -1 on error
764 Gets the current timestamp of kvmclock as seen by the current guest. In
765 conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios
768 struct kvm_clock_data {
769 __u64 clock; /* kvmclock current value */
777 Capability: KVM_CAP_ADJUST_CLOCK
780 Parameters: struct kvm_clock_data (in)
781 Returns: 0 on success, -1 on error
783 Sets the current timestamp of kvmclock to the value specified in its parameter.
784 In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios
787 struct kvm_clock_data {
788 __u64 clock; /* kvmclock current value */
794 4.31 KVM_GET_VCPU_EVENTS
796 Capability: KVM_CAP_VCPU_EVENTS
797 Extended by: KVM_CAP_INTR_SHADOW
800 Parameters: struct kvm_vcpu_event (out)
801 Returns: 0 on success, -1 on error
803 Gets currently pending exceptions, interrupts, and NMIs as well as related
806 struct kvm_vcpu_events {
836 Only two fields are defined in the flags field:
838 - KVM_VCPUEVENT_VALID_SHADOW may be set in the flags field to signal that
839 interrupt.shadow contains a valid state.
841 - KVM_VCPUEVENT_VALID_SMM may be set in the flags field to signal that
842 smi contains a valid state.
844 4.32 KVM_SET_VCPU_EVENTS
846 Capability: KVM_CAP_VCPU_EVENTS
847 Extended by: KVM_CAP_INTR_SHADOW
850 Parameters: struct kvm_vcpu_event (in)
851 Returns: 0 on success, -1 on error
853 Set pending exceptions, interrupts, and NMIs as well as related states of the
856 See KVM_GET_VCPU_EVENTS for the data structure.
858 Fields that may be modified asynchronously by running VCPUs can be excluded
859 from the update. These fields are nmi.pending, sipi_vector, smi.smm,
860 smi.pending. Keep the corresponding bits in the flags field cleared to
861 suppress overwriting the current in-kernel state. The bits are:
863 KVM_VCPUEVENT_VALID_NMI_PENDING - transfer nmi.pending to the kernel
864 KVM_VCPUEVENT_VALID_SIPI_VECTOR - transfer sipi_vector
865 KVM_VCPUEVENT_VALID_SMM - transfer the smi sub-struct.
867 If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in
868 the flags field to signal that interrupt.shadow contains a valid state and
869 shall be written into the VCPU.
871 KVM_VCPUEVENT_VALID_SMM can only be set if KVM_CAP_X86_SMM is available.
874 4.33 KVM_GET_DEBUGREGS
876 Capability: KVM_CAP_DEBUGREGS
879 Parameters: struct kvm_debugregs (out)
880 Returns: 0 on success, -1 on error
882 Reads debug registers from the vcpu.
884 struct kvm_debugregs {
893 4.34 KVM_SET_DEBUGREGS
895 Capability: KVM_CAP_DEBUGREGS
898 Parameters: struct kvm_debugregs (in)
899 Returns: 0 on success, -1 on error
901 Writes debug registers into the vcpu.
903 See KVM_GET_DEBUGREGS for the data structure. The flags field is unused
904 yet and must be cleared on entry.
907 4.35 KVM_SET_USER_MEMORY_REGION
909 Capability: KVM_CAP_USER_MEM
912 Parameters: struct kvm_userspace_memory_region (in)
913 Returns: 0 on success, -1 on error
915 struct kvm_userspace_memory_region {
918 __u64 guest_phys_addr;
919 __u64 memory_size; /* bytes */
920 __u64 userspace_addr; /* start of the userspace allocated memory */
923 /* for kvm_memory_region::flags */
924 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0)
925 #define KVM_MEM_READONLY (1UL << 1)
927 This ioctl allows the user to create or modify a guest physical memory
928 slot. When changing an existing slot, it may be moved in the guest
929 physical memory space, or its flags may be modified. It may not be
930 resized. Slots may not overlap in guest physical address space.
932 If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of "slot"
933 specifies the address space which is being modified. They must be
934 less than the value that KVM_CHECK_EXTENSION returns for the
935 KVM_CAP_MULTI_ADDRESS_SPACE capability. Slots in separate address spaces
936 are unrelated; the restriction on overlapping slots only applies within
939 Memory for the region is taken starting at the address denoted by the
940 field userspace_addr, which must point at user addressable memory for
941 the entire memory slot size. Any object may back this memory, including
942 anonymous memory, ordinary files, and hugetlbfs.
944 It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr
945 be identical. This allows large pages in the guest to be backed by large
948 The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and
949 KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of
950 writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to
951 use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it,
952 to make a new slot read-only. In this case, writes to this memory will be
953 posted to userspace as KVM_EXIT_MMIO exits.
955 When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of
956 the memory region are automatically reflected into the guest. For example, an
957 mmap() that affects the region will be made visible immediately. Another
958 example is madvise(MADV_DROP).
960 It is recommended to use this API instead of the KVM_SET_MEMORY_REGION ioctl.
961 The KVM_SET_MEMORY_REGION does not allow fine grained control over memory
962 allocation and is deprecated.
965 4.36 KVM_SET_TSS_ADDR
967 Capability: KVM_CAP_SET_TSS_ADDR
970 Parameters: unsigned long tss_address (in)
971 Returns: 0 on success, -1 on error
973 This ioctl defines the physical address of a three-page region in the guest
974 physical address space. The region must be within the first 4GB of the
975 guest physical address space and must not conflict with any memory slot
976 or any mmio address. The guest may malfunction if it accesses this memory
979 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
980 because of a quirk in the virtualization implementation (see the internals
981 documentation when it pops into existence).
986 Capability: KVM_CAP_ENABLE_CAP, KVM_CAP_ENABLE_CAP_VM
987 Architectures: x86 (only KVM_CAP_ENABLE_CAP_VM),
988 mips (only KVM_CAP_ENABLE_CAP), ppc, s390
989 Type: vcpu ioctl, vm ioctl (with KVM_CAP_ENABLE_CAP_VM)
990 Parameters: struct kvm_enable_cap (in)
991 Returns: 0 on success; -1 on error
993 +Not all extensions are enabled by default. Using this ioctl the application
994 can enable an extension, making it available to the guest.
996 On systems that do not support this ioctl, it always fails. On systems that
997 do support it, it only works for extensions that are supported for enablement.
999 To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should
1002 struct kvm_enable_cap {
1006 The capability that is supposed to get enabled.
1010 A bitfield indicating future enhancements. Has to be 0 for now.
1014 Arguments for enabling a feature. If a feature needs initial values to
1015 function properly, this is the place to put them.
1020 The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl
1021 for vm-wide capabilities.
1023 4.38 KVM_GET_MP_STATE
1025 Capability: KVM_CAP_MP_STATE
1026 Architectures: x86, s390, arm, arm64
1028 Parameters: struct kvm_mp_state (out)
1029 Returns: 0 on success; -1 on error
1031 struct kvm_mp_state {
1035 Returns the vcpu's current "multiprocessing state" (though also valid on
1036 uniprocessor guests).
1038 Possible values are:
1040 - KVM_MP_STATE_RUNNABLE: the vcpu is currently running [x86,arm/arm64]
1041 - KVM_MP_STATE_UNINITIALIZED: the vcpu is an application processor (AP)
1042 which has not yet received an INIT signal [x86]
1043 - KVM_MP_STATE_INIT_RECEIVED: the vcpu has received an INIT signal, and is
1044 now ready for a SIPI [x86]
1045 - KVM_MP_STATE_HALTED: the vcpu has executed a HLT instruction and
1046 is waiting for an interrupt [x86]
1047 - KVM_MP_STATE_SIPI_RECEIVED: the vcpu has just received a SIPI (vector
1048 accessible via KVM_GET_VCPU_EVENTS) [x86]
1049 - KVM_MP_STATE_STOPPED: the vcpu is stopped [s390,arm/arm64]
1050 - KVM_MP_STATE_CHECK_STOP: the vcpu is in a special error state [s390]
1051 - KVM_MP_STATE_OPERATING: the vcpu is operating (running or halted)
1053 - KVM_MP_STATE_LOAD: the vcpu is in a special load/startup state
1056 On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an
1057 in-kernel irqchip, the multiprocessing state must be maintained by userspace on
1058 these architectures.
1062 The only states that are valid are KVM_MP_STATE_STOPPED and
1063 KVM_MP_STATE_RUNNABLE which reflect if the vcpu is paused or not.
1065 4.39 KVM_SET_MP_STATE
1067 Capability: KVM_CAP_MP_STATE
1068 Architectures: x86, s390, arm, arm64
1070 Parameters: struct kvm_mp_state (in)
1071 Returns: 0 on success; -1 on error
1073 Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for
1076 On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an
1077 in-kernel irqchip, the multiprocessing state must be maintained by userspace on
1078 these architectures.
1082 The only states that are valid are KVM_MP_STATE_STOPPED and
1083 KVM_MP_STATE_RUNNABLE which reflect if the vcpu should be paused or not.
1085 4.40 KVM_SET_IDENTITY_MAP_ADDR
1087 Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR
1090 Parameters: unsigned long identity (in)
1091 Returns: 0 on success, -1 on error
1093 This ioctl defines the physical address of a one-page region in the guest
1094 physical address space. The region must be within the first 4GB of the
1095 guest physical address space and must not conflict with any memory slot
1096 or any mmio address. The guest may malfunction if it accesses this memory
1099 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
1100 because of a quirk in the virtualization implementation (see the internals
1101 documentation when it pops into existence).
1104 4.41 KVM_SET_BOOT_CPU_ID
1106 Capability: KVM_CAP_SET_BOOT_CPU_ID
1109 Parameters: unsigned long vcpu_id
1110 Returns: 0 on success, -1 on error
1112 Define which vcpu is the Bootstrap Processor (BSP). Values are the same
1113 as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default
1119 Capability: KVM_CAP_XSAVE
1122 Parameters: struct kvm_xsave (out)
1123 Returns: 0 on success, -1 on error
1129 This ioctl would copy current vcpu's xsave struct to the userspace.
1134 Capability: KVM_CAP_XSAVE
1137 Parameters: struct kvm_xsave (in)
1138 Returns: 0 on success, -1 on error
1144 This ioctl would copy userspace's xsave struct to the kernel.
1149 Capability: KVM_CAP_XCRS
1152 Parameters: struct kvm_xcrs (out)
1153 Returns: 0 on success, -1 on error
1164 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1168 This ioctl would copy current vcpu's xcrs to the userspace.
1173 Capability: KVM_CAP_XCRS
1176 Parameters: struct kvm_xcrs (in)
1177 Returns: 0 on success, -1 on error
1188 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1192 This ioctl would set vcpu's xcr to the value userspace specified.
1195 4.46 KVM_GET_SUPPORTED_CPUID
1197 Capability: KVM_CAP_EXT_CPUID
1200 Parameters: struct kvm_cpuid2 (in/out)
1201 Returns: 0 on success, -1 on error
1206 struct kvm_cpuid_entry2 entries[0];
1209 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
1210 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
1211 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
1213 struct kvm_cpuid_entry2 {
1224 This ioctl returns x86 cpuid features which are supported by both the hardware
1225 and kvm. Userspace can use the information returned by this ioctl to
1226 construct cpuid information (for KVM_SET_CPUID2) that is consistent with
1227 hardware, kernel, and userspace capabilities, and with user requirements (for
1228 example, the user may wish to constrain cpuid to emulate older hardware,
1229 or for feature consistency across a cluster).
1231 Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure
1232 with the 'nent' field indicating the number of entries in the variable-size
1233 array 'entries'. If the number of entries is too low to describe the cpu
1234 capabilities, an error (E2BIG) is returned. If the number is too high,
1235 the 'nent' field is adjusted and an error (ENOMEM) is returned. If the
1236 number is just right, the 'nent' field is adjusted to the number of valid
1237 entries in the 'entries' array, which is then filled.
1239 The entries returned are the host cpuid as returned by the cpuid instruction,
1240 with unknown or unsupported features masked out. Some features (for example,
1241 x2apic), may not be present in the host cpu, but are exposed by kvm if it can
1242 emulate them efficiently. The fields in each entry are defined as follows:
1244 function: the eax value used to obtain the entry
1245 index: the ecx value used to obtain the entry (for entries that are
1247 flags: an OR of zero or more of the following:
1248 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
1249 if the index field is valid
1250 KVM_CPUID_FLAG_STATEFUL_FUNC:
1251 if cpuid for this function returns different values for successive
1252 invocations; there will be several entries with the same function,
1253 all with this flag set
1254 KVM_CPUID_FLAG_STATE_READ_NEXT:
1255 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
1256 the first entry to be read by a cpu
1257 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
1258 this function/index combination
1260 The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned
1261 as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC
1262 support. Instead it is reported via
1264 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER)
1266 if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the
1267 feature in userspace, then you can enable the feature for KVM_SET_CPUID2.
1270 4.47 KVM_PPC_GET_PVINFO
1272 Capability: KVM_CAP_PPC_GET_PVINFO
1275 Parameters: struct kvm_ppc_pvinfo (out)
1276 Returns: 0 on success, !0 on error
1278 struct kvm_ppc_pvinfo {
1284 This ioctl fetches PV specific information that need to be passed to the guest
1285 using the device tree or other means from vm context.
1287 The hcall array defines 4 instructions that make up a hypercall.
1289 If any additional field gets added to this structure later on, a bit for that
1290 additional piece of information will be set in the flags bitmap.
1292 The flags bitmap is defined as:
1294 /* the host supports the ePAPR idle hcall
1295 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0)
1297 4.48 KVM_ASSIGN_PCI_DEVICE (deprecated)
1302 Parameters: struct kvm_assigned_pci_dev (in)
1303 Returns: 0 on success, -1 on error
1305 Assigns a host PCI device to the VM.
1307 struct kvm_assigned_pci_dev {
1308 __u32 assigned_dev_id;
1318 The PCI device is specified by the triple segnr, busnr, and devfn.
1319 Identification in succeeding service requests is done via assigned_dev_id. The
1320 following flags are specified:
1322 /* Depends on KVM_CAP_IOMMU */
1323 #define KVM_DEV_ASSIGN_ENABLE_IOMMU (1 << 0)
1324 /* The following two depend on KVM_CAP_PCI_2_3 */
1325 #define KVM_DEV_ASSIGN_PCI_2_3 (1 << 1)
1326 #define KVM_DEV_ASSIGN_MASK_INTX (1 << 2)
1328 If KVM_DEV_ASSIGN_PCI_2_3 is set, the kernel will manage legacy INTx interrupts
1329 via the PCI-2.3-compliant device-level mask, thus enable IRQ sharing with other
1330 assigned devices or host devices. KVM_DEV_ASSIGN_MASK_INTX specifies the
1331 guest's view on the INTx mask, see KVM_ASSIGN_SET_INTX_MASK for details.
1333 The KVM_DEV_ASSIGN_ENABLE_IOMMU flag is a mandatory option to ensure
1334 isolation of the device. Usages not specifying this flag are deprecated.
1336 Only PCI header type 0 devices with PCI BAR resources are supported by
1337 device assignment. The user requesting this ioctl must have read/write
1338 access to the PCI sysfs resource files associated with the device.
1341 ENOTTY: kernel does not support this ioctl
1343 Other error conditions may be defined by individual device types or
1344 have their standard meanings.
1347 4.49 KVM_DEASSIGN_PCI_DEVICE (deprecated)
1352 Parameters: struct kvm_assigned_pci_dev (in)
1353 Returns: 0 on success, -1 on error
1355 Ends PCI device assignment, releasing all associated resources.
1357 See KVM_ASSIGN_PCI_DEVICE for the data structure. Only assigned_dev_id is
1358 used in kvm_assigned_pci_dev to identify the device.
1361 ENOTTY: kernel does not support this ioctl
1363 Other error conditions may be defined by individual device types or
1364 have their standard meanings.
1366 4.50 KVM_ASSIGN_DEV_IRQ (deprecated)
1368 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1371 Parameters: struct kvm_assigned_irq (in)
1372 Returns: 0 on success, -1 on error
1374 Assigns an IRQ to a passed-through device.
1376 struct kvm_assigned_irq {
1377 __u32 assigned_dev_id;
1378 __u32 host_irq; /* ignored (legacy field) */
1386 The following flags are defined:
1388 #define KVM_DEV_IRQ_HOST_INTX (1 << 0)
1389 #define KVM_DEV_IRQ_HOST_MSI (1 << 1)
1390 #define KVM_DEV_IRQ_HOST_MSIX (1 << 2)
1392 #define KVM_DEV_IRQ_GUEST_INTX (1 << 8)
1393 #define KVM_DEV_IRQ_GUEST_MSI (1 << 9)
1394 #define KVM_DEV_IRQ_GUEST_MSIX (1 << 10)
1396 It is not valid to specify multiple types per host or guest IRQ. However, the
1397 IRQ type of host and guest can differ or can even be null.
1400 ENOTTY: kernel does not support this ioctl
1402 Other error conditions may be defined by individual device types or
1403 have their standard meanings.
1406 4.51 KVM_DEASSIGN_DEV_IRQ (deprecated)
1408 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1411 Parameters: struct kvm_assigned_irq (in)
1412 Returns: 0 on success, -1 on error
1414 Ends an IRQ assignment to a passed-through device.
1416 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1417 by assigned_dev_id, flags must correspond to the IRQ type specified on
1418 KVM_ASSIGN_DEV_IRQ. Partial deassignment of host or guest IRQ is allowed.
1421 4.52 KVM_SET_GSI_ROUTING
1423 Capability: KVM_CAP_IRQ_ROUTING
1424 Architectures: x86 s390
1426 Parameters: struct kvm_irq_routing (in)
1427 Returns: 0 on success, -1 on error
1429 Sets the GSI routing table entries, overwriting any previously set entries.
1431 struct kvm_irq_routing {
1434 struct kvm_irq_routing_entry entries[0];
1437 No flags are specified so far, the corresponding field must be set to zero.
1439 struct kvm_irq_routing_entry {
1445 struct kvm_irq_routing_irqchip irqchip;
1446 struct kvm_irq_routing_msi msi;
1447 struct kvm_irq_routing_s390_adapter adapter;
1452 /* gsi routing entry types */
1453 #define KVM_IRQ_ROUTING_IRQCHIP 1
1454 #define KVM_IRQ_ROUTING_MSI 2
1455 #define KVM_IRQ_ROUTING_S390_ADAPTER 3
1457 No flags are specified so far, the corresponding field must be set to zero.
1459 struct kvm_irq_routing_irqchip {
1464 struct kvm_irq_routing_msi {
1471 struct kvm_irq_routing_s390_adapter {
1475 __u32 summary_offset;
1480 4.53 KVM_ASSIGN_SET_MSIX_NR (deprecated)
1485 Parameters: struct kvm_assigned_msix_nr (in)
1486 Returns: 0 on success, -1 on error
1488 Set the number of MSI-X interrupts for an assigned device. The number is
1489 reset again by terminating the MSI-X assignment of the device via
1490 KVM_DEASSIGN_DEV_IRQ. Calling this service more than once at any earlier
1493 struct kvm_assigned_msix_nr {
1494 __u32 assigned_dev_id;
1499 #define KVM_MAX_MSIX_PER_DEV 256
1502 4.54 KVM_ASSIGN_SET_MSIX_ENTRY (deprecated)
1507 Parameters: struct kvm_assigned_msix_entry (in)
1508 Returns: 0 on success, -1 on error
1510 Specifies the routing of an MSI-X assigned device interrupt to a GSI. Setting
1511 the GSI vector to zero means disabling the interrupt.
1513 struct kvm_assigned_msix_entry {
1514 __u32 assigned_dev_id;
1516 __u16 entry; /* The index of entry in the MSI-X table */
1521 ENOTTY: kernel does not support this ioctl
1523 Other error conditions may be defined by individual device types or
1524 have their standard meanings.
1527 4.55 KVM_SET_TSC_KHZ
1529 Capability: KVM_CAP_TSC_CONTROL
1532 Parameters: virtual tsc_khz
1533 Returns: 0 on success, -1 on error
1535 Specifies the tsc frequency for the virtual machine. The unit of the
1539 4.56 KVM_GET_TSC_KHZ
1541 Capability: KVM_CAP_GET_TSC_KHZ
1545 Returns: virtual tsc-khz on success, negative value on error
1547 Returns the tsc frequency of the guest. The unit of the return value is
1548 KHz. If the host has unstable tsc this ioctl returns -EIO instead as an
1554 Capability: KVM_CAP_IRQCHIP
1557 Parameters: struct kvm_lapic_state (out)
1558 Returns: 0 on success, -1 on error
1560 #define KVM_APIC_REG_SIZE 0x400
1561 struct kvm_lapic_state {
1562 char regs[KVM_APIC_REG_SIZE];
1565 Reads the Local APIC registers and copies them into the input argument. The
1566 data format and layout are the same as documented in the architecture manual.
1571 Capability: KVM_CAP_IRQCHIP
1574 Parameters: struct kvm_lapic_state (in)
1575 Returns: 0 on success, -1 on error
1577 #define KVM_APIC_REG_SIZE 0x400
1578 struct kvm_lapic_state {
1579 char regs[KVM_APIC_REG_SIZE];
1582 Copies the input argument into the Local APIC registers. The data format
1583 and layout are the same as documented in the architecture manual.
1588 Capability: KVM_CAP_IOEVENTFD
1591 Parameters: struct kvm_ioeventfd (in)
1592 Returns: 0 on success, !0 on error
1594 This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address
1595 within the guest. A guest write in the registered address will signal the
1596 provided event instead of triggering an exit.
1598 struct kvm_ioeventfd {
1600 __u64 addr; /* legal pio/mmio address */
1601 __u32 len; /* 1, 2, 4, or 8 bytes */
1607 For the special case of virtio-ccw devices on s390, the ioevent is matched
1608 to a subchannel/virtqueue tuple instead.
1610 The following flags are defined:
1612 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch)
1613 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio)
1614 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign)
1615 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \
1616 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify)
1618 If datamatch flag is set, the event will be signaled only if the written value
1619 to the registered address is equal to datamatch in struct kvm_ioeventfd.
1621 For virtio-ccw devices, addr contains the subchannel id and datamatch the
1627 Capability: KVM_CAP_SW_TLB
1630 Parameters: struct kvm_dirty_tlb (in)
1631 Returns: 0 on success, -1 on error
1633 struct kvm_dirty_tlb {
1638 This must be called whenever userspace has changed an entry in the shared
1639 TLB, prior to calling KVM_RUN on the associated vcpu.
1641 The "bitmap" field is the userspace address of an array. This array
1642 consists of a number of bits, equal to the total number of TLB entries as
1643 determined by the last successful call to KVM_CONFIG_TLB, rounded up to the
1644 nearest multiple of 64.
1646 Each bit corresponds to one TLB entry, ordered the same as in the shared TLB
1649 The array is little-endian: the bit 0 is the least significant bit of the
1650 first byte, bit 8 is the least significant bit of the second byte, etc.
1651 This avoids any complications with differing word sizes.
1653 The "num_dirty" field is a performance hint for KVM to determine whether it
1654 should skip processing the bitmap and just invalidate everything. It must
1655 be set to the number of set bits in the bitmap.
1658 4.61 KVM_ASSIGN_SET_INTX_MASK (deprecated)
1660 Capability: KVM_CAP_PCI_2_3
1663 Parameters: struct kvm_assigned_pci_dev (in)
1664 Returns: 0 on success, -1 on error
1666 Allows userspace to mask PCI INTx interrupts from the assigned device. The
1667 kernel will not deliver INTx interrupts to the guest between setting and
1668 clearing of KVM_ASSIGN_SET_INTX_MASK via this interface. This enables use of
1669 and emulation of PCI 2.3 INTx disable command register behavior.
1671 This may be used for both PCI 2.3 devices supporting INTx disable natively and
1672 older devices lacking this support. Userspace is responsible for emulating the
1673 read value of the INTx disable bit in the guest visible PCI command register.
1674 When modifying the INTx disable state, userspace should precede updating the
1675 physical device command register by calling this ioctl to inform the kernel of
1676 the new intended INTx mask state.
1678 Note that the kernel uses the device INTx disable bit to internally manage the
1679 device interrupt state for PCI 2.3 devices. Reads of this register may
1680 therefore not match the expected value. Writes should always use the guest
1681 intended INTx disable value rather than attempting to read-copy-update the
1682 current physical device state. Races between user and kernel updates to the
1683 INTx disable bit are handled lazily in the kernel. It's possible the device
1684 may generate unintended interrupts, but they will not be injected into the
1687 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1688 by assigned_dev_id. In the flags field, only KVM_DEV_ASSIGN_MASK_INTX is
1692 4.62 KVM_CREATE_SPAPR_TCE
1694 Capability: KVM_CAP_SPAPR_TCE
1695 Architectures: powerpc
1697 Parameters: struct kvm_create_spapr_tce (in)
1698 Returns: file descriptor for manipulating the created TCE table
1700 This creates a virtual TCE (translation control entry) table, which
1701 is an IOMMU for PAPR-style virtual I/O. It is used to translate
1702 logical addresses used in virtual I/O into guest physical addresses,
1703 and provides a scatter/gather capability for PAPR virtual I/O.
1705 /* for KVM_CAP_SPAPR_TCE */
1706 struct kvm_create_spapr_tce {
1711 The liobn field gives the logical IO bus number for which to create a
1712 TCE table. The window_size field specifies the size of the DMA window
1713 which this TCE table will translate - the table will contain one 64
1714 bit TCE entry for every 4kiB of the DMA window.
1716 When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE
1717 table has been created using this ioctl(), the kernel will handle it
1718 in real mode, updating the TCE table. H_PUT_TCE calls for other
1719 liobns will cause a vm exit and must be handled by userspace.
1721 The return value is a file descriptor which can be passed to mmap(2)
1722 to map the created TCE table into userspace. This lets userspace read
1723 the entries written by kernel-handled H_PUT_TCE calls, and also lets
1724 userspace update the TCE table directly which is useful in some
1728 4.63 KVM_ALLOCATE_RMA
1730 Capability: KVM_CAP_PPC_RMA
1731 Architectures: powerpc
1733 Parameters: struct kvm_allocate_rma (out)
1734 Returns: file descriptor for mapping the allocated RMA
1736 This allocates a Real Mode Area (RMA) from the pool allocated at boot
1737 time by the kernel. An RMA is a physically-contiguous, aligned region
1738 of memory used on older POWER processors to provide the memory which
1739 will be accessed by real-mode (MMU off) accesses in a KVM guest.
1740 POWER processors support a set of sizes for the RMA that usually
1741 includes 64MB, 128MB, 256MB and some larger powers of two.
1743 /* for KVM_ALLOCATE_RMA */
1744 struct kvm_allocate_rma {
1748 The return value is a file descriptor which can be passed to mmap(2)
1749 to map the allocated RMA into userspace. The mapped area can then be
1750 passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the
1751 RMA for a virtual machine. The size of the RMA in bytes (which is
1752 fixed at host kernel boot time) is returned in the rma_size field of
1753 the argument structure.
1755 The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl
1756 is supported; 2 if the processor requires all virtual machines to have
1757 an RMA, or 1 if the processor can use an RMA but doesn't require it,
1758 because it supports the Virtual RMA (VRMA) facility.
1763 Capability: KVM_CAP_USER_NMI
1767 Returns: 0 on success, -1 on error
1769 Queues an NMI on the thread's vcpu. Note this is well defined only
1770 when KVM_CREATE_IRQCHIP has not been called, since this is an interface
1771 between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP
1772 has been called, this interface is completely emulated within the kernel.
1774 To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the
1775 following algorithm:
1778 - read the local APIC's state (KVM_GET_LAPIC)
1779 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1)
1780 - if so, issue KVM_NMI
1783 Some guests configure the LINT1 NMI input to cause a panic, aiding in
1787 4.65 KVM_S390_UCAS_MAP
1789 Capability: KVM_CAP_S390_UCONTROL
1792 Parameters: struct kvm_s390_ucas_mapping (in)
1793 Returns: 0 in case of success
1795 The parameter is defined like this:
1796 struct kvm_s390_ucas_mapping {
1802 This ioctl maps the memory at "user_addr" with the length "length" to
1803 the vcpu's address space starting at "vcpu_addr". All parameters need to
1804 be aligned by 1 megabyte.
1807 4.66 KVM_S390_UCAS_UNMAP
1809 Capability: KVM_CAP_S390_UCONTROL
1812 Parameters: struct kvm_s390_ucas_mapping (in)
1813 Returns: 0 in case of success
1815 The parameter is defined like this:
1816 struct kvm_s390_ucas_mapping {
1822 This ioctl unmaps the memory in the vcpu's address space starting at
1823 "vcpu_addr" with the length "length". The field "user_addr" is ignored.
1824 All parameters need to be aligned by 1 megabyte.
1827 4.67 KVM_S390_VCPU_FAULT
1829 Capability: KVM_CAP_S390_UCONTROL
1832 Parameters: vcpu absolute address (in)
1833 Returns: 0 in case of success
1835 This call creates a page table entry on the virtual cpu's address space
1836 (for user controlled virtual machines) or the virtual machine's address
1837 space (for regular virtual machines). This only works for minor faults,
1838 thus it's recommended to access subject memory page via the user page
1839 table upfront. This is useful to handle validity intercepts for user
1840 controlled virtual machines to fault in the virtual cpu's lowcore pages
1841 prior to calling the KVM_RUN ioctl.
1844 4.68 KVM_SET_ONE_REG
1846 Capability: KVM_CAP_ONE_REG
1849 Parameters: struct kvm_one_reg (in)
1850 Returns: 0 on success, negative value on failure
1852 struct kvm_one_reg {
1857 Using this ioctl, a single vcpu register can be set to a specific value
1858 defined by user space with the passed in struct kvm_one_reg, where id
1859 refers to the register identifier as described below and addr is a pointer
1860 to a variable with the respective size. There can be architecture agnostic
1861 and architecture specific registers. Each have their own range of operation
1862 and their own constants and width. To keep track of the implemented
1863 registers, find a list below:
1865 Arch | Register | Width (bits)
1867 PPC | KVM_REG_PPC_HIOR | 64
1868 PPC | KVM_REG_PPC_IAC1 | 64
1869 PPC | KVM_REG_PPC_IAC2 | 64
1870 PPC | KVM_REG_PPC_IAC3 | 64
1871 PPC | KVM_REG_PPC_IAC4 | 64
1872 PPC | KVM_REG_PPC_DAC1 | 64
1873 PPC | KVM_REG_PPC_DAC2 | 64
1874 PPC | KVM_REG_PPC_DABR | 64
1875 PPC | KVM_REG_PPC_DSCR | 64
1876 PPC | KVM_REG_PPC_PURR | 64
1877 PPC | KVM_REG_PPC_SPURR | 64
1878 PPC | KVM_REG_PPC_DAR | 64
1879 PPC | KVM_REG_PPC_DSISR | 32
1880 PPC | KVM_REG_PPC_AMR | 64
1881 PPC | KVM_REG_PPC_UAMOR | 64
1882 PPC | KVM_REG_PPC_MMCR0 | 64
1883 PPC | KVM_REG_PPC_MMCR1 | 64
1884 PPC | KVM_REG_PPC_MMCRA | 64
1885 PPC | KVM_REG_PPC_MMCR2 | 64
1886 PPC | KVM_REG_PPC_MMCRS | 64
1887 PPC | KVM_REG_PPC_SIAR | 64
1888 PPC | KVM_REG_PPC_SDAR | 64
1889 PPC | KVM_REG_PPC_SIER | 64
1890 PPC | KVM_REG_PPC_PMC1 | 32
1891 PPC | KVM_REG_PPC_PMC2 | 32
1892 PPC | KVM_REG_PPC_PMC3 | 32
1893 PPC | KVM_REG_PPC_PMC4 | 32
1894 PPC | KVM_REG_PPC_PMC5 | 32
1895 PPC | KVM_REG_PPC_PMC6 | 32
1896 PPC | KVM_REG_PPC_PMC7 | 32
1897 PPC | KVM_REG_PPC_PMC8 | 32
1898 PPC | KVM_REG_PPC_FPR0 | 64
1900 PPC | KVM_REG_PPC_FPR31 | 64
1901 PPC | KVM_REG_PPC_VR0 | 128
1903 PPC | KVM_REG_PPC_VR31 | 128
1904 PPC | KVM_REG_PPC_VSR0 | 128
1906 PPC | KVM_REG_PPC_VSR31 | 128
1907 PPC | KVM_REG_PPC_FPSCR | 64
1908 PPC | KVM_REG_PPC_VSCR | 32
1909 PPC | KVM_REG_PPC_VPA_ADDR | 64
1910 PPC | KVM_REG_PPC_VPA_SLB | 128
1911 PPC | KVM_REG_PPC_VPA_DTL | 128
1912 PPC | KVM_REG_PPC_EPCR | 32
1913 PPC | KVM_REG_PPC_EPR | 32
1914 PPC | KVM_REG_PPC_TCR | 32
1915 PPC | KVM_REG_PPC_TSR | 32
1916 PPC | KVM_REG_PPC_OR_TSR | 32
1917 PPC | KVM_REG_PPC_CLEAR_TSR | 32
1918 PPC | KVM_REG_PPC_MAS0 | 32
1919 PPC | KVM_REG_PPC_MAS1 | 32
1920 PPC | KVM_REG_PPC_MAS2 | 64
1921 PPC | KVM_REG_PPC_MAS7_3 | 64
1922 PPC | KVM_REG_PPC_MAS4 | 32
1923 PPC | KVM_REG_PPC_MAS6 | 32
1924 PPC | KVM_REG_PPC_MMUCFG | 32
1925 PPC | KVM_REG_PPC_TLB0CFG | 32
1926 PPC | KVM_REG_PPC_TLB1CFG | 32
1927 PPC | KVM_REG_PPC_TLB2CFG | 32
1928 PPC | KVM_REG_PPC_TLB3CFG | 32
1929 PPC | KVM_REG_PPC_TLB0PS | 32
1930 PPC | KVM_REG_PPC_TLB1PS | 32
1931 PPC | KVM_REG_PPC_TLB2PS | 32
1932 PPC | KVM_REG_PPC_TLB3PS | 32
1933 PPC | KVM_REG_PPC_EPTCFG | 32
1934 PPC | KVM_REG_PPC_ICP_STATE | 64
1935 PPC | KVM_REG_PPC_TB_OFFSET | 64
1936 PPC | KVM_REG_PPC_SPMC1 | 32
1937 PPC | KVM_REG_PPC_SPMC2 | 32
1938 PPC | KVM_REG_PPC_IAMR | 64
1939 PPC | KVM_REG_PPC_TFHAR | 64
1940 PPC | KVM_REG_PPC_TFIAR | 64
1941 PPC | KVM_REG_PPC_TEXASR | 64
1942 PPC | KVM_REG_PPC_FSCR | 64
1943 PPC | KVM_REG_PPC_PSPB | 32
1944 PPC | KVM_REG_PPC_EBBHR | 64
1945 PPC | KVM_REG_PPC_EBBRR | 64
1946 PPC | KVM_REG_PPC_BESCR | 64
1947 PPC | KVM_REG_PPC_TAR | 64
1948 PPC | KVM_REG_PPC_DPDES | 64
1949 PPC | KVM_REG_PPC_DAWR | 64
1950 PPC | KVM_REG_PPC_DAWRX | 64
1951 PPC | KVM_REG_PPC_CIABR | 64
1952 PPC | KVM_REG_PPC_IC | 64
1953 PPC | KVM_REG_PPC_VTB | 64
1954 PPC | KVM_REG_PPC_CSIGR | 64
1955 PPC | KVM_REG_PPC_TACR | 64
1956 PPC | KVM_REG_PPC_TCSCR | 64
1957 PPC | KVM_REG_PPC_PID | 64
1958 PPC | KVM_REG_PPC_ACOP | 64
1959 PPC | KVM_REG_PPC_VRSAVE | 32
1960 PPC | KVM_REG_PPC_LPCR | 32
1961 PPC | KVM_REG_PPC_LPCR_64 | 64
1962 PPC | KVM_REG_PPC_PPR | 64
1963 PPC | KVM_REG_PPC_ARCH_COMPAT | 32
1964 PPC | KVM_REG_PPC_DABRX | 32
1965 PPC | KVM_REG_PPC_WORT | 64
1966 PPC | KVM_REG_PPC_SPRG9 | 64
1967 PPC | KVM_REG_PPC_DBSR | 32
1968 PPC | KVM_REG_PPC_TM_GPR0 | 64
1970 PPC | KVM_REG_PPC_TM_GPR31 | 64
1971 PPC | KVM_REG_PPC_TM_VSR0 | 128
1973 PPC | KVM_REG_PPC_TM_VSR63 | 128
1974 PPC | KVM_REG_PPC_TM_CR | 64
1975 PPC | KVM_REG_PPC_TM_LR | 64
1976 PPC | KVM_REG_PPC_TM_CTR | 64
1977 PPC | KVM_REG_PPC_TM_FPSCR | 64
1978 PPC | KVM_REG_PPC_TM_AMR | 64
1979 PPC | KVM_REG_PPC_TM_PPR | 64
1980 PPC | KVM_REG_PPC_TM_VRSAVE | 64
1981 PPC | KVM_REG_PPC_TM_VSCR | 32
1982 PPC | KVM_REG_PPC_TM_DSCR | 64
1983 PPC | KVM_REG_PPC_TM_TAR | 64
1985 MIPS | KVM_REG_MIPS_R0 | 64
1987 MIPS | KVM_REG_MIPS_R31 | 64
1988 MIPS | KVM_REG_MIPS_HI | 64
1989 MIPS | KVM_REG_MIPS_LO | 64
1990 MIPS | KVM_REG_MIPS_PC | 64
1991 MIPS | KVM_REG_MIPS_CP0_INDEX | 32
1992 MIPS | KVM_REG_MIPS_CP0_CONTEXT | 64
1993 MIPS | KVM_REG_MIPS_CP0_USERLOCAL | 64
1994 MIPS | KVM_REG_MIPS_CP0_PAGEMASK | 32
1995 MIPS | KVM_REG_MIPS_CP0_WIRED | 32
1996 MIPS | KVM_REG_MIPS_CP0_HWRENA | 32
1997 MIPS | KVM_REG_MIPS_CP0_BADVADDR | 64
1998 MIPS | KVM_REG_MIPS_CP0_COUNT | 32
1999 MIPS | KVM_REG_MIPS_CP0_ENTRYHI | 64
2000 MIPS | KVM_REG_MIPS_CP0_COMPARE | 32
2001 MIPS | KVM_REG_MIPS_CP0_STATUS | 32
2002 MIPS | KVM_REG_MIPS_CP0_CAUSE | 32
2003 MIPS | KVM_REG_MIPS_CP0_EPC | 64
2004 MIPS | KVM_REG_MIPS_CP0_PRID | 32
2005 MIPS | KVM_REG_MIPS_CP0_CONFIG | 32
2006 MIPS | KVM_REG_MIPS_CP0_CONFIG1 | 32
2007 MIPS | KVM_REG_MIPS_CP0_CONFIG2 | 32
2008 MIPS | KVM_REG_MIPS_CP0_CONFIG3 | 32
2009 MIPS | KVM_REG_MIPS_CP0_CONFIG4 | 32
2010 MIPS | KVM_REG_MIPS_CP0_CONFIG5 | 32
2011 MIPS | KVM_REG_MIPS_CP0_CONFIG7 | 32
2012 MIPS | KVM_REG_MIPS_CP0_ERROREPC | 64
2013 MIPS | KVM_REG_MIPS_COUNT_CTL | 64
2014 MIPS | KVM_REG_MIPS_COUNT_RESUME | 64
2015 MIPS | KVM_REG_MIPS_COUNT_HZ | 64
2016 MIPS | KVM_REG_MIPS_FPR_32(0..31) | 32
2017 MIPS | KVM_REG_MIPS_FPR_64(0..31) | 64
2018 MIPS | KVM_REG_MIPS_VEC_128(0..31) | 128
2019 MIPS | KVM_REG_MIPS_FCR_IR | 32
2020 MIPS | KVM_REG_MIPS_FCR_CSR | 32
2021 MIPS | KVM_REG_MIPS_MSA_IR | 32
2022 MIPS | KVM_REG_MIPS_MSA_CSR | 32
2024 ARM registers are mapped using the lower 32 bits. The upper 16 of that
2025 is the register group type, or coprocessor number:
2027 ARM core registers have the following id bit patterns:
2028 0x4020 0000 0010 <index into the kvm_regs struct:16>
2030 ARM 32-bit CP15 registers have the following id bit patterns:
2031 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3>
2033 ARM 64-bit CP15 registers have the following id bit patterns:
2034 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3>
2036 ARM CCSIDR registers are demultiplexed by CSSELR value:
2037 0x4020 0000 0011 00 <csselr:8>
2039 ARM 32-bit VFP control registers have the following id bit patterns:
2040 0x4020 0000 0012 1 <regno:12>
2042 ARM 64-bit FP registers have the following id bit patterns:
2043 0x4030 0000 0012 0 <regno:12>
2046 arm64 registers are mapped using the lower 32 bits. The upper 16 of
2047 that is the register group type, or coprocessor number:
2049 arm64 core/FP-SIMD registers have the following id bit patterns. Note
2050 that the size of the access is variable, as the kvm_regs structure
2051 contains elements ranging from 32 to 128 bits. The index is a 32bit
2052 value in the kvm_regs structure seen as a 32bit array.
2053 0x60x0 0000 0010 <index into the kvm_regs struct:16>
2055 arm64 CCSIDR registers are demultiplexed by CSSELR value:
2056 0x6020 0000 0011 00 <csselr:8>
2058 arm64 system registers have the following id bit patterns:
2059 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3>
2062 MIPS registers are mapped using the lower 32 bits. The upper 16 of that is
2063 the register group type:
2065 MIPS core registers (see above) have the following id bit patterns:
2066 0x7030 0000 0000 <reg:16>
2068 MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit
2069 patterns depending on whether they're 32-bit or 64-bit registers:
2070 0x7020 0000 0001 00 <reg:5> <sel:3> (32-bit)
2071 0x7030 0000 0001 00 <reg:5> <sel:3> (64-bit)
2073 MIPS KVM control registers (see above) have the following id bit patterns:
2074 0x7030 0000 0002 <reg:16>
2076 MIPS FPU registers (see KVM_REG_MIPS_FPR_{32,64}() above) have the following
2077 id bit patterns depending on the size of the register being accessed. They are
2078 always accessed according to the current guest FPU mode (Status.FR and
2079 Config5.FRE), i.e. as the guest would see them, and they become unpredictable
2080 if the guest FPU mode is changed. MIPS SIMD Architecture (MSA) vector
2081 registers (see KVM_REG_MIPS_VEC_128() above) have similar patterns as they
2082 overlap the FPU registers:
2083 0x7020 0000 0003 00 <0:3> <reg:5> (32-bit FPU registers)
2084 0x7030 0000 0003 00 <0:3> <reg:5> (64-bit FPU registers)
2085 0x7040 0000 0003 00 <0:3> <reg:5> (128-bit MSA vector registers)
2087 MIPS FPU control registers (see KVM_REG_MIPS_FCR_{IR,CSR} above) have the
2088 following id bit patterns:
2089 0x7020 0000 0003 01 <0:3> <reg:5>
2091 MIPS MSA control registers (see KVM_REG_MIPS_MSA_{IR,CSR} above) have the
2092 following id bit patterns:
2093 0x7020 0000 0003 02 <0:3> <reg:5>
2096 4.69 KVM_GET_ONE_REG
2098 Capability: KVM_CAP_ONE_REG
2101 Parameters: struct kvm_one_reg (in and out)
2102 Returns: 0 on success, negative value on failure
2104 This ioctl allows to receive the value of a single register implemented
2105 in a vcpu. The register to read is indicated by the "id" field of the
2106 kvm_one_reg struct passed in. On success, the register value can be found
2107 at the memory location pointed to by "addr".
2109 The list of registers accessible using this interface is identical to the
2113 4.70 KVM_KVMCLOCK_CTRL
2115 Capability: KVM_CAP_KVMCLOCK_CTRL
2116 Architectures: Any that implement pvclocks (currently x86 only)
2119 Returns: 0 on success, -1 on error
2121 This signals to the host kernel that the specified guest is being paused by
2122 userspace. The host will set a flag in the pvclock structure that is checked
2123 from the soft lockup watchdog. The flag is part of the pvclock structure that
2124 is shared between guest and host, specifically the second bit of the flags
2125 field of the pvclock_vcpu_time_info structure. It will be set exclusively by
2126 the host and read/cleared exclusively by the guest. The guest operation of
2127 checking and clearing the flag must an atomic operation so
2128 load-link/store-conditional, or equivalent must be used. There are two cases
2129 where the guest will clear the flag: when the soft lockup watchdog timer resets
2130 itself or when a soft lockup is detected. This ioctl can be called any time
2131 after pausing the vcpu, but before it is resumed.
2136 Capability: KVM_CAP_SIGNAL_MSI
2139 Parameters: struct kvm_msi (in)
2140 Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error
2142 Directly inject a MSI message. Only valid with in-kernel irqchip that handles
2153 No flags are defined so far. The corresponding field must be 0.
2156 4.71 KVM_CREATE_PIT2
2158 Capability: KVM_CAP_PIT2
2161 Parameters: struct kvm_pit_config (in)
2162 Returns: 0 on success, -1 on error
2164 Creates an in-kernel device model for the i8254 PIT. This call is only valid
2165 after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following
2166 parameters have to be passed:
2168 struct kvm_pit_config {
2175 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */
2177 PIT timer interrupts may use a per-VM kernel thread for injection. If it
2178 exists, this thread will have a name of the following pattern:
2180 kvm-pit/<owner-process-pid>
2182 When running a guest with elevated priorities, the scheduling parameters of
2183 this thread may have to be adjusted accordingly.
2185 This IOCTL replaces the obsolete KVM_CREATE_PIT.
2190 Capability: KVM_CAP_PIT_STATE2
2193 Parameters: struct kvm_pit_state2 (out)
2194 Returns: 0 on success, -1 on error
2196 Retrieves the state of the in-kernel PIT model. Only valid after
2197 KVM_CREATE_PIT2. The state is returned in the following structure:
2199 struct kvm_pit_state2 {
2200 struct kvm_pit_channel_state channels[3];
2207 /* disable PIT in HPET legacy mode */
2208 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001
2210 This IOCTL replaces the obsolete KVM_GET_PIT.
2215 Capability: KVM_CAP_PIT_STATE2
2218 Parameters: struct kvm_pit_state2 (in)
2219 Returns: 0 on success, -1 on error
2221 Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2.
2222 See KVM_GET_PIT2 for details on struct kvm_pit_state2.
2224 This IOCTL replaces the obsolete KVM_SET_PIT.
2227 4.74 KVM_PPC_GET_SMMU_INFO
2229 Capability: KVM_CAP_PPC_GET_SMMU_INFO
2230 Architectures: powerpc
2233 Returns: 0 on success, -1 on error
2235 This populates and returns a structure describing the features of
2236 the "Server" class MMU emulation supported by KVM.
2237 This can in turn be used by userspace to generate the appropriate
2238 device-tree properties for the guest operating system.
2240 The structure contains some global information, followed by an
2241 array of supported segment page sizes:
2243 struct kvm_ppc_smmu_info {
2247 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ];
2250 The supported flags are:
2252 - KVM_PPC_PAGE_SIZES_REAL:
2253 When that flag is set, guest page sizes must "fit" the backing
2254 store page sizes. When not set, any page size in the list can
2255 be used regardless of how they are backed by userspace.
2257 - KVM_PPC_1T_SEGMENTS
2258 The emulated MMU supports 1T segments in addition to the
2261 The "slb_size" field indicates how many SLB entries are supported
2263 The "sps" array contains 8 entries indicating the supported base
2264 page sizes for a segment in increasing order. Each entry is defined
2267 struct kvm_ppc_one_seg_page_size {
2268 __u32 page_shift; /* Base page shift of segment (or 0) */
2269 __u32 slb_enc; /* SLB encoding for BookS */
2270 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ];
2273 An entry with a "page_shift" of 0 is unused. Because the array is
2274 organized in increasing order, a lookup can stop when encoutering
2277 The "slb_enc" field provides the encoding to use in the SLB for the
2278 page size. The bits are in positions such as the value can directly
2279 be OR'ed into the "vsid" argument of the slbmte instruction.
2281 The "enc" array is a list which for each of those segment base page
2282 size provides the list of supported actual page sizes (which can be
2283 only larger or equal to the base page size), along with the
2284 corresponding encoding in the hash PTE. Similarly, the array is
2285 8 entries sorted by increasing sizes and an entry with a "0" shift
2286 is an empty entry and a terminator:
2288 struct kvm_ppc_one_page_size {
2289 __u32 page_shift; /* Page shift (or 0) */
2290 __u32 pte_enc; /* Encoding in the HPTE (>>12) */
2293 The "pte_enc" field provides a value that can OR'ed into the hash
2294 PTE's RPN field (ie, it needs to be shifted left by 12 to OR it
2295 into the hash PTE second double word).
2299 Capability: KVM_CAP_IRQFD
2300 Architectures: x86 s390 arm arm64
2302 Parameters: struct kvm_irqfd (in)
2303 Returns: 0 on success, -1 on error
2305 Allows setting an eventfd to directly trigger a guest interrupt.
2306 kvm_irqfd.fd specifies the file descriptor to use as the eventfd and
2307 kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When
2308 an event is triggered on the eventfd, an interrupt is injected into
2309 the guest using the specified gsi pin. The irqfd is removed using
2310 the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd
2313 With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify
2314 mechanism allowing emulation of level-triggered, irqfd-based
2315 interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an
2316 additional eventfd in the kvm_irqfd.resamplefd field. When operating
2317 in resample mode, posting of an interrupt through kvm_irq.fd asserts
2318 the specified gsi in the irqchip. When the irqchip is resampled, such
2319 as from an EOI, the gsi is de-asserted and the user is notified via
2320 kvm_irqfd.resamplefd. It is the user's responsibility to re-queue
2321 the interrupt if the device making use of it still requires service.
2322 Note that closing the resamplefd is not sufficient to disable the
2323 irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment
2324 and need not be specified with KVM_IRQFD_FLAG_DEASSIGN.
2326 On ARM/ARM64, the gsi field in the kvm_irqfd struct specifies the Shared
2327 Peripheral Interrupt (SPI) index, such that the GIC interrupt ID is
2330 4.76 KVM_PPC_ALLOCATE_HTAB
2332 Capability: KVM_CAP_PPC_ALLOC_HTAB
2333 Architectures: powerpc
2335 Parameters: Pointer to u32 containing hash table order (in/out)
2336 Returns: 0 on success, -1 on error
2338 This requests the host kernel to allocate an MMU hash table for a
2339 guest using the PAPR paravirtualization interface. This only does
2340 anything if the kernel is configured to use the Book 3S HV style of
2341 virtualization. Otherwise the capability doesn't exist and the ioctl
2342 returns an ENOTTY error. The rest of this description assumes Book 3S
2345 There must be no vcpus running when this ioctl is called; if there
2346 are, it will do nothing and return an EBUSY error.
2348 The parameter is a pointer to a 32-bit unsigned integer variable
2349 containing the order (log base 2) of the desired size of the hash
2350 table, which must be between 18 and 46. On successful return from the
2351 ioctl, it will have been updated with the order of the hash table that
2354 If no hash table has been allocated when any vcpu is asked to run
2355 (with the KVM_RUN ioctl), the host kernel will allocate a
2356 default-sized hash table (16 MB).
2358 If this ioctl is called when a hash table has already been allocated,
2359 the kernel will clear out the existing hash table (zero all HPTEs) and
2360 return the hash table order in the parameter. (If the guest is using
2361 the virtualized real-mode area (VRMA) facility, the kernel will
2362 re-create the VMRA HPTEs on the next KVM_RUN of any vcpu.)
2364 4.77 KVM_S390_INTERRUPT
2368 Type: vm ioctl, vcpu ioctl
2369 Parameters: struct kvm_s390_interrupt (in)
2370 Returns: 0 on success, -1 on error
2372 Allows to inject an interrupt to the guest. Interrupts can be floating
2373 (vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type.
2375 Interrupt parameters are passed via kvm_s390_interrupt:
2377 struct kvm_s390_interrupt {
2383 type can be one of the following:
2385 KVM_S390_SIGP_STOP (vcpu) - sigp stop; optional flags in parm
2386 KVM_S390_PROGRAM_INT (vcpu) - program check; code in parm
2387 KVM_S390_SIGP_SET_PREFIX (vcpu) - sigp set prefix; prefix address in parm
2388 KVM_S390_RESTART (vcpu) - restart
2389 KVM_S390_INT_CLOCK_COMP (vcpu) - clock comparator interrupt
2390 KVM_S390_INT_CPU_TIMER (vcpu) - CPU timer interrupt
2391 KVM_S390_INT_VIRTIO (vm) - virtio external interrupt; external interrupt
2392 parameters in parm and parm64
2393 KVM_S390_INT_SERVICE (vm) - sclp external interrupt; sclp parameter in parm
2394 KVM_S390_INT_EMERGENCY (vcpu) - sigp emergency; source cpu in parm
2395 KVM_S390_INT_EXTERNAL_CALL (vcpu) - sigp external call; source cpu in parm
2396 KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) - compound value to indicate an
2397 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel);
2398 I/O interruption parameters in parm (subchannel) and parm64 (intparm,
2399 interruption subclass)
2400 KVM_S390_MCHK (vm, vcpu) - machine check interrupt; cr 14 bits in parm,
2401 machine check interrupt code in parm64 (note that
2402 machine checks needing further payload are not
2403 supported by this ioctl)
2405 Note that the vcpu ioctl is asynchronous to vcpu execution.
2407 4.78 KVM_PPC_GET_HTAB_FD
2409 Capability: KVM_CAP_PPC_HTAB_FD
2410 Architectures: powerpc
2412 Parameters: Pointer to struct kvm_get_htab_fd (in)
2413 Returns: file descriptor number (>= 0) on success, -1 on error
2415 This returns a file descriptor that can be used either to read out the
2416 entries in the guest's hashed page table (HPT), or to write entries to
2417 initialize the HPT. The returned fd can only be written to if the
2418 KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and
2419 can only be read if that bit is clear. The argument struct looks like
2422 /* For KVM_PPC_GET_HTAB_FD */
2423 struct kvm_get_htab_fd {
2429 /* Values for kvm_get_htab_fd.flags */
2430 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1)
2431 #define KVM_GET_HTAB_WRITE ((__u64)0x2)
2433 The `start_index' field gives the index in the HPT of the entry at
2434 which to start reading. It is ignored when writing.
2436 Reads on the fd will initially supply information about all
2437 "interesting" HPT entries. Interesting entries are those with the
2438 bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise
2439 all entries. When the end of the HPT is reached, the read() will
2440 return. If read() is called again on the fd, it will start again from
2441 the beginning of the HPT, but will only return HPT entries that have
2442 changed since they were last read.
2444 Data read or written is structured as a header (8 bytes) followed by a
2445 series of valid HPT entries (16 bytes) each. The header indicates how
2446 many valid HPT entries there are and how many invalid entries follow
2447 the valid entries. The invalid entries are not represented explicitly
2448 in the stream. The header format is:
2450 struct kvm_get_htab_header {
2456 Writes to the fd create HPT entries starting at the index given in the
2457 header; first `n_valid' valid entries with contents from the data
2458 written, then `n_invalid' invalid entries, invalidating any previously
2459 valid entries found.
2461 4.79 KVM_CREATE_DEVICE
2463 Capability: KVM_CAP_DEVICE_CTRL
2465 Parameters: struct kvm_create_device (in/out)
2466 Returns: 0 on success, -1 on error
2468 ENODEV: The device type is unknown or unsupported
2469 EEXIST: Device already created, and this type of device may not
2470 be instantiated multiple times
2472 Other error conditions may be defined by individual device types or
2473 have their standard meanings.
2475 Creates an emulated device in the kernel. The file descriptor returned
2476 in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR.
2478 If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the
2479 device type is supported (not necessarily whether it can be created
2482 Individual devices should not define flags. Attributes should be used
2483 for specifying any behavior that is not implied by the device type
2486 struct kvm_create_device {
2487 __u32 type; /* in: KVM_DEV_TYPE_xxx */
2488 __u32 fd; /* out: device handle */
2489 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */
2492 4.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR
2494 Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device
2495 Type: device ioctl, vm ioctl
2496 Parameters: struct kvm_device_attr
2497 Returns: 0 on success, -1 on error
2499 ENXIO: The group or attribute is unknown/unsupported for this device
2500 EPERM: The attribute cannot (currently) be accessed this way
2501 (e.g. read-only attribute, or attribute that only makes
2502 sense when the device is in a different state)
2504 Other error conditions may be defined by individual device types.
2506 Gets/sets a specified piece of device configuration and/or state. The
2507 semantics are device-specific. See individual device documentation in
2508 the "devices" directory. As with ONE_REG, the size of the data
2509 transferred is defined by the particular attribute.
2511 struct kvm_device_attr {
2512 __u32 flags; /* no flags currently defined */
2513 __u32 group; /* device-defined */
2514 __u64 attr; /* group-defined */
2515 __u64 addr; /* userspace address of attr data */
2518 4.81 KVM_HAS_DEVICE_ATTR
2520 Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device
2521 Type: device ioctl, vm ioctl
2522 Parameters: struct kvm_device_attr
2523 Returns: 0 on success, -1 on error
2525 ENXIO: The group or attribute is unknown/unsupported for this device
2527 Tests whether a device supports a particular attribute. A successful
2528 return indicates the attribute is implemented. It does not necessarily
2529 indicate that the attribute can be read or written in the device's
2530 current state. "addr" is ignored.
2532 4.82 KVM_ARM_VCPU_INIT
2535 Architectures: arm, arm64
2537 Parameters: struct kvm_vcpu_init (in)
2538 Returns: 0 on success; -1 on error
2540 Â EINVAL: Â Â Â the target is unknown, or the combination of features is invalid.
2541 Â ENOENT: Â Â Â a features bit specified is unknown.
2543 This tells KVM what type of CPU to present to the guest, and what
2544 optional features it should have. Â This will cause a reset of the cpu
2545 registers to their initial values. Â If this is not called, KVM_RUN will
2546 return ENOEXEC for that vcpu.
2548 Note that because some registers reflect machine topology, all vcpus
2549 should be created before this ioctl is invoked.
2551 Userspace can call this function multiple times for a given vcpu, including
2552 after the vcpu has been run. This will reset the vcpu to its initial
2553 state. All calls to this function after the initial call must use the same
2554 target and same set of feature flags, otherwise EINVAL will be returned.
2557 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state.
2558 Depends on KVM_CAP_ARM_PSCI. If not set, the CPU will be powered on
2559 and execute guest code when KVM_RUN is called.
2560 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode.
2561 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only).
2562 - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 for the CPU.
2563 Depends on KVM_CAP_ARM_PSCI_0_2.
2566 4.83 KVM_ARM_PREFERRED_TARGET
2569 Architectures: arm, arm64
2571 Parameters: struct struct kvm_vcpu_init (out)
2572 Returns: 0 on success; -1 on error
2574 ENODEV: no preferred target available for the host
2576 This queries KVM for preferred CPU target type which can be emulated
2577 by KVM on underlying host.
2579 The ioctl returns struct kvm_vcpu_init instance containing information
2580 about preferred CPU target type and recommended features for it. The
2581 kvm_vcpu_init->features bitmap returned will have feature bits set if
2582 the preferred target recommends setting these features, but this is
2585 The information returned by this ioctl can be used to prepare an instance
2586 of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in
2587 in VCPU matching underlying host.
2590 4.84 KVM_GET_REG_LIST
2593 Architectures: arm, arm64, mips
2595 Parameters: struct kvm_reg_list (in/out)
2596 Returns: 0 on success; -1 on error
2598 Â E2BIG: Â Â Â Â the reg index list is too big to fit in the array specified by
2599 Â Â Â Â Â Â Â Â Â Â Â Â the user (the number required will be written into n).
2601 struct kvm_reg_list {
2602 __u64 n; /* number of registers in reg[] */
2606 This ioctl returns the guest registers that are supported for the
2607 KVM_GET_ONE_REG/KVM_SET_ONE_REG calls.
2610 4.85 KVM_ARM_SET_DEVICE_ADDR (deprecated)
2612 Capability: KVM_CAP_ARM_SET_DEVICE_ADDR
2613 Architectures: arm, arm64
2615 Parameters: struct kvm_arm_device_address (in)
2616 Returns: 0 on success, -1 on error
2618 ENODEV: The device id is unknown
2619 ENXIO: Device not supported on current system
2620 EEXIST: Address already set
2621 E2BIG: Address outside guest physical address space
2622 EBUSY: Address overlaps with other device range
2624 struct kvm_arm_device_addr {
2629 Specify a device address in the guest's physical address space where guests
2630 can access emulated or directly exposed devices, which the host kernel needs
2631 to know about. The id field is an architecture specific identifier for a
2634 ARM/arm64 divides the id field into two parts, a device id and an
2635 address type id specific to the individual device.
2637 Â bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 |
2638 field: | 0x00000000 | device id | addr type id |
2640 ARM/arm64 currently only require this when using the in-kernel GIC
2641 support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2
2642 as the device id. When setting the base address for the guest's
2643 mapping of the VGIC virtual CPU and distributor interface, the ioctl
2644 must be called after calling KVM_CREATE_IRQCHIP, but before calling
2645 KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the
2646 base addresses will return -EEXIST.
2648 Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API
2649 should be used instead.
2652 4.86 KVM_PPC_RTAS_DEFINE_TOKEN
2654 Capability: KVM_CAP_PPC_RTAS
2657 Parameters: struct kvm_rtas_token_args
2658 Returns: 0 on success, -1 on error
2660 Defines a token value for a RTAS (Run Time Abstraction Services)
2661 service in order to allow it to be handled in the kernel. The
2662 argument struct gives the name of the service, which must be the name
2663 of a service that has a kernel-side implementation. If the token
2664 value is non-zero, it will be associated with that service, and
2665 subsequent RTAS calls by the guest specifying that token will be
2666 handled by the kernel. If the token value is 0, then any token
2667 associated with the service will be forgotten, and subsequent RTAS
2668 calls by the guest for that service will be passed to userspace to be
2671 4.87 KVM_SET_GUEST_DEBUG
2673 Capability: KVM_CAP_SET_GUEST_DEBUG
2674 Architectures: x86, s390, ppc, arm64
2676 Parameters: struct kvm_guest_debug (in)
2677 Returns: 0 on success; -1 on error
2679 struct kvm_guest_debug {
2682 struct kvm_guest_debug_arch arch;
2685 Set up the processor specific debug registers and configure vcpu for
2686 handling guest debug events. There are two parts to the structure, the
2687 first a control bitfield indicates the type of debug events to handle
2688 when running. Common control bits are:
2690 - KVM_GUESTDBG_ENABLE: guest debugging is enabled
2691 - KVM_GUESTDBG_SINGLESTEP: the next run should single-step
2693 The top 16 bits of the control field are architecture specific control
2694 flags which can include the following:
2696 - KVM_GUESTDBG_USE_SW_BP: using software breakpoints [x86, arm64]
2697 - KVM_GUESTDBG_USE_HW_BP: using hardware breakpoints [x86, s390, arm64]
2698 - KVM_GUESTDBG_INJECT_DB: inject DB type exception [x86]
2699 - KVM_GUESTDBG_INJECT_BP: inject BP type exception [x86]
2700 - KVM_GUESTDBG_EXIT_PENDING: trigger an immediate guest exit [s390]
2702 For example KVM_GUESTDBG_USE_SW_BP indicates that software breakpoints
2703 are enabled in memory so we need to ensure breakpoint exceptions are
2704 correctly trapped and the KVM run loop exits at the breakpoint and not
2705 running off into the normal guest vector. For KVM_GUESTDBG_USE_HW_BP
2706 we need to ensure the guest vCPUs architecture specific registers are
2707 updated to the correct (supplied) values.
2709 The second part of the structure is architecture specific and
2710 typically contains a set of debug registers.
2712 For arm64 the number of debug registers is implementation defined and
2713 can be determined by querying the KVM_CAP_GUEST_DEBUG_HW_BPS and
2714 KVM_CAP_GUEST_DEBUG_HW_WPS capabilities which return a positive number
2715 indicating the number of supported registers.
2717 When debug events exit the main run loop with the reason
2718 KVM_EXIT_DEBUG with the kvm_debug_exit_arch part of the kvm_run
2719 structure containing architecture specific debug information.
2721 4.88 KVM_GET_EMULATED_CPUID
2723 Capability: KVM_CAP_EXT_EMUL_CPUID
2726 Parameters: struct kvm_cpuid2 (in/out)
2727 Returns: 0 on success, -1 on error
2732 struct kvm_cpuid_entry2 entries[0];
2735 The member 'flags' is used for passing flags from userspace.
2737 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
2738 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
2739 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
2741 struct kvm_cpuid_entry2 {
2752 This ioctl returns x86 cpuid features which are emulated by
2753 kvm.Userspace can use the information returned by this ioctl to query
2754 which features are emulated by kvm instead of being present natively.
2756 Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2
2757 structure with the 'nent' field indicating the number of entries in
2758 the variable-size array 'entries'. If the number of entries is too low
2759 to describe the cpu capabilities, an error (E2BIG) is returned. If the
2760 number is too high, the 'nent' field is adjusted and an error (ENOMEM)
2761 is returned. If the number is just right, the 'nent' field is adjusted
2762 to the number of valid entries in the 'entries' array, which is then
2765 The entries returned are the set CPUID bits of the respective features
2766 which kvm emulates, as returned by the CPUID instruction, with unknown
2767 or unsupported feature bits cleared.
2769 Features like x2apic, for example, may not be present in the host cpu
2770 but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be
2771 emulated efficiently and thus not included here.
2773 The fields in each entry are defined as follows:
2775 function: the eax value used to obtain the entry
2776 index: the ecx value used to obtain the entry (for entries that are
2778 flags: an OR of zero or more of the following:
2779 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
2780 if the index field is valid
2781 KVM_CPUID_FLAG_STATEFUL_FUNC:
2782 if cpuid for this function returns different values for successive
2783 invocations; there will be several entries with the same function,
2784 all with this flag set
2785 KVM_CPUID_FLAG_STATE_READ_NEXT:
2786 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
2787 the first entry to be read by a cpu
2788 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
2789 this function/index combination
2791 4.89 KVM_S390_MEM_OP
2793 Capability: KVM_CAP_S390_MEM_OP
2796 Parameters: struct kvm_s390_mem_op (in)
2797 Returns: = 0 on success,
2798 < 0 on generic error (e.g. -EFAULT or -ENOMEM),
2799 > 0 if an exception occurred while walking the page tables
2801 Read or write data from/to the logical (virtual) memory of a VPCU.
2803 Parameters are specified via the following structure:
2805 struct kvm_s390_mem_op {
2806 __u64 gaddr; /* the guest address */
2807 __u64 flags; /* flags */
2808 __u32 size; /* amount of bytes */
2809 __u32 op; /* type of operation */
2810 __u64 buf; /* buffer in userspace */
2811 __u8 ar; /* the access register number */
2812 __u8 reserved[31]; /* should be set to 0 */
2815 The type of operation is specified in the "op" field. It is either
2816 KVM_S390_MEMOP_LOGICAL_READ for reading from logical memory space or
2817 KVM_S390_MEMOP_LOGICAL_WRITE for writing to logical memory space. The
2818 KVM_S390_MEMOP_F_CHECK_ONLY flag can be set in the "flags" field to check
2819 whether the corresponding memory access would create an access exception
2820 (without touching the data in the memory at the destination). In case an
2821 access exception occurred while walking the MMU tables of the guest, the
2822 ioctl returns a positive error number to indicate the type of exception.
2823 This exception is also raised directly at the corresponding VCPU if the
2824 flag KVM_S390_MEMOP_F_INJECT_EXCEPTION is set in the "flags" field.
2826 The start address of the memory region has to be specified in the "gaddr"
2827 field, and the length of the region in the "size" field. "buf" is the buffer
2828 supplied by the userspace application where the read data should be written
2829 to for KVM_S390_MEMOP_LOGICAL_READ, or where the data that should be written
2830 is stored for a KVM_S390_MEMOP_LOGICAL_WRITE. "buf" is unused and can be NULL
2831 when KVM_S390_MEMOP_F_CHECK_ONLY is specified. "ar" designates the access
2832 register number to be used.
2834 The "reserved" field is meant for future extensions. It is not used by
2835 KVM with the currently defined set of flags.
2837 4.90 KVM_S390_GET_SKEYS
2839 Capability: KVM_CAP_S390_SKEYS
2842 Parameters: struct kvm_s390_skeys
2843 Returns: 0 on success, KVM_S390_GET_KEYS_NONE if guest is not using storage
2844 keys, negative value on error
2846 This ioctl is used to get guest storage key values on the s390
2847 architecture. The ioctl takes parameters via the kvm_s390_skeys struct.
2849 struct kvm_s390_skeys {
2852 __u64 skeydata_addr;
2857 The start_gfn field is the number of the first guest frame whose storage keys
2860 The count field is the number of consecutive frames (starting from start_gfn)
2861 whose storage keys to get. The count field must be at least 1 and the maximum
2862 allowed value is defined as KVM_S390_SKEYS_ALLOC_MAX. Values outside this range
2863 will cause the ioctl to return -EINVAL.
2865 The skeydata_addr field is the address to a buffer large enough to hold count
2866 bytes. This buffer will be filled with storage key data by the ioctl.
2868 4.91 KVM_S390_SET_SKEYS
2870 Capability: KVM_CAP_S390_SKEYS
2873 Parameters: struct kvm_s390_skeys
2874 Returns: 0 on success, negative value on error
2876 This ioctl is used to set guest storage key values on the s390
2877 architecture. The ioctl takes parameters via the kvm_s390_skeys struct.
2878 See section on KVM_S390_GET_SKEYS for struct definition.
2880 The start_gfn field is the number of the first guest frame whose storage keys
2883 The count field is the number of consecutive frames (starting from start_gfn)
2884 whose storage keys to get. The count field must be at least 1 and the maximum
2885 allowed value is defined as KVM_S390_SKEYS_ALLOC_MAX. Values outside this range
2886 will cause the ioctl to return -EINVAL.
2888 The skeydata_addr field is the address to a buffer containing count bytes of
2889 storage keys. Each byte in the buffer will be set as the storage key for a
2890 single frame starting at start_gfn for count frames.
2892 Note: If any architecturally invalid key value is found in the given data then
2893 the ioctl will return -EINVAL.
2897 Capability: KVM_CAP_S390_INJECT_IRQ
2900 Parameters: struct kvm_s390_irq (in)
2901 Returns: 0 on success, -1 on error
2903 EINVAL: interrupt type is invalid
2904 type is KVM_S390_SIGP_STOP and flag parameter is invalid value
2905 type is KVM_S390_INT_EXTERNAL_CALL and code is bigger
2906 than the maximum of VCPUs
2907 EBUSY: type is KVM_S390_SIGP_SET_PREFIX and vcpu is not stopped
2908 type is KVM_S390_SIGP_STOP and a stop irq is already pending
2909 type is KVM_S390_INT_EXTERNAL_CALL and an external call interrupt
2912 Allows to inject an interrupt to the guest.
2914 Using struct kvm_s390_irq as a parameter allows
2915 to inject additional payload which is not
2916 possible via KVM_S390_INTERRUPT.
2918 Interrupt parameters are passed via kvm_s390_irq:
2920 struct kvm_s390_irq {
2923 struct kvm_s390_io_info io;
2924 struct kvm_s390_ext_info ext;
2925 struct kvm_s390_pgm_info pgm;
2926 struct kvm_s390_emerg_info emerg;
2927 struct kvm_s390_extcall_info extcall;
2928 struct kvm_s390_prefix_info prefix;
2929 struct kvm_s390_stop_info stop;
2930 struct kvm_s390_mchk_info mchk;
2935 type can be one of the following:
2937 KVM_S390_SIGP_STOP - sigp stop; parameter in .stop
2938 KVM_S390_PROGRAM_INT - program check; parameters in .pgm
2939 KVM_S390_SIGP_SET_PREFIX - sigp set prefix; parameters in .prefix
2940 KVM_S390_RESTART - restart; no parameters
2941 KVM_S390_INT_CLOCK_COMP - clock comparator interrupt; no parameters
2942 KVM_S390_INT_CPU_TIMER - CPU timer interrupt; no parameters
2943 KVM_S390_INT_EMERGENCY - sigp emergency; parameters in .emerg
2944 KVM_S390_INT_EXTERNAL_CALL - sigp external call; parameters in .extcall
2945 KVM_S390_MCHK - machine check interrupt; parameters in .mchk
2948 Note that the vcpu ioctl is asynchronous to vcpu execution.
2950 4.94 KVM_S390_GET_IRQ_STATE
2952 Capability: KVM_CAP_S390_IRQ_STATE
2955 Parameters: struct kvm_s390_irq_state (out)
2956 Returns: >= number of bytes copied into buffer,
2957 -EINVAL if buffer size is 0,
2958 -ENOBUFS if buffer size is too small to fit all pending interrupts,
2959 -EFAULT if the buffer address was invalid
2961 This ioctl allows userspace to retrieve the complete state of all currently
2962 pending interrupts in a single buffer. Use cases include migration
2963 and introspection. The parameter structure contains the address of a
2964 userspace buffer and its length:
2966 struct kvm_s390_irq_state {
2973 Userspace passes in the above struct and for each pending interrupt a
2974 struct kvm_s390_irq is copied to the provided buffer.
2976 If -ENOBUFS is returned the buffer provided was too small and userspace
2977 may retry with a bigger buffer.
2979 4.95 KVM_S390_SET_IRQ_STATE
2981 Capability: KVM_CAP_S390_IRQ_STATE
2984 Parameters: struct kvm_s390_irq_state (in)
2985 Returns: 0 on success,
2986 -EFAULT if the buffer address was invalid,
2987 -EINVAL for an invalid buffer length (see below),
2988 -EBUSY if there were already interrupts pending,
2989 errors occurring when actually injecting the
2990 interrupt. See KVM_S390_IRQ.
2992 This ioctl allows userspace to set the complete state of all cpu-local
2993 interrupts currently pending for the vcpu. It is intended for restoring
2994 interrupt state after a migration. The input parameter is a userspace buffer
2995 containing a struct kvm_s390_irq_state:
2997 struct kvm_s390_irq_state {
3003 The userspace memory referenced by buf contains a struct kvm_s390_irq
3004 for each interrupt to be injected into the guest.
3005 If one of the interrupts could not be injected for some reason the
3008 len must be a multiple of sizeof(struct kvm_s390_irq). It must be > 0
3009 and it must not exceed (max_vcpus + 32) * sizeof(struct kvm_s390_irq),
3010 which is the maximum number of possibly pending cpu-local interrupts.
3014 Capability: KVM_CAP_X86_SMM
3018 Returns: 0 on success, -1 on error
3020 Queues an SMI on the thread's vcpu.
3022 5. The kvm_run structure
3023 ------------------------
3025 Application code obtains a pointer to the kvm_run structure by
3026 mmap()ing a vcpu fd. From that point, application code can control
3027 execution by changing fields in kvm_run prior to calling the KVM_RUN
3028 ioctl, and obtain information about the reason KVM_RUN returned by
3029 looking up structure members.
3033 __u8 request_interrupt_window;
3035 Request that KVM_RUN return when it becomes possible to inject external
3036 interrupts into the guest. Useful in conjunction with KVM_INTERRUPT.
3043 When KVM_RUN has returned successfully (return value 0), this informs
3044 application code why KVM_RUN has returned. Allowable values for this
3045 field are detailed below.
3047 __u8 ready_for_interrupt_injection;
3049 If request_interrupt_window has been specified, this field indicates
3050 an interrupt can be injected now with KVM_INTERRUPT.
3054 The value of the current interrupt flag. Only valid if in-kernel
3055 local APIC is not used.
3059 More architecture-specific flags detailing state of the VCPU that may
3060 affect the device's behavior. The only currently defined flag is
3061 KVM_RUN_X86_SMM, which is valid on x86 machines and is set if the
3062 VCPU is in system management mode.
3064 /* in (pre_kvm_run), out (post_kvm_run) */
3067 The value of the cr8 register. Only valid if in-kernel local APIC is
3068 not used. Both input and output.
3072 The value of the APIC BASE msr. Only valid if in-kernel local
3073 APIC is not used. Both input and output.
3076 /* KVM_EXIT_UNKNOWN */
3078 __u64 hardware_exit_reason;
3081 If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown
3082 reasons. Further architecture-specific information is available in
3083 hardware_exit_reason.
3085 /* KVM_EXIT_FAIL_ENTRY */
3087 __u64 hardware_entry_failure_reason;
3090 If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due
3091 to unknown reasons. Further architecture-specific information is
3092 available in hardware_entry_failure_reason.
3094 /* KVM_EXIT_EXCEPTION */
3104 #define KVM_EXIT_IO_IN 0
3105 #define KVM_EXIT_IO_OUT 1
3107 __u8 size; /* bytes */
3110 __u64 data_offset; /* relative to kvm_run start */
3113 If exit_reason is KVM_EXIT_IO, then the vcpu has
3114 executed a port I/O instruction which could not be satisfied by kvm.
3115 data_offset describes where the data is located (KVM_EXIT_IO_OUT) or
3116 where kvm expects application code to place the data for the next
3117 KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array.
3119 /* KVM_EXIT_DEBUG */
3121 struct kvm_debug_exit_arch arch;
3124 If the exit_reason is KVM_EXIT_DEBUG, then a vcpu is processing a debug event
3125 for which architecture specific information is returned.
3135 If exit_reason is KVM_EXIT_MMIO, then the vcpu has
3136 executed a memory-mapped I/O instruction which could not be satisfied
3137 by kvm. The 'data' member contains the written data if 'is_write' is
3138 true, and should be filled by application code otherwise.
3140 The 'data' member contains, in its first 'len' bytes, the value as it would
3141 appear if the VCPU performed a load or store of the appropriate width directly
3144 NOTE: For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_PAPR and
3145 KVM_EXIT_EPR the corresponding
3146 operations are complete (and guest state is consistent) only after userspace
3147 has re-entered the kernel with KVM_RUN. The kernel side will first finish
3148 incomplete operations and then check for pending signals. Userspace
3149 can re-enter the guest with an unmasked signal pending to complete
3152 /* KVM_EXIT_HYPERCALL */
3161 Unused. This was once used for 'hypercall to userspace'. To implement
3162 such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390).
3163 Note KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO.
3165 /* KVM_EXIT_TPR_ACCESS */
3172 To be documented (KVM_TPR_ACCESS_REPORTING).
3174 /* KVM_EXIT_S390_SIEIC */
3177 __u64 mask; /* psw upper half */
3178 __u64 addr; /* psw lower half */
3185 /* KVM_EXIT_S390_RESET */
3186 #define KVM_S390_RESET_POR 1
3187 #define KVM_S390_RESET_CLEAR 2
3188 #define KVM_S390_RESET_SUBSYSTEM 4
3189 #define KVM_S390_RESET_CPU_INIT 8
3190 #define KVM_S390_RESET_IPL 16
3191 __u64 s390_reset_flags;
3195 /* KVM_EXIT_S390_UCONTROL */
3197 __u64 trans_exc_code;
3201 s390 specific. A page fault has occurred for a user controlled virtual
3202 machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be
3203 resolved by the kernel.
3204 The program code and the translation exception code that were placed
3205 in the cpu's lowcore are presented here as defined by the z Architecture
3206 Principles of Operation Book in the Chapter for Dynamic Address Translation
3216 Deprecated - was used for 440 KVM.
3223 MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch
3224 hypercalls and exit with this exit struct that contains all the guest gprs.
3226 If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall.
3227 Userspace can now handle the hypercall and when it's done modify the gprs as
3228 necessary. Upon guest entry all guest GPRs will then be replaced by the values
3231 /* KVM_EXIT_PAPR_HCALL */
3238 This is used on 64-bit PowerPC when emulating a pSeries partition,
3239 e.g. with the 'pseries' machine type in qemu. It occurs when the
3240 guest does a hypercall using the 'sc 1' instruction. The 'nr' field
3241 contains the hypercall number (from the guest R3), and 'args' contains
3242 the arguments (from the guest R4 - R12). Userspace should put the
3243 return code in 'ret' and any extra returned values in args[].
3244 The possible hypercalls are defined in the Power Architecture Platform
3245 Requirements (PAPR) document available from www.power.org (free
3246 developer registration required to access it).
3248 /* KVM_EXIT_S390_TSCH */
3250 __u16 subchannel_id;
3251 __u16 subchannel_nr;
3258 s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled
3259 and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O
3260 interrupt for the target subchannel has been dequeued and subchannel_id,
3261 subchannel_nr, io_int_parm and io_int_word contain the parameters for that
3262 interrupt. ipb is needed for instruction parameter decoding.
3269 On FSL BookE PowerPC chips, the interrupt controller has a fast patch
3270 interrupt acknowledge path to the core. When the core successfully
3271 delivers an interrupt, it automatically populates the EPR register with
3272 the interrupt vector number and acknowledges the interrupt inside
3273 the interrupt controller.
3275 In case the interrupt controller lives in user space, we need to do
3276 the interrupt acknowledge cycle through it to fetch the next to be
3277 delivered interrupt vector using this exit.
3279 It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an
3280 external interrupt has just been delivered into the guest. User space
3281 should put the acknowledged interrupt vector into the 'epr' field.
3283 /* KVM_EXIT_SYSTEM_EVENT */
3285 #define KVM_SYSTEM_EVENT_SHUTDOWN 1
3286 #define KVM_SYSTEM_EVENT_RESET 2
3287 #define KVM_SYSTEM_EVENT_CRASH 3
3292 If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered
3293 a system-level event using some architecture specific mechanism (hypercall
3294 or some special instruction). In case of ARM/ARM64, this is triggered using
3295 HVC instruction based PSCI call from the vcpu. The 'type' field describes
3296 the system-level event type. The 'flags' field describes architecture
3297 specific flags for the system-level event.
3299 Valid values for 'type' are:
3300 KVM_SYSTEM_EVENT_SHUTDOWN -- the guest has requested a shutdown of the
3301 VM. Userspace is not obliged to honour this, and if it does honour
3302 this does not need to destroy the VM synchronously (ie it may call
3303 KVM_RUN again before shutdown finally occurs).
3304 KVM_SYSTEM_EVENT_RESET -- the guest has requested a reset of the VM.
3305 As with SHUTDOWN, userspace can choose to ignore the request, or
3306 to schedule the reset to occur in the future and may call KVM_RUN again.
3307 KVM_SYSTEM_EVENT_CRASH -- the guest crash occurred and the guest
3308 has requested a crash condition maintenance. Userspace can choose
3309 to ignore the request, or to gather VM memory core dump and/or
3310 reset/shutdown of the VM.
3312 /* KVM_EXIT_IOAPIC_EOI */
3317 Indicates that the VCPU's in-kernel local APIC received an EOI for a
3318 level-triggered IOAPIC interrupt. This exit only triggers when the
3319 IOAPIC is implemented in userspace (i.e. KVM_CAP_SPLIT_IRQCHIP is enabled);
3320 the userspace IOAPIC should process the EOI and retrigger the interrupt if
3321 it is still asserted. Vector is the LAPIC interrupt vector for which the
3324 /* Fix the size of the union. */
3329 * shared registers between kvm and userspace.
3330 * kvm_valid_regs specifies the register classes set by the host
3331 * kvm_dirty_regs specified the register classes dirtied by userspace
3332 * struct kvm_sync_regs is architecture specific, as well as the
3333 * bits for kvm_valid_regs and kvm_dirty_regs
3335 __u64 kvm_valid_regs;
3336 __u64 kvm_dirty_regs;
3338 struct kvm_sync_regs regs;
3342 If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access
3343 certain guest registers without having to call SET/GET_*REGS. Thus we can
3344 avoid some system call overhead if userspace has to handle the exit.
3345 Userspace can query the validity of the structure by checking
3346 kvm_valid_regs for specific bits. These bits are architecture specific
3347 and usually define the validity of a groups of registers. (e.g. one bit
3348 for general purpose registers)
3350 Please note that the kernel is allowed to use the kvm_run structure as the
3351 primary storage for certain register types. Therefore, the kernel may use the
3352 values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set.
3358 6. Capabilities that can be enabled on vCPUs
3359 --------------------------------------------
3361 There are certain capabilities that change the behavior of the virtual CPU or
3362 the virtual machine when enabled. To enable them, please see section 4.37.
3363 Below you can find a list of capabilities and what their effect on the vCPU or
3364 the virtual machine is when enabling them.
3366 The following information is provided along with the description:
3368 Architectures: which instruction set architectures provide this ioctl.
3369 x86 includes both i386 and x86_64.
3371 Target: whether this is a per-vcpu or per-vm capability.
3373 Parameters: what parameters are accepted by the capability.
3375 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
3376 are not detailed, but errors with specific meanings are.
3384 Returns: 0 on success; -1 on error
3386 This capability enables interception of OSI hypercalls that otherwise would
3387 be treated as normal system calls to be injected into the guest. OSI hypercalls
3388 were invented by Mac-on-Linux to have a standardized communication mechanism
3389 between the guest and the host.
3391 When this capability is enabled, KVM_EXIT_OSI can occur.
3394 6.2 KVM_CAP_PPC_PAPR
3399 Returns: 0 on success; -1 on error
3401 This capability enables interception of PAPR hypercalls. PAPR hypercalls are
3402 done using the hypercall instruction "sc 1".
3404 It also sets the guest privilege level to "supervisor" mode. Usually the guest
3405 runs in "hypervisor" privilege mode with a few missing features.
3407 In addition to the above, it changes the semantics of SDR1. In this mode, the
3408 HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the
3409 HTAB invisible to the guest.
3411 When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur.
3418 Parameters: args[0] is the address of a struct kvm_config_tlb
3419 Returns: 0 on success; -1 on error
3421 struct kvm_config_tlb {
3428 Configures the virtual CPU's TLB array, establishing a shared memory area
3429 between userspace and KVM. The "params" and "array" fields are userspace
3430 addresses of mmu-type-specific data structures. The "array_len" field is an
3431 safety mechanism, and should be set to the size in bytes of the memory that
3432 userspace has reserved for the array. It must be at least the size dictated
3433 by "mmu_type" and "params".
3435 While KVM_RUN is active, the shared region is under control of KVM. Its
3436 contents are undefined, and any modification by userspace results in
3437 boundedly undefined behavior.
3439 On return from KVM_RUN, the shared region will reflect the current state of
3440 the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB
3441 to tell KVM which entries have been changed, prior to calling KVM_RUN again
3444 For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV:
3445 - The "params" field is of type "struct kvm_book3e_206_tlb_params".
3446 - The "array" field points to an array of type "struct
3447 kvm_book3e_206_tlb_entry".
3448 - The array consists of all entries in the first TLB, followed by all
3449 entries in the second TLB.
3450 - Within a TLB, entries are ordered first by increasing set number. Within a
3451 set, entries are ordered by way (increasing ESEL).
3452 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1)
3453 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value.
3454 - The tsize field of mas1 shall be set to 4K on TLB0, even though the
3455 hardware ignores this value for TLB0.
3457 6.4 KVM_CAP_S390_CSS_SUPPORT
3462 Returns: 0 on success; -1 on error
3464 This capability enables support for handling of channel I/O instructions.
3466 TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are
3467 handled in-kernel, while the other I/O instructions are passed to userspace.
3469 When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST
3470 SUBCHANNEL intercepts.
3472 Note that even though this capability is enabled per-vcpu, the complete
3473 virtual machine is affected.
3479 Parameters: args[0] defines whether the proxy facility is active
3480 Returns: 0 on success; -1 on error
3482 This capability enables or disables the delivery of interrupts through the
3483 external proxy facility.
3485 When enabled (args[0] != 0), every time the guest gets an external interrupt
3486 delivered, it automatically exits into user space with a KVM_EXIT_EPR exit
3487 to receive the topmost interrupt vector.
3489 When disabled (args[0] == 0), behavior is as if this facility is unsupported.
3491 When this capability is enabled, KVM_EXIT_EPR can occur.
3493 6.6 KVM_CAP_IRQ_MPIC
3496 Parameters: args[0] is the MPIC device fd
3497 args[1] is the MPIC CPU number for this vcpu
3499 This capability connects the vcpu to an in-kernel MPIC device.
3501 6.7 KVM_CAP_IRQ_XICS
3505 Parameters: args[0] is the XICS device fd
3506 args[1] is the XICS CPU number (server ID) for this vcpu
3508 This capability connects the vcpu to an in-kernel XICS device.
3510 6.8 KVM_CAP_S390_IRQCHIP
3516 This capability enables the in-kernel irqchip for s390. Please refer to
3517 "4.24 KVM_CREATE_IRQCHIP" for details.
3519 6.9 KVM_CAP_MIPS_FPU
3523 Parameters: args[0] is reserved for future use (should be 0).
3525 This capability allows the use of the host Floating Point Unit by the guest. It
3526 allows the Config1.FP bit to be set to enable the FPU in the guest. Once this is
3527 done the KVM_REG_MIPS_FPR_* and KVM_REG_MIPS_FCR_* registers can be accessed
3528 (depending on the current guest FPU register mode), and the Status.FR,
3529 Config5.FRE bits are accessible via the KVM API and also from the guest,
3530 depending on them being supported by the FPU.
3532 6.10 KVM_CAP_MIPS_MSA
3536 Parameters: args[0] is reserved for future use (should be 0).
3538 This capability allows the use of the MIPS SIMD Architecture (MSA) by the guest.
3539 It allows the Config3.MSAP bit to be set to enable the use of MSA by the guest.
3540 Once this is done the KVM_REG_MIPS_VEC_* and KVM_REG_MIPS_MSA_* registers can be
3541 accessed, and the Config5.MSAEn bit is accessible via the KVM API and also from
3544 7. Capabilities that can be enabled on VMs
3545 ------------------------------------------
3547 There are certain capabilities that change the behavior of the virtual
3548 machine when enabled. To enable them, please see section 4.37. Below
3549 you can find a list of capabilities and what their effect on the VM
3550 is when enabling them.
3552 The following information is provided along with the description:
3554 Architectures: which instruction set architectures provide this ioctl.
3555 x86 includes both i386 and x86_64.
3557 Parameters: what parameters are accepted by the capability.
3559 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
3560 are not detailed, but errors with specific meanings are.
3563 7.1 KVM_CAP_PPC_ENABLE_HCALL
3566 Parameters: args[0] is the sPAPR hcall number
3567 args[1] is 0 to disable, 1 to enable in-kernel handling
3569 This capability controls whether individual sPAPR hypercalls (hcalls)
3570 get handled by the kernel or not. Enabling or disabling in-kernel
3571 handling of an hcall is effective across the VM. On creation, an
3572 initial set of hcalls are enabled for in-kernel handling, which
3573 consists of those hcalls for which in-kernel handlers were implemented
3574 before this capability was implemented. If disabled, the kernel will
3575 not to attempt to handle the hcall, but will always exit to userspace
3576 to handle it. Note that it may not make sense to enable some and
3577 disable others of a group of related hcalls, but KVM does not prevent
3578 userspace from doing that.
3580 If the hcall number specified is not one that has an in-kernel
3581 implementation, the KVM_ENABLE_CAP ioctl will fail with an EINVAL
3584 7.2 KVM_CAP_S390_USER_SIGP
3589 This capability controls which SIGP orders will be handled completely in user
3590 space. With this capability enabled, all fast orders will be handled completely
3596 - CONDITIONAL EMERGENCY SIGNAL
3598 All other orders will be handled completely in user space.
3600 Only privileged operation exceptions will be checked for in the kernel (or even
3601 in the hardware prior to interception). If this capability is not enabled, the
3602 old way of handling SIGP orders is used (partially in kernel and user space).
3604 7.3 KVM_CAP_S390_VECTOR_REGISTERS
3608 Returns: 0 on success, negative value on error
3610 Allows use of the vector registers introduced with z13 processor, and
3611 provides for the synchronization between host and user space. Will
3612 return -EINVAL if the machine does not support vectors.
3614 7.4 KVM_CAP_S390_USER_STSI
3619 This capability allows post-handlers for the STSI instruction. After
3620 initial handling in the kernel, KVM exits to user space with
3621 KVM_EXIT_S390_STSI to allow user space to insert further data.
3623 Before exiting to userspace, kvm handlers should fill in s390_stsi field of
3634 @addr - guest address of STSI SYSIB
3638 @ar - access register number
3640 KVM handlers should exit to userspace with rc = -EREMOTE.
3642 7.5 KVM_CAP_SPLIT_IRQCHIP
3646 Returns: 0 on success, -1 on error
3648 Create a local apic for each processor in the kernel. This can be used
3649 instead of KVM_CREATE_IRQCHIP if the userspace VMM wishes to emulate the
3650 IOAPIC and PIC (and also the PIT, even though this has to be enabled
3653 This supersedes KVM_CREATE_IRQCHIP, creating only local APICs, but no in kernel
3654 IOAPIC or PIC. This also enables in kernel routing of interrupt requests.
3656 Fails if VCPU has already been created, or if the irqchip is already in the
3657 kernel (i.e. KVM_CREATE_IRQCHIP has already been called).
3660 8. Other capabilities.
3661 ----------------------
3663 This section lists capabilities that give information about other
3664 features of the KVM implementation.
3666 8.1 KVM_CAP_PPC_HWRNG
3670 This capability, if KVM_CHECK_EXTENSION indicates that it is
3671 available, means that that the kernel has an implementation of the
3672 H_RANDOM hypercall backed by a hardware random-number generator.
3673 If present, the kernel H_RANDOM handler can be enabled for guest use
3674 with the KVM_CAP_PPC_ENABLE_HCALL capability.