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
258 4.9 KVM_SET_MEMORY_ALIAS
263 Parameters: struct kvm_memory_alias (in)
264 Returns: 0 (success), -1 (error)
266 This ioctl is obsolete and has been removed.
275 Returns: 0 on success, -1 on error
277 EINTR: an unmasked signal is pending
279 This ioctl is used to run a guest virtual cpu. While there are no
280 explicit parameters, there is an implicit parameter block that can be
281 obtained by mmap()ing the vcpu fd at offset 0, with the size given by
282 KVM_GET_VCPU_MMAP_SIZE. The parameter block is formatted as a 'struct
283 kvm_run' (see below).
289 Architectures: all except ARM, arm64
291 Parameters: struct kvm_regs (out)
292 Returns: 0 on success, -1 on error
294 Reads the general purpose registers from the vcpu.
298 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
299 __u64 rax, rbx, rcx, rdx;
300 __u64 rsi, rdi, rsp, rbp;
301 __u64 r8, r9, r10, r11;
302 __u64 r12, r13, r14, r15;
308 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
319 Architectures: all except ARM, arm64
321 Parameters: struct kvm_regs (in)
322 Returns: 0 on success, -1 on error
324 Writes the general purpose registers into the vcpu.
326 See KVM_GET_REGS for the data structure.
332 Architectures: x86, ppc
334 Parameters: struct kvm_sregs (out)
335 Returns: 0 on success, -1 on error
337 Reads special registers from the vcpu.
341 struct kvm_segment cs, ds, es, fs, gs, ss;
342 struct kvm_segment tr, ldt;
343 struct kvm_dtable gdt, idt;
344 __u64 cr0, cr2, cr3, cr4, cr8;
347 __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64];
350 /* ppc -- see arch/powerpc/include/uapi/asm/kvm.h */
352 interrupt_bitmap is a bitmap of pending external interrupts. At most
353 one bit may be set. This interrupt has been acknowledged by the APIC
354 but not yet injected into the cpu core.
360 Architectures: x86, ppc
362 Parameters: struct kvm_sregs (in)
363 Returns: 0 on success, -1 on error
365 Writes special registers into the vcpu. See KVM_GET_SREGS for the
374 Parameters: struct kvm_translation (in/out)
375 Returns: 0 on success, -1 on error
377 Translates a virtual address according to the vcpu's current address
380 struct kvm_translation {
382 __u64 linear_address;
385 __u64 physical_address;
396 Architectures: x86, ppc, mips
398 Parameters: struct kvm_interrupt (in)
399 Returns: 0 on success, -1 on error
401 Queues a hardware interrupt vector to be injected. This is only
402 useful if in-kernel local APIC or equivalent is not used.
404 /* for KVM_INTERRUPT */
405 struct kvm_interrupt {
412 Note 'irq' is an interrupt vector, not an interrupt pin or line.
416 Queues an external interrupt to be injected. This ioctl is overleaded
417 with 3 different irq values:
421 This injects an edge type external interrupt into the guest once it's ready
422 to receive interrupts. When injected, the interrupt is done.
424 b) KVM_INTERRUPT_UNSET
426 This unsets any pending interrupt.
428 Only available with KVM_CAP_PPC_UNSET_IRQ.
430 c) KVM_INTERRUPT_SET_LEVEL
432 This injects a level type external interrupt into the guest context. The
433 interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET
436 Only available with KVM_CAP_PPC_IRQ_LEVEL.
438 Note that any value for 'irq' other than the ones stated above is invalid
439 and incurs unexpected behavior.
443 Queues an external interrupt to be injected into the virtual CPU. A negative
444 interrupt number dequeues the interrupt.
455 Support for this has been removed. Use KVM_SET_GUEST_DEBUG instead.
463 Parameters: struct kvm_msrs (in/out)
464 Returns: 0 on success, -1 on error
466 Reads model-specific registers from the vcpu. Supported msr indices can
467 be obtained using KVM_GET_MSR_INDEX_LIST.
470 __u32 nmsrs; /* number of msrs in entries */
473 struct kvm_msr_entry entries[0];
476 struct kvm_msr_entry {
482 Application code should set the 'nmsrs' member (which indicates the
483 size of the entries array) and the 'index' member of each array entry.
484 kvm will fill in the 'data' member.
492 Parameters: struct kvm_msrs (in)
493 Returns: 0 on success, -1 on error
495 Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the
498 Application code should set the 'nmsrs' member (which indicates the
499 size of the entries array), and the 'index' and 'data' members of each
508 Parameters: struct kvm_cpuid (in)
509 Returns: 0 on success, -1 on error
511 Defines the vcpu responses to the cpuid instruction. Applications
512 should use the KVM_SET_CPUID2 ioctl if available.
515 struct kvm_cpuid_entry {
524 /* for KVM_SET_CPUID */
528 struct kvm_cpuid_entry entries[0];
532 4.21 KVM_SET_SIGNAL_MASK
537 Parameters: struct kvm_signal_mask (in)
538 Returns: 0 on success, -1 on error
540 Defines which signals are blocked during execution of KVM_RUN. This
541 signal mask temporarily overrides the threads signal mask. Any
542 unblocked signal received (except SIGKILL and SIGSTOP, which retain
543 their traditional behaviour) will cause KVM_RUN to return with -EINTR.
545 Note the signal will only be delivered if not blocked by the original
548 /* for KVM_SET_SIGNAL_MASK */
549 struct kvm_signal_mask {
560 Parameters: struct kvm_fpu (out)
561 Returns: 0 on success, -1 on error
563 Reads the floating point state from the vcpu.
565 /* for KVM_GET_FPU and KVM_SET_FPU */
570 __u8 ftwx; /* in fxsave format */
586 Parameters: struct kvm_fpu (in)
587 Returns: 0 on success, -1 on error
589 Writes the floating point state to the vcpu.
591 /* for KVM_GET_FPU and KVM_SET_FPU */
596 __u8 ftwx; /* in fxsave format */
607 4.24 KVM_CREATE_IRQCHIP
609 Capability: KVM_CAP_IRQCHIP, KVM_CAP_S390_IRQCHIP (s390)
610 Architectures: x86, ARM, arm64, s390
613 Returns: 0 on success, -1 on error
615 Creates an interrupt controller model in the kernel.
616 On x86, creates a virtual ioapic, a virtual PIC (two PICs, nested), and sets up
617 future vcpus to have a local APIC. IRQ routing for GSIs 0-15 is set to both
618 PIC and IOAPIC; GSI 16-23 only go to the IOAPIC.
619 On ARM/arm64, a GICv2 is created. Any other GIC versions require the usage of
620 KVM_CREATE_DEVICE, which also supports creating a GICv2. Using
621 KVM_CREATE_DEVICE is preferred over KVM_CREATE_IRQCHIP for GICv2.
622 On s390, a dummy irq routing table is created.
624 Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled
625 before KVM_CREATE_IRQCHIP can be used.
630 Capability: KVM_CAP_IRQCHIP
631 Architectures: x86, arm, arm64
633 Parameters: struct kvm_irq_level
634 Returns: 0 on success, -1 on error
636 Sets the level of a GSI input to the interrupt controller model in the kernel.
637 On some architectures it is required that an interrupt controller model has
638 been previously created with KVM_CREATE_IRQCHIP. Note that edge-triggered
639 interrupts require the level to be set to 1 and then back to 0.
641 On real hardware, interrupt pins can be active-low or active-high. This
642 does not matter for the level field of struct kvm_irq_level: 1 always
643 means active (asserted), 0 means inactive (deasserted).
645 x86 allows the operating system to program the interrupt polarity
646 (active-low/active-high) for level-triggered interrupts, and KVM used
647 to consider the polarity. However, due to bitrot in the handling of
648 active-low interrupts, the above convention is now valid on x86 too.
649 This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED. Userspace
650 should not present interrupts to the guest as active-low unless this
651 capability is present (or unless it is not using the in-kernel irqchip,
655 ARM/arm64 can signal an interrupt either at the CPU level, or at the
656 in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to
657 use PPIs designated for specific cpus. The irq field is interpreted
660 Â bits: | 31 ... 24 | 23 ... 16 | 15 ... 0 |
661 field: | irq_type | vcpu_index | irq_id |
663 The irq_type field has the following values:
664 - irq_type[0]: out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ
665 - irq_type[1]: in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.)
666 (the vcpu_index field is ignored)
667 - irq_type[2]: in-kernel GIC: PPI, irq_id between 16 and 31 (incl.)
669 (The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs)
671 In both cases, level is used to assert/deassert the line.
673 struct kvm_irq_level {
676 __s32 status; /* not used for KVM_IRQ_LEVEL */
678 __u32 level; /* 0 or 1 */
684 Capability: KVM_CAP_IRQCHIP
687 Parameters: struct kvm_irqchip (in/out)
688 Returns: 0 on success, -1 on error
690 Reads the state of a kernel interrupt controller created with
691 KVM_CREATE_IRQCHIP into a buffer provided by the caller.
694 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
697 char dummy[512]; /* reserving space */
698 struct kvm_pic_state pic;
699 struct kvm_ioapic_state ioapic;
706 Capability: KVM_CAP_IRQCHIP
709 Parameters: struct kvm_irqchip (in)
710 Returns: 0 on success, -1 on error
712 Sets the state of a kernel interrupt controller created with
713 KVM_CREATE_IRQCHIP from a buffer provided by the caller.
716 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
719 char dummy[512]; /* reserving space */
720 struct kvm_pic_state pic;
721 struct kvm_ioapic_state ioapic;
726 4.28 KVM_XEN_HVM_CONFIG
728 Capability: KVM_CAP_XEN_HVM
731 Parameters: struct kvm_xen_hvm_config (in)
732 Returns: 0 on success, -1 on error
734 Sets the MSR that the Xen HVM guest uses to initialize its hypercall
735 page, and provides the starting address and size of the hypercall
736 blobs in userspace. When the guest writes the MSR, kvm copies one
737 page of a blob (32- or 64-bit, depending on the vcpu mode) to guest
740 struct kvm_xen_hvm_config {
753 Capability: KVM_CAP_ADJUST_CLOCK
756 Parameters: struct kvm_clock_data (out)
757 Returns: 0 on success, -1 on error
759 Gets the current timestamp of kvmclock as seen by the current guest. In
760 conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios
763 struct kvm_clock_data {
764 __u64 clock; /* kvmclock current value */
772 Capability: KVM_CAP_ADJUST_CLOCK
775 Parameters: struct kvm_clock_data (in)
776 Returns: 0 on success, -1 on error
778 Sets the current timestamp of kvmclock to the value specified in its parameter.
779 In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios
782 struct kvm_clock_data {
783 __u64 clock; /* kvmclock current value */
789 4.31 KVM_GET_VCPU_EVENTS
791 Capability: KVM_CAP_VCPU_EVENTS
792 Extended by: KVM_CAP_INTR_SHADOW
795 Parameters: struct kvm_vcpu_event (out)
796 Returns: 0 on success, -1 on error
798 Gets currently pending exceptions, interrupts, and NMIs as well as related
801 struct kvm_vcpu_events {
825 KVM_VCPUEVENT_VALID_SHADOW may be set in the flags field to signal that
826 interrupt.shadow contains a valid state. Otherwise, this field is undefined.
829 4.32 KVM_SET_VCPU_EVENTS
831 Capability: KVM_CAP_VCPU_EVENTS
832 Extended by: KVM_CAP_INTR_SHADOW
835 Parameters: struct kvm_vcpu_event (in)
836 Returns: 0 on success, -1 on error
838 Set pending exceptions, interrupts, and NMIs as well as related states of the
841 See KVM_GET_VCPU_EVENTS for the data structure.
843 Fields that may be modified asynchronously by running VCPUs can be excluded
844 from the update. These fields are nmi.pending and sipi_vector. Keep the
845 corresponding bits in the flags field cleared to suppress overwriting the
846 current in-kernel state. The bits are:
848 KVM_VCPUEVENT_VALID_NMI_PENDING - transfer nmi.pending to the kernel
849 KVM_VCPUEVENT_VALID_SIPI_VECTOR - transfer sipi_vector
851 If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in
852 the flags field to signal that interrupt.shadow contains a valid state and
853 shall be written into the VCPU.
856 4.33 KVM_GET_DEBUGREGS
858 Capability: KVM_CAP_DEBUGREGS
861 Parameters: struct kvm_debugregs (out)
862 Returns: 0 on success, -1 on error
864 Reads debug registers from the vcpu.
866 struct kvm_debugregs {
875 4.34 KVM_SET_DEBUGREGS
877 Capability: KVM_CAP_DEBUGREGS
880 Parameters: struct kvm_debugregs (in)
881 Returns: 0 on success, -1 on error
883 Writes debug registers into the vcpu.
885 See KVM_GET_DEBUGREGS for the data structure. The flags field is unused
886 yet and must be cleared on entry.
889 4.35 KVM_SET_USER_MEMORY_REGION
891 Capability: KVM_CAP_USER_MEM
894 Parameters: struct kvm_userspace_memory_region (in)
895 Returns: 0 on success, -1 on error
897 struct kvm_userspace_memory_region {
900 __u64 guest_phys_addr;
901 __u64 memory_size; /* bytes */
902 __u64 userspace_addr; /* start of the userspace allocated memory */
905 /* for kvm_memory_region::flags */
906 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0)
907 #define KVM_MEM_READONLY (1UL << 1)
909 This ioctl allows the user to create or modify a guest physical memory
910 slot. When changing an existing slot, it may be moved in the guest
911 physical memory space, or its flags may be modified. It may not be
912 resized. Slots may not overlap in guest physical address space.
914 Memory for the region is taken starting at the address denoted by the
915 field userspace_addr, which must point at user addressable memory for
916 the entire memory slot size. Any object may back this memory, including
917 anonymous memory, ordinary files, and hugetlbfs.
919 It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr
920 be identical. This allows large pages in the guest to be backed by large
923 The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and
924 KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of
925 writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to
926 use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it,
927 to make a new slot read-only. In this case, writes to this memory will be
928 posted to userspace as KVM_EXIT_MMIO exits.
930 When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of
931 the memory region are automatically reflected into the guest. For example, an
932 mmap() that affects the region will be made visible immediately. Another
933 example is madvise(MADV_DROP).
935 It is recommended to use this API instead of the KVM_SET_MEMORY_REGION ioctl.
936 The KVM_SET_MEMORY_REGION does not allow fine grained control over memory
937 allocation and is deprecated.
940 4.36 KVM_SET_TSS_ADDR
942 Capability: KVM_CAP_SET_TSS_ADDR
945 Parameters: unsigned long tss_address (in)
946 Returns: 0 on success, -1 on error
948 This ioctl defines the physical address of a three-page region in the guest
949 physical address space. The region must be within the first 4GB of the
950 guest physical address space and must not conflict with any memory slot
951 or any mmio address. The guest may malfunction if it accesses this memory
954 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
955 because of a quirk in the virtualization implementation (see the internals
956 documentation when it pops into existence).
961 Capability: KVM_CAP_ENABLE_CAP, KVM_CAP_ENABLE_CAP_VM
962 Architectures: x86 (only KVM_CAP_ENABLE_CAP_VM),
963 mips (only KVM_CAP_ENABLE_CAP), ppc, s390
964 Type: vcpu ioctl, vm ioctl (with KVM_CAP_ENABLE_CAP_VM)
965 Parameters: struct kvm_enable_cap (in)
966 Returns: 0 on success; -1 on error
968 +Not all extensions are enabled by default. Using this ioctl the application
969 can enable an extension, making it available to the guest.
971 On systems that do not support this ioctl, it always fails. On systems that
972 do support it, it only works for extensions that are supported for enablement.
974 To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should
977 struct kvm_enable_cap {
981 The capability that is supposed to get enabled.
985 A bitfield indicating future enhancements. Has to be 0 for now.
989 Arguments for enabling a feature. If a feature needs initial values to
990 function properly, this is the place to put them.
995 The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl
996 for vm-wide capabilities.
998 4.38 KVM_GET_MP_STATE
1000 Capability: KVM_CAP_MP_STATE
1001 Architectures: x86, s390, arm, arm64
1003 Parameters: struct kvm_mp_state (out)
1004 Returns: 0 on success; -1 on error
1006 struct kvm_mp_state {
1010 Returns the vcpu's current "multiprocessing state" (though also valid on
1011 uniprocessor guests).
1013 Possible values are:
1015 - KVM_MP_STATE_RUNNABLE: the vcpu is currently running [x86,arm/arm64]
1016 - KVM_MP_STATE_UNINITIALIZED: the vcpu is an application processor (AP)
1017 which has not yet received an INIT signal [x86]
1018 - KVM_MP_STATE_INIT_RECEIVED: the vcpu has received an INIT signal, and is
1019 now ready for a SIPI [x86]
1020 - KVM_MP_STATE_HALTED: the vcpu has executed a HLT instruction and
1021 is waiting for an interrupt [x86]
1022 - KVM_MP_STATE_SIPI_RECEIVED: the vcpu has just received a SIPI (vector
1023 accessible via KVM_GET_VCPU_EVENTS) [x86]
1024 - KVM_MP_STATE_STOPPED: the vcpu is stopped [s390,arm/arm64]
1025 - KVM_MP_STATE_CHECK_STOP: the vcpu is in a special error state [s390]
1026 - KVM_MP_STATE_OPERATING: the vcpu is operating (running or halted)
1028 - KVM_MP_STATE_LOAD: the vcpu is in a special load/startup state
1031 On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an
1032 in-kernel irqchip, the multiprocessing state must be maintained by userspace on
1033 these architectures.
1037 The only states that are valid are KVM_MP_STATE_STOPPED and
1038 KVM_MP_STATE_RUNNABLE which reflect if the vcpu is paused or not.
1040 4.39 KVM_SET_MP_STATE
1042 Capability: KVM_CAP_MP_STATE
1043 Architectures: x86, s390, arm, arm64
1045 Parameters: struct kvm_mp_state (in)
1046 Returns: 0 on success; -1 on error
1048 Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for
1051 On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an
1052 in-kernel irqchip, the multiprocessing state must be maintained by userspace on
1053 these architectures.
1057 The only states that are valid are KVM_MP_STATE_STOPPED and
1058 KVM_MP_STATE_RUNNABLE which reflect if the vcpu should be paused or not.
1060 4.40 KVM_SET_IDENTITY_MAP_ADDR
1062 Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR
1065 Parameters: unsigned long identity (in)
1066 Returns: 0 on success, -1 on error
1068 This ioctl defines the physical address of a one-page region in the guest
1069 physical address space. The region must be within the first 4GB of the
1070 guest physical address space and must not conflict with any memory slot
1071 or any mmio address. The guest may malfunction if it accesses this memory
1074 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
1075 because of a quirk in the virtualization implementation (see the internals
1076 documentation when it pops into existence).
1079 4.41 KVM_SET_BOOT_CPU_ID
1081 Capability: KVM_CAP_SET_BOOT_CPU_ID
1084 Parameters: unsigned long vcpu_id
1085 Returns: 0 on success, -1 on error
1087 Define which vcpu is the Bootstrap Processor (BSP). Values are the same
1088 as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default
1094 Capability: KVM_CAP_XSAVE
1097 Parameters: struct kvm_xsave (out)
1098 Returns: 0 on success, -1 on error
1104 This ioctl would copy current vcpu's xsave struct to the userspace.
1109 Capability: KVM_CAP_XSAVE
1112 Parameters: struct kvm_xsave (in)
1113 Returns: 0 on success, -1 on error
1119 This ioctl would copy userspace's xsave struct to the kernel.
1124 Capability: KVM_CAP_XCRS
1127 Parameters: struct kvm_xcrs (out)
1128 Returns: 0 on success, -1 on error
1139 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1143 This ioctl would copy current vcpu's xcrs to the userspace.
1148 Capability: KVM_CAP_XCRS
1151 Parameters: struct kvm_xcrs (in)
1152 Returns: 0 on success, -1 on error
1163 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1167 This ioctl would set vcpu's xcr to the value userspace specified.
1170 4.46 KVM_GET_SUPPORTED_CPUID
1172 Capability: KVM_CAP_EXT_CPUID
1175 Parameters: struct kvm_cpuid2 (in/out)
1176 Returns: 0 on success, -1 on error
1181 struct kvm_cpuid_entry2 entries[0];
1184 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
1185 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
1186 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
1188 struct kvm_cpuid_entry2 {
1199 This ioctl returns x86 cpuid features which are supported by both the hardware
1200 and kvm. Userspace can use the information returned by this ioctl to
1201 construct cpuid information (for KVM_SET_CPUID2) that is consistent with
1202 hardware, kernel, and userspace capabilities, and with user requirements (for
1203 example, the user may wish to constrain cpuid to emulate older hardware,
1204 or for feature consistency across a cluster).
1206 Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure
1207 with the 'nent' field indicating the number of entries in the variable-size
1208 array 'entries'. If the number of entries is too low to describe the cpu
1209 capabilities, an error (E2BIG) is returned. If the number is too high,
1210 the 'nent' field is adjusted and an error (ENOMEM) is returned. If the
1211 number is just right, the 'nent' field is adjusted to the number of valid
1212 entries in the 'entries' array, which is then filled.
1214 The entries returned are the host cpuid as returned by the cpuid instruction,
1215 with unknown or unsupported features masked out. Some features (for example,
1216 x2apic), may not be present in the host cpu, but are exposed by kvm if it can
1217 emulate them efficiently. The fields in each entry are defined as follows:
1219 function: the eax value used to obtain the entry
1220 index: the ecx value used to obtain the entry (for entries that are
1222 flags: an OR of zero or more of the following:
1223 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
1224 if the index field is valid
1225 KVM_CPUID_FLAG_STATEFUL_FUNC:
1226 if cpuid for this function returns different values for successive
1227 invocations; there will be several entries with the same function,
1228 all with this flag set
1229 KVM_CPUID_FLAG_STATE_READ_NEXT:
1230 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
1231 the first entry to be read by a cpu
1232 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
1233 this function/index combination
1235 The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned
1236 as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC
1237 support. Instead it is reported via
1239 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER)
1241 if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the
1242 feature in userspace, then you can enable the feature for KVM_SET_CPUID2.
1245 4.47 KVM_PPC_GET_PVINFO
1247 Capability: KVM_CAP_PPC_GET_PVINFO
1250 Parameters: struct kvm_ppc_pvinfo (out)
1251 Returns: 0 on success, !0 on error
1253 struct kvm_ppc_pvinfo {
1259 This ioctl fetches PV specific information that need to be passed to the guest
1260 using the device tree or other means from vm context.
1262 The hcall array defines 4 instructions that make up a hypercall.
1264 If any additional field gets added to this structure later on, a bit for that
1265 additional piece of information will be set in the flags bitmap.
1267 The flags bitmap is defined as:
1269 /* the host supports the ePAPR idle hcall
1270 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0)
1272 4.48 KVM_ASSIGN_PCI_DEVICE
1277 Parameters: struct kvm_assigned_pci_dev (in)
1278 Returns: 0 on success, -1 on error
1280 Assigns a host PCI device to the VM.
1282 struct kvm_assigned_pci_dev {
1283 __u32 assigned_dev_id;
1293 The PCI device is specified by the triple segnr, busnr, and devfn.
1294 Identification in succeeding service requests is done via assigned_dev_id. The
1295 following flags are specified:
1297 /* Depends on KVM_CAP_IOMMU */
1298 #define KVM_DEV_ASSIGN_ENABLE_IOMMU (1 << 0)
1299 /* The following two depend on KVM_CAP_PCI_2_3 */
1300 #define KVM_DEV_ASSIGN_PCI_2_3 (1 << 1)
1301 #define KVM_DEV_ASSIGN_MASK_INTX (1 << 2)
1303 If KVM_DEV_ASSIGN_PCI_2_3 is set, the kernel will manage legacy INTx interrupts
1304 via the PCI-2.3-compliant device-level mask, thus enable IRQ sharing with other
1305 assigned devices or host devices. KVM_DEV_ASSIGN_MASK_INTX specifies the
1306 guest's view on the INTx mask, see KVM_ASSIGN_SET_INTX_MASK for details.
1308 The KVM_DEV_ASSIGN_ENABLE_IOMMU flag is a mandatory option to ensure
1309 isolation of the device. Usages not specifying this flag are deprecated.
1311 Only PCI header type 0 devices with PCI BAR resources are supported by
1312 device assignment. The user requesting this ioctl must have read/write
1313 access to the PCI sysfs resource files associated with the device.
1316 ENOTTY: kernel does not support this ioctl
1318 Other error conditions may be defined by individual device types or
1319 have their standard meanings.
1322 4.49 KVM_DEASSIGN_PCI_DEVICE
1327 Parameters: struct kvm_assigned_pci_dev (in)
1328 Returns: 0 on success, -1 on error
1330 Ends PCI device assignment, releasing all associated resources.
1332 See KVM_ASSIGN_PCI_DEVICE for the data structure. Only assigned_dev_id is
1333 used in kvm_assigned_pci_dev to identify the device.
1336 ENOTTY: kernel does not support this ioctl
1338 Other error conditions may be defined by individual device types or
1339 have their standard meanings.
1341 4.50 KVM_ASSIGN_DEV_IRQ
1343 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1346 Parameters: struct kvm_assigned_irq (in)
1347 Returns: 0 on success, -1 on error
1349 Assigns an IRQ to a passed-through device.
1351 struct kvm_assigned_irq {
1352 __u32 assigned_dev_id;
1353 __u32 host_irq; /* ignored (legacy field) */
1361 The following flags are defined:
1363 #define KVM_DEV_IRQ_HOST_INTX (1 << 0)
1364 #define KVM_DEV_IRQ_HOST_MSI (1 << 1)
1365 #define KVM_DEV_IRQ_HOST_MSIX (1 << 2)
1367 #define KVM_DEV_IRQ_GUEST_INTX (1 << 8)
1368 #define KVM_DEV_IRQ_GUEST_MSI (1 << 9)
1369 #define KVM_DEV_IRQ_GUEST_MSIX (1 << 10)
1371 It is not valid to specify multiple types per host or guest IRQ. However, the
1372 IRQ type of host and guest can differ or can even be null.
1375 ENOTTY: kernel does not support this ioctl
1377 Other error conditions may be defined by individual device types or
1378 have their standard meanings.
1381 4.51 KVM_DEASSIGN_DEV_IRQ
1383 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1386 Parameters: struct kvm_assigned_irq (in)
1387 Returns: 0 on success, -1 on error
1389 Ends an IRQ assignment to a passed-through device.
1391 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1392 by assigned_dev_id, flags must correspond to the IRQ type specified on
1393 KVM_ASSIGN_DEV_IRQ. Partial deassignment of host or guest IRQ is allowed.
1396 4.52 KVM_SET_GSI_ROUTING
1398 Capability: KVM_CAP_IRQ_ROUTING
1399 Architectures: x86 s390
1401 Parameters: struct kvm_irq_routing (in)
1402 Returns: 0 on success, -1 on error
1404 Sets the GSI routing table entries, overwriting any previously set entries.
1406 struct kvm_irq_routing {
1409 struct kvm_irq_routing_entry entries[0];
1412 No flags are specified so far, the corresponding field must be set to zero.
1414 struct kvm_irq_routing_entry {
1420 struct kvm_irq_routing_irqchip irqchip;
1421 struct kvm_irq_routing_msi msi;
1422 struct kvm_irq_routing_s390_adapter adapter;
1427 /* gsi routing entry types */
1428 #define KVM_IRQ_ROUTING_IRQCHIP 1
1429 #define KVM_IRQ_ROUTING_MSI 2
1430 #define KVM_IRQ_ROUTING_S390_ADAPTER 3
1432 No flags are specified so far, the corresponding field must be set to zero.
1434 struct kvm_irq_routing_irqchip {
1439 struct kvm_irq_routing_msi {
1446 struct kvm_irq_routing_s390_adapter {
1450 __u32 summary_offset;
1455 4.53 KVM_ASSIGN_SET_MSIX_NR
1460 Parameters: struct kvm_assigned_msix_nr (in)
1461 Returns: 0 on success, -1 on error
1463 Set the number of MSI-X interrupts for an assigned device. The number is
1464 reset again by terminating the MSI-X assignment of the device via
1465 KVM_DEASSIGN_DEV_IRQ. Calling this service more than once at any earlier
1468 struct kvm_assigned_msix_nr {
1469 __u32 assigned_dev_id;
1474 #define KVM_MAX_MSIX_PER_DEV 256
1477 4.54 KVM_ASSIGN_SET_MSIX_ENTRY
1482 Parameters: struct kvm_assigned_msix_entry (in)
1483 Returns: 0 on success, -1 on error
1485 Specifies the routing of an MSI-X assigned device interrupt to a GSI. Setting
1486 the GSI vector to zero means disabling the interrupt.
1488 struct kvm_assigned_msix_entry {
1489 __u32 assigned_dev_id;
1491 __u16 entry; /* The index of entry in the MSI-X table */
1496 ENOTTY: kernel does not support this ioctl
1498 Other error conditions may be defined by individual device types or
1499 have their standard meanings.
1502 4.55 KVM_SET_TSC_KHZ
1504 Capability: KVM_CAP_TSC_CONTROL
1507 Parameters: virtual tsc_khz
1508 Returns: 0 on success, -1 on error
1510 Specifies the tsc frequency for the virtual machine. The unit of the
1514 4.56 KVM_GET_TSC_KHZ
1516 Capability: KVM_CAP_GET_TSC_KHZ
1520 Returns: virtual tsc-khz on success, negative value on error
1522 Returns the tsc frequency of the guest. The unit of the return value is
1523 KHz. If the host has unstable tsc this ioctl returns -EIO instead as an
1529 Capability: KVM_CAP_IRQCHIP
1532 Parameters: struct kvm_lapic_state (out)
1533 Returns: 0 on success, -1 on error
1535 #define KVM_APIC_REG_SIZE 0x400
1536 struct kvm_lapic_state {
1537 char regs[KVM_APIC_REG_SIZE];
1540 Reads the Local APIC registers and copies them into the input argument. The
1541 data format and layout are the same as documented in the architecture manual.
1546 Capability: KVM_CAP_IRQCHIP
1549 Parameters: struct kvm_lapic_state (in)
1550 Returns: 0 on success, -1 on error
1552 #define KVM_APIC_REG_SIZE 0x400
1553 struct kvm_lapic_state {
1554 char regs[KVM_APIC_REG_SIZE];
1557 Copies the input argument into the Local APIC registers. The data format
1558 and layout are the same as documented in the architecture manual.
1563 Capability: KVM_CAP_IOEVENTFD
1566 Parameters: struct kvm_ioeventfd (in)
1567 Returns: 0 on success, !0 on error
1569 This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address
1570 within the guest. A guest write in the registered address will signal the
1571 provided event instead of triggering an exit.
1573 struct kvm_ioeventfd {
1575 __u64 addr; /* legal pio/mmio address */
1576 __u32 len; /* 1, 2, 4, or 8 bytes */
1582 For the special case of virtio-ccw devices on s390, the ioevent is matched
1583 to a subchannel/virtqueue tuple instead.
1585 The following flags are defined:
1587 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch)
1588 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio)
1589 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign)
1590 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \
1591 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify)
1593 If datamatch flag is set, the event will be signaled only if the written value
1594 to the registered address is equal to datamatch in struct kvm_ioeventfd.
1596 For virtio-ccw devices, addr contains the subchannel id and datamatch the
1602 Capability: KVM_CAP_SW_TLB
1605 Parameters: struct kvm_dirty_tlb (in)
1606 Returns: 0 on success, -1 on error
1608 struct kvm_dirty_tlb {
1613 This must be called whenever userspace has changed an entry in the shared
1614 TLB, prior to calling KVM_RUN on the associated vcpu.
1616 The "bitmap" field is the userspace address of an array. This array
1617 consists of a number of bits, equal to the total number of TLB entries as
1618 determined by the last successful call to KVM_CONFIG_TLB, rounded up to the
1619 nearest multiple of 64.
1621 Each bit corresponds to one TLB entry, ordered the same as in the shared TLB
1624 The array is little-endian: the bit 0 is the least significant bit of the
1625 first byte, bit 8 is the least significant bit of the second byte, etc.
1626 This avoids any complications with differing word sizes.
1628 The "num_dirty" field is a performance hint for KVM to determine whether it
1629 should skip processing the bitmap and just invalidate everything. It must
1630 be set to the number of set bits in the bitmap.
1633 4.61 KVM_ASSIGN_SET_INTX_MASK
1635 Capability: KVM_CAP_PCI_2_3
1638 Parameters: struct kvm_assigned_pci_dev (in)
1639 Returns: 0 on success, -1 on error
1641 Allows userspace to mask PCI INTx interrupts from the assigned device. The
1642 kernel will not deliver INTx interrupts to the guest between setting and
1643 clearing of KVM_ASSIGN_SET_INTX_MASK via this interface. This enables use of
1644 and emulation of PCI 2.3 INTx disable command register behavior.
1646 This may be used for both PCI 2.3 devices supporting INTx disable natively and
1647 older devices lacking this support. Userspace is responsible for emulating the
1648 read value of the INTx disable bit in the guest visible PCI command register.
1649 When modifying the INTx disable state, userspace should precede updating the
1650 physical device command register by calling this ioctl to inform the kernel of
1651 the new intended INTx mask state.
1653 Note that the kernel uses the device INTx disable bit to internally manage the
1654 device interrupt state for PCI 2.3 devices. Reads of this register may
1655 therefore not match the expected value. Writes should always use the guest
1656 intended INTx disable value rather than attempting to read-copy-update the
1657 current physical device state. Races between user and kernel updates to the
1658 INTx disable bit are handled lazily in the kernel. It's possible the device
1659 may generate unintended interrupts, but they will not be injected into the
1662 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1663 by assigned_dev_id. In the flags field, only KVM_DEV_ASSIGN_MASK_INTX is
1667 4.62 KVM_CREATE_SPAPR_TCE
1669 Capability: KVM_CAP_SPAPR_TCE
1670 Architectures: powerpc
1672 Parameters: struct kvm_create_spapr_tce (in)
1673 Returns: file descriptor for manipulating the created TCE table
1675 This creates a virtual TCE (translation control entry) table, which
1676 is an IOMMU for PAPR-style virtual I/O. It is used to translate
1677 logical addresses used in virtual I/O into guest physical addresses,
1678 and provides a scatter/gather capability for PAPR virtual I/O.
1680 /* for KVM_CAP_SPAPR_TCE */
1681 struct kvm_create_spapr_tce {
1686 The liobn field gives the logical IO bus number for which to create a
1687 TCE table. The window_size field specifies the size of the DMA window
1688 which this TCE table will translate - the table will contain one 64
1689 bit TCE entry for every 4kiB of the DMA window.
1691 When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE
1692 table has been created using this ioctl(), the kernel will handle it
1693 in real mode, updating the TCE table. H_PUT_TCE calls for other
1694 liobns will cause a vm exit and must be handled by userspace.
1696 The return value is a file descriptor which can be passed to mmap(2)
1697 to map the created TCE table into userspace. This lets userspace read
1698 the entries written by kernel-handled H_PUT_TCE calls, and also lets
1699 userspace update the TCE table directly which is useful in some
1703 4.63 KVM_ALLOCATE_RMA
1705 Capability: KVM_CAP_PPC_RMA
1706 Architectures: powerpc
1708 Parameters: struct kvm_allocate_rma (out)
1709 Returns: file descriptor for mapping the allocated RMA
1711 This allocates a Real Mode Area (RMA) from the pool allocated at boot
1712 time by the kernel. An RMA is a physically-contiguous, aligned region
1713 of memory used on older POWER processors to provide the memory which
1714 will be accessed by real-mode (MMU off) accesses in a KVM guest.
1715 POWER processors support a set of sizes for the RMA that usually
1716 includes 64MB, 128MB, 256MB and some larger powers of two.
1718 /* for KVM_ALLOCATE_RMA */
1719 struct kvm_allocate_rma {
1723 The return value is a file descriptor which can be passed to mmap(2)
1724 to map the allocated RMA into userspace. The mapped area can then be
1725 passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the
1726 RMA for a virtual machine. The size of the RMA in bytes (which is
1727 fixed at host kernel boot time) is returned in the rma_size field of
1728 the argument structure.
1730 The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl
1731 is supported; 2 if the processor requires all virtual machines to have
1732 an RMA, or 1 if the processor can use an RMA but doesn't require it,
1733 because it supports the Virtual RMA (VRMA) facility.
1738 Capability: KVM_CAP_USER_NMI
1742 Returns: 0 on success, -1 on error
1744 Queues an NMI on the thread's vcpu. Note this is well defined only
1745 when KVM_CREATE_IRQCHIP has not been called, since this is an interface
1746 between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP
1747 has been called, this interface is completely emulated within the kernel.
1749 To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the
1750 following algorithm:
1753 - read the local APIC's state (KVM_GET_LAPIC)
1754 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1)
1755 - if so, issue KVM_NMI
1758 Some guests configure the LINT1 NMI input to cause a panic, aiding in
1762 4.65 KVM_S390_UCAS_MAP
1764 Capability: KVM_CAP_S390_UCONTROL
1767 Parameters: struct kvm_s390_ucas_mapping (in)
1768 Returns: 0 in case of success
1770 The parameter is defined like this:
1771 struct kvm_s390_ucas_mapping {
1777 This ioctl maps the memory at "user_addr" with the length "length" to
1778 the vcpu's address space starting at "vcpu_addr". All parameters need to
1779 be aligned by 1 megabyte.
1782 4.66 KVM_S390_UCAS_UNMAP
1784 Capability: KVM_CAP_S390_UCONTROL
1787 Parameters: struct kvm_s390_ucas_mapping (in)
1788 Returns: 0 in case of success
1790 The parameter is defined like this:
1791 struct kvm_s390_ucas_mapping {
1797 This ioctl unmaps the memory in the vcpu's address space starting at
1798 "vcpu_addr" with the length "length". The field "user_addr" is ignored.
1799 All parameters need to be aligned by 1 megabyte.
1802 4.67 KVM_S390_VCPU_FAULT
1804 Capability: KVM_CAP_S390_UCONTROL
1807 Parameters: vcpu absolute address (in)
1808 Returns: 0 in case of success
1810 This call creates a page table entry on the virtual cpu's address space
1811 (for user controlled virtual machines) or the virtual machine's address
1812 space (for regular virtual machines). This only works for minor faults,
1813 thus it's recommended to access subject memory page via the user page
1814 table upfront. This is useful to handle validity intercepts for user
1815 controlled virtual machines to fault in the virtual cpu's lowcore pages
1816 prior to calling the KVM_RUN ioctl.
1819 4.68 KVM_SET_ONE_REG
1821 Capability: KVM_CAP_ONE_REG
1824 Parameters: struct kvm_one_reg (in)
1825 Returns: 0 on success, negative value on failure
1827 struct kvm_one_reg {
1832 Using this ioctl, a single vcpu register can be set to a specific value
1833 defined by user space with the passed in struct kvm_one_reg, where id
1834 refers to the register identifier as described below and addr is a pointer
1835 to a variable with the respective size. There can be architecture agnostic
1836 and architecture specific registers. Each have their own range of operation
1837 and their own constants and width. To keep track of the implemented
1838 registers, find a list below:
1840 Arch | Register | Width (bits)
1842 PPC | KVM_REG_PPC_HIOR | 64
1843 PPC | KVM_REG_PPC_IAC1 | 64
1844 PPC | KVM_REG_PPC_IAC2 | 64
1845 PPC | KVM_REG_PPC_IAC3 | 64
1846 PPC | KVM_REG_PPC_IAC4 | 64
1847 PPC | KVM_REG_PPC_DAC1 | 64
1848 PPC | KVM_REG_PPC_DAC2 | 64
1849 PPC | KVM_REG_PPC_DABR | 64
1850 PPC | KVM_REG_PPC_DSCR | 64
1851 PPC | KVM_REG_PPC_PURR | 64
1852 PPC | KVM_REG_PPC_SPURR | 64
1853 PPC | KVM_REG_PPC_DAR | 64
1854 PPC | KVM_REG_PPC_DSISR | 32
1855 PPC | KVM_REG_PPC_AMR | 64
1856 PPC | KVM_REG_PPC_UAMOR | 64
1857 PPC | KVM_REG_PPC_MMCR0 | 64
1858 PPC | KVM_REG_PPC_MMCR1 | 64
1859 PPC | KVM_REG_PPC_MMCRA | 64
1860 PPC | KVM_REG_PPC_MMCR2 | 64
1861 PPC | KVM_REG_PPC_MMCRS | 64
1862 PPC | KVM_REG_PPC_SIAR | 64
1863 PPC | KVM_REG_PPC_SDAR | 64
1864 PPC | KVM_REG_PPC_SIER | 64
1865 PPC | KVM_REG_PPC_PMC1 | 32
1866 PPC | KVM_REG_PPC_PMC2 | 32
1867 PPC | KVM_REG_PPC_PMC3 | 32
1868 PPC | KVM_REG_PPC_PMC4 | 32
1869 PPC | KVM_REG_PPC_PMC5 | 32
1870 PPC | KVM_REG_PPC_PMC6 | 32
1871 PPC | KVM_REG_PPC_PMC7 | 32
1872 PPC | KVM_REG_PPC_PMC8 | 32
1873 PPC | KVM_REG_PPC_FPR0 | 64
1875 PPC | KVM_REG_PPC_FPR31 | 64
1876 PPC | KVM_REG_PPC_VR0 | 128
1878 PPC | KVM_REG_PPC_VR31 | 128
1879 PPC | KVM_REG_PPC_VSR0 | 128
1881 PPC | KVM_REG_PPC_VSR31 | 128
1882 PPC | KVM_REG_PPC_FPSCR | 64
1883 PPC | KVM_REG_PPC_VSCR | 32
1884 PPC | KVM_REG_PPC_VPA_ADDR | 64
1885 PPC | KVM_REG_PPC_VPA_SLB | 128
1886 PPC | KVM_REG_PPC_VPA_DTL | 128
1887 PPC | KVM_REG_PPC_EPCR | 32
1888 PPC | KVM_REG_PPC_EPR | 32
1889 PPC | KVM_REG_PPC_TCR | 32
1890 PPC | KVM_REG_PPC_TSR | 32
1891 PPC | KVM_REG_PPC_OR_TSR | 32
1892 PPC | KVM_REG_PPC_CLEAR_TSR | 32
1893 PPC | KVM_REG_PPC_MAS0 | 32
1894 PPC | KVM_REG_PPC_MAS1 | 32
1895 PPC | KVM_REG_PPC_MAS2 | 64
1896 PPC | KVM_REG_PPC_MAS7_3 | 64
1897 PPC | KVM_REG_PPC_MAS4 | 32
1898 PPC | KVM_REG_PPC_MAS6 | 32
1899 PPC | KVM_REG_PPC_MMUCFG | 32
1900 PPC | KVM_REG_PPC_TLB0CFG | 32
1901 PPC | KVM_REG_PPC_TLB1CFG | 32
1902 PPC | KVM_REG_PPC_TLB2CFG | 32
1903 PPC | KVM_REG_PPC_TLB3CFG | 32
1904 PPC | KVM_REG_PPC_TLB0PS | 32
1905 PPC | KVM_REG_PPC_TLB1PS | 32
1906 PPC | KVM_REG_PPC_TLB2PS | 32
1907 PPC | KVM_REG_PPC_TLB3PS | 32
1908 PPC | KVM_REG_PPC_EPTCFG | 32
1909 PPC | KVM_REG_PPC_ICP_STATE | 64
1910 PPC | KVM_REG_PPC_TB_OFFSET | 64
1911 PPC | KVM_REG_PPC_SPMC1 | 32
1912 PPC | KVM_REG_PPC_SPMC2 | 32
1913 PPC | KVM_REG_PPC_IAMR | 64
1914 PPC | KVM_REG_PPC_TFHAR | 64
1915 PPC | KVM_REG_PPC_TFIAR | 64
1916 PPC | KVM_REG_PPC_TEXASR | 64
1917 PPC | KVM_REG_PPC_FSCR | 64
1918 PPC | KVM_REG_PPC_PSPB | 32
1919 PPC | KVM_REG_PPC_EBBHR | 64
1920 PPC | KVM_REG_PPC_EBBRR | 64
1921 PPC | KVM_REG_PPC_BESCR | 64
1922 PPC | KVM_REG_PPC_TAR | 64
1923 PPC | KVM_REG_PPC_DPDES | 64
1924 PPC | KVM_REG_PPC_DAWR | 64
1925 PPC | KVM_REG_PPC_DAWRX | 64
1926 PPC | KVM_REG_PPC_CIABR | 64
1927 PPC | KVM_REG_PPC_IC | 64
1928 PPC | KVM_REG_PPC_VTB | 64
1929 PPC | KVM_REG_PPC_CSIGR | 64
1930 PPC | KVM_REG_PPC_TACR | 64
1931 PPC | KVM_REG_PPC_TCSCR | 64
1932 PPC | KVM_REG_PPC_PID | 64
1933 PPC | KVM_REG_PPC_ACOP | 64
1934 PPC | KVM_REG_PPC_VRSAVE | 32
1935 PPC | KVM_REG_PPC_LPCR | 32
1936 PPC | KVM_REG_PPC_LPCR_64 | 64
1937 PPC | KVM_REG_PPC_PPR | 64
1938 PPC | KVM_REG_PPC_ARCH_COMPAT | 32
1939 PPC | KVM_REG_PPC_DABRX | 32
1940 PPC | KVM_REG_PPC_WORT | 64
1941 PPC | KVM_REG_PPC_SPRG9 | 64
1942 PPC | KVM_REG_PPC_DBSR | 32
1943 PPC | KVM_REG_PPC_TM_GPR0 | 64
1945 PPC | KVM_REG_PPC_TM_GPR31 | 64
1946 PPC | KVM_REG_PPC_TM_VSR0 | 128
1948 PPC | KVM_REG_PPC_TM_VSR63 | 128
1949 PPC | KVM_REG_PPC_TM_CR | 64
1950 PPC | KVM_REG_PPC_TM_LR | 64
1951 PPC | KVM_REG_PPC_TM_CTR | 64
1952 PPC | KVM_REG_PPC_TM_FPSCR | 64
1953 PPC | KVM_REG_PPC_TM_AMR | 64
1954 PPC | KVM_REG_PPC_TM_PPR | 64
1955 PPC | KVM_REG_PPC_TM_VRSAVE | 64
1956 PPC | KVM_REG_PPC_TM_VSCR | 32
1957 PPC | KVM_REG_PPC_TM_DSCR | 64
1958 PPC | KVM_REG_PPC_TM_TAR | 64
1960 MIPS | KVM_REG_MIPS_R0 | 64
1962 MIPS | KVM_REG_MIPS_R31 | 64
1963 MIPS | KVM_REG_MIPS_HI | 64
1964 MIPS | KVM_REG_MIPS_LO | 64
1965 MIPS | KVM_REG_MIPS_PC | 64
1966 MIPS | KVM_REG_MIPS_CP0_INDEX | 32
1967 MIPS | KVM_REG_MIPS_CP0_CONTEXT | 64
1968 MIPS | KVM_REG_MIPS_CP0_USERLOCAL | 64
1969 MIPS | KVM_REG_MIPS_CP0_PAGEMASK | 32
1970 MIPS | KVM_REG_MIPS_CP0_WIRED | 32
1971 MIPS | KVM_REG_MIPS_CP0_HWRENA | 32
1972 MIPS | KVM_REG_MIPS_CP0_BADVADDR | 64
1973 MIPS | KVM_REG_MIPS_CP0_COUNT | 32
1974 MIPS | KVM_REG_MIPS_CP0_ENTRYHI | 64
1975 MIPS | KVM_REG_MIPS_CP0_COMPARE | 32
1976 MIPS | KVM_REG_MIPS_CP0_STATUS | 32
1977 MIPS | KVM_REG_MIPS_CP0_CAUSE | 32
1978 MIPS | KVM_REG_MIPS_CP0_EPC | 64
1979 MIPS | KVM_REG_MIPS_CP0_PRID | 32
1980 MIPS | KVM_REG_MIPS_CP0_CONFIG | 32
1981 MIPS | KVM_REG_MIPS_CP0_CONFIG1 | 32
1982 MIPS | KVM_REG_MIPS_CP0_CONFIG2 | 32
1983 MIPS | KVM_REG_MIPS_CP0_CONFIG3 | 32
1984 MIPS | KVM_REG_MIPS_CP0_CONFIG4 | 32
1985 MIPS | KVM_REG_MIPS_CP0_CONFIG5 | 32
1986 MIPS | KVM_REG_MIPS_CP0_CONFIG7 | 32
1987 MIPS | KVM_REG_MIPS_CP0_ERROREPC | 64
1988 MIPS | KVM_REG_MIPS_COUNT_CTL | 64
1989 MIPS | KVM_REG_MIPS_COUNT_RESUME | 64
1990 MIPS | KVM_REG_MIPS_COUNT_HZ | 64
1991 MIPS | KVM_REG_MIPS_FPR_32(0..31) | 32
1992 MIPS | KVM_REG_MIPS_FPR_64(0..31) | 64
1993 MIPS | KVM_REG_MIPS_VEC_128(0..31) | 128
1994 MIPS | KVM_REG_MIPS_FCR_IR | 32
1995 MIPS | KVM_REG_MIPS_FCR_CSR | 32
1996 MIPS | KVM_REG_MIPS_MSA_IR | 32
1997 MIPS | KVM_REG_MIPS_MSA_CSR | 32
1999 ARM registers are mapped using the lower 32 bits. The upper 16 of that
2000 is the register group type, or coprocessor number:
2002 ARM core registers have the following id bit patterns:
2003 0x4020 0000 0010 <index into the kvm_regs struct:16>
2005 ARM 32-bit CP15 registers have the following id bit patterns:
2006 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3>
2008 ARM 64-bit CP15 registers have the following id bit patterns:
2009 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3>
2011 ARM CCSIDR registers are demultiplexed by CSSELR value:
2012 0x4020 0000 0011 00 <csselr:8>
2014 ARM 32-bit VFP control registers have the following id bit patterns:
2015 0x4020 0000 0012 1 <regno:12>
2017 ARM 64-bit FP registers have the following id bit patterns:
2018 0x4030 0000 0012 0 <regno:12>
2021 arm64 registers are mapped using the lower 32 bits. The upper 16 of
2022 that is the register group type, or coprocessor number:
2024 arm64 core/FP-SIMD registers have the following id bit patterns. Note
2025 that the size of the access is variable, as the kvm_regs structure
2026 contains elements ranging from 32 to 128 bits. The index is a 32bit
2027 value in the kvm_regs structure seen as a 32bit array.
2028 0x60x0 0000 0010 <index into the kvm_regs struct:16>
2030 arm64 CCSIDR registers are demultiplexed by CSSELR value:
2031 0x6020 0000 0011 00 <csselr:8>
2033 arm64 system registers have the following id bit patterns:
2034 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3>
2037 MIPS registers are mapped using the lower 32 bits. The upper 16 of that is
2038 the register group type:
2040 MIPS core registers (see above) have the following id bit patterns:
2041 0x7030 0000 0000 <reg:16>
2043 MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit
2044 patterns depending on whether they're 32-bit or 64-bit registers:
2045 0x7020 0000 0001 00 <reg:5> <sel:3> (32-bit)
2046 0x7030 0000 0001 00 <reg:5> <sel:3> (64-bit)
2048 MIPS KVM control registers (see above) have the following id bit patterns:
2049 0x7030 0000 0002 <reg:16>
2051 MIPS FPU registers (see KVM_REG_MIPS_FPR_{32,64}() above) have the following
2052 id bit patterns depending on the size of the register being accessed. They are
2053 always accessed according to the current guest FPU mode (Status.FR and
2054 Config5.FRE), i.e. as the guest would see them, and they become unpredictable
2055 if the guest FPU mode is changed. MIPS SIMD Architecture (MSA) vector
2056 registers (see KVM_REG_MIPS_VEC_128() above) have similar patterns as they
2057 overlap the FPU registers:
2058 0x7020 0000 0003 00 <0:3> <reg:5> (32-bit FPU registers)
2059 0x7030 0000 0003 00 <0:3> <reg:5> (64-bit FPU registers)
2060 0x7040 0000 0003 00 <0:3> <reg:5> (128-bit MSA vector registers)
2062 MIPS FPU control registers (see KVM_REG_MIPS_FCR_{IR,CSR} above) have the
2063 following id bit patterns:
2064 0x7020 0000 0003 01 <0:3> <reg:5>
2066 MIPS MSA control registers (see KVM_REG_MIPS_MSA_{IR,CSR} above) have the
2067 following id bit patterns:
2068 0x7020 0000 0003 02 <0:3> <reg:5>
2071 4.69 KVM_GET_ONE_REG
2073 Capability: KVM_CAP_ONE_REG
2076 Parameters: struct kvm_one_reg (in and out)
2077 Returns: 0 on success, negative value on failure
2079 This ioctl allows to receive the value of a single register implemented
2080 in a vcpu. The register to read is indicated by the "id" field of the
2081 kvm_one_reg struct passed in. On success, the register value can be found
2082 at the memory location pointed to by "addr".
2084 The list of registers accessible using this interface is identical to the
2088 4.70 KVM_KVMCLOCK_CTRL
2090 Capability: KVM_CAP_KVMCLOCK_CTRL
2091 Architectures: Any that implement pvclocks (currently x86 only)
2094 Returns: 0 on success, -1 on error
2096 This signals to the host kernel that the specified guest is being paused by
2097 userspace. The host will set a flag in the pvclock structure that is checked
2098 from the soft lockup watchdog. The flag is part of the pvclock structure that
2099 is shared between guest and host, specifically the second bit of the flags
2100 field of the pvclock_vcpu_time_info structure. It will be set exclusively by
2101 the host and read/cleared exclusively by the guest. The guest operation of
2102 checking and clearing the flag must an atomic operation so
2103 load-link/store-conditional, or equivalent must be used. There are two cases
2104 where the guest will clear the flag: when the soft lockup watchdog timer resets
2105 itself or when a soft lockup is detected. This ioctl can be called any time
2106 after pausing the vcpu, but before it is resumed.
2111 Capability: KVM_CAP_SIGNAL_MSI
2114 Parameters: struct kvm_msi (in)
2115 Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error
2117 Directly inject a MSI message. Only valid with in-kernel irqchip that handles
2128 No flags are defined so far. The corresponding field must be 0.
2131 4.71 KVM_CREATE_PIT2
2133 Capability: KVM_CAP_PIT2
2136 Parameters: struct kvm_pit_config (in)
2137 Returns: 0 on success, -1 on error
2139 Creates an in-kernel device model for the i8254 PIT. This call is only valid
2140 after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following
2141 parameters have to be passed:
2143 struct kvm_pit_config {
2150 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */
2152 PIT timer interrupts may use a per-VM kernel thread for injection. If it
2153 exists, this thread will have a name of the following pattern:
2155 kvm-pit/<owner-process-pid>
2157 When running a guest with elevated priorities, the scheduling parameters of
2158 this thread may have to be adjusted accordingly.
2160 This IOCTL replaces the obsolete KVM_CREATE_PIT.
2165 Capability: KVM_CAP_PIT_STATE2
2168 Parameters: struct kvm_pit_state2 (out)
2169 Returns: 0 on success, -1 on error
2171 Retrieves the state of the in-kernel PIT model. Only valid after
2172 KVM_CREATE_PIT2. The state is returned in the following structure:
2174 struct kvm_pit_state2 {
2175 struct kvm_pit_channel_state channels[3];
2182 /* disable PIT in HPET legacy mode */
2183 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001
2185 This IOCTL replaces the obsolete KVM_GET_PIT.
2190 Capability: KVM_CAP_PIT_STATE2
2193 Parameters: struct kvm_pit_state2 (in)
2194 Returns: 0 on success, -1 on error
2196 Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2.
2197 See KVM_GET_PIT2 for details on struct kvm_pit_state2.
2199 This IOCTL replaces the obsolete KVM_SET_PIT.
2202 4.74 KVM_PPC_GET_SMMU_INFO
2204 Capability: KVM_CAP_PPC_GET_SMMU_INFO
2205 Architectures: powerpc
2208 Returns: 0 on success, -1 on error
2210 This populates and returns a structure describing the features of
2211 the "Server" class MMU emulation supported by KVM.
2212 This can in turn be used by userspace to generate the appropriate
2213 device-tree properties for the guest operating system.
2215 The structure contains some global information, followed by an
2216 array of supported segment page sizes:
2218 struct kvm_ppc_smmu_info {
2222 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ];
2225 The supported flags are:
2227 - KVM_PPC_PAGE_SIZES_REAL:
2228 When that flag is set, guest page sizes must "fit" the backing
2229 store page sizes. When not set, any page size in the list can
2230 be used regardless of how they are backed by userspace.
2232 - KVM_PPC_1T_SEGMENTS
2233 The emulated MMU supports 1T segments in addition to the
2236 The "slb_size" field indicates how many SLB entries are supported
2238 The "sps" array contains 8 entries indicating the supported base
2239 page sizes for a segment in increasing order. Each entry is defined
2242 struct kvm_ppc_one_seg_page_size {
2243 __u32 page_shift; /* Base page shift of segment (or 0) */
2244 __u32 slb_enc; /* SLB encoding for BookS */
2245 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ];
2248 An entry with a "page_shift" of 0 is unused. Because the array is
2249 organized in increasing order, a lookup can stop when encoutering
2252 The "slb_enc" field provides the encoding to use in the SLB for the
2253 page size. The bits are in positions such as the value can directly
2254 be OR'ed into the "vsid" argument of the slbmte instruction.
2256 The "enc" array is a list which for each of those segment base page
2257 size provides the list of supported actual page sizes (which can be
2258 only larger or equal to the base page size), along with the
2259 corresponding encoding in the hash PTE. Similarly, the array is
2260 8 entries sorted by increasing sizes and an entry with a "0" shift
2261 is an empty entry and a terminator:
2263 struct kvm_ppc_one_page_size {
2264 __u32 page_shift; /* Page shift (or 0) */
2265 __u32 pte_enc; /* Encoding in the HPTE (>>12) */
2268 The "pte_enc" field provides a value that can OR'ed into the hash
2269 PTE's RPN field (ie, it needs to be shifted left by 12 to OR it
2270 into the hash PTE second double word).
2274 Capability: KVM_CAP_IRQFD
2275 Architectures: x86 s390 arm arm64
2277 Parameters: struct kvm_irqfd (in)
2278 Returns: 0 on success, -1 on error
2280 Allows setting an eventfd to directly trigger a guest interrupt.
2281 kvm_irqfd.fd specifies the file descriptor to use as the eventfd and
2282 kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When
2283 an event is triggered on the eventfd, an interrupt is injected into
2284 the guest using the specified gsi pin. The irqfd is removed using
2285 the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd
2288 With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify
2289 mechanism allowing emulation of level-triggered, irqfd-based
2290 interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an
2291 additional eventfd in the kvm_irqfd.resamplefd field. When operating
2292 in resample mode, posting of an interrupt through kvm_irq.fd asserts
2293 the specified gsi in the irqchip. When the irqchip is resampled, such
2294 as from an EOI, the gsi is de-asserted and the user is notified via
2295 kvm_irqfd.resamplefd. It is the user's responsibility to re-queue
2296 the interrupt if the device making use of it still requires service.
2297 Note that closing the resamplefd is not sufficient to disable the
2298 irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment
2299 and need not be specified with KVM_IRQFD_FLAG_DEASSIGN.
2301 On ARM/ARM64, the gsi field in the kvm_irqfd struct specifies the Shared
2302 Peripheral Interrupt (SPI) index, such that the GIC interrupt ID is
2305 4.76 KVM_PPC_ALLOCATE_HTAB
2307 Capability: KVM_CAP_PPC_ALLOC_HTAB
2308 Architectures: powerpc
2310 Parameters: Pointer to u32 containing hash table order (in/out)
2311 Returns: 0 on success, -1 on error
2313 This requests the host kernel to allocate an MMU hash table for a
2314 guest using the PAPR paravirtualization interface. This only does
2315 anything if the kernel is configured to use the Book 3S HV style of
2316 virtualization. Otherwise the capability doesn't exist and the ioctl
2317 returns an ENOTTY error. The rest of this description assumes Book 3S
2320 There must be no vcpus running when this ioctl is called; if there
2321 are, it will do nothing and return an EBUSY error.
2323 The parameter is a pointer to a 32-bit unsigned integer variable
2324 containing the order (log base 2) of the desired size of the hash
2325 table, which must be between 18 and 46. On successful return from the
2326 ioctl, it will have been updated with the order of the hash table that
2329 If no hash table has been allocated when any vcpu is asked to run
2330 (with the KVM_RUN ioctl), the host kernel will allocate a
2331 default-sized hash table (16 MB).
2333 If this ioctl is called when a hash table has already been allocated,
2334 the kernel will clear out the existing hash table (zero all HPTEs) and
2335 return the hash table order in the parameter. (If the guest is using
2336 the virtualized real-mode area (VRMA) facility, the kernel will
2337 re-create the VMRA HPTEs on the next KVM_RUN of any vcpu.)
2339 4.77 KVM_S390_INTERRUPT
2343 Type: vm ioctl, vcpu ioctl
2344 Parameters: struct kvm_s390_interrupt (in)
2345 Returns: 0 on success, -1 on error
2347 Allows to inject an interrupt to the guest. Interrupts can be floating
2348 (vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type.
2350 Interrupt parameters are passed via kvm_s390_interrupt:
2352 struct kvm_s390_interrupt {
2358 type can be one of the following:
2360 KVM_S390_SIGP_STOP (vcpu) - sigp stop; optional flags in parm
2361 KVM_S390_PROGRAM_INT (vcpu) - program check; code in parm
2362 KVM_S390_SIGP_SET_PREFIX (vcpu) - sigp set prefix; prefix address in parm
2363 KVM_S390_RESTART (vcpu) - restart
2364 KVM_S390_INT_CLOCK_COMP (vcpu) - clock comparator interrupt
2365 KVM_S390_INT_CPU_TIMER (vcpu) - CPU timer interrupt
2366 KVM_S390_INT_VIRTIO (vm) - virtio external interrupt; external interrupt
2367 parameters in parm and parm64
2368 KVM_S390_INT_SERVICE (vm) - sclp external interrupt; sclp parameter in parm
2369 KVM_S390_INT_EMERGENCY (vcpu) - sigp emergency; source cpu in parm
2370 KVM_S390_INT_EXTERNAL_CALL (vcpu) - sigp external call; source cpu in parm
2371 KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) - compound value to indicate an
2372 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel);
2373 I/O interruption parameters in parm (subchannel) and parm64 (intparm,
2374 interruption subclass)
2375 KVM_S390_MCHK (vm, vcpu) - machine check interrupt; cr 14 bits in parm,
2376 machine check interrupt code in parm64 (note that
2377 machine checks needing further payload are not
2378 supported by this ioctl)
2380 Note that the vcpu ioctl is asynchronous to vcpu execution.
2382 4.78 KVM_PPC_GET_HTAB_FD
2384 Capability: KVM_CAP_PPC_HTAB_FD
2385 Architectures: powerpc
2387 Parameters: Pointer to struct kvm_get_htab_fd (in)
2388 Returns: file descriptor number (>= 0) on success, -1 on error
2390 This returns a file descriptor that can be used either to read out the
2391 entries in the guest's hashed page table (HPT), or to write entries to
2392 initialize the HPT. The returned fd can only be written to if the
2393 KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and
2394 can only be read if that bit is clear. The argument struct looks like
2397 /* For KVM_PPC_GET_HTAB_FD */
2398 struct kvm_get_htab_fd {
2404 /* Values for kvm_get_htab_fd.flags */
2405 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1)
2406 #define KVM_GET_HTAB_WRITE ((__u64)0x2)
2408 The `start_index' field gives the index in the HPT of the entry at
2409 which to start reading. It is ignored when writing.
2411 Reads on the fd will initially supply information about all
2412 "interesting" HPT entries. Interesting entries are those with the
2413 bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise
2414 all entries. When the end of the HPT is reached, the read() will
2415 return. If read() is called again on the fd, it will start again from
2416 the beginning of the HPT, but will only return HPT entries that have
2417 changed since they were last read.
2419 Data read or written is structured as a header (8 bytes) followed by a
2420 series of valid HPT entries (16 bytes) each. The header indicates how
2421 many valid HPT entries there are and how many invalid entries follow
2422 the valid entries. The invalid entries are not represented explicitly
2423 in the stream. The header format is:
2425 struct kvm_get_htab_header {
2431 Writes to the fd create HPT entries starting at the index given in the
2432 header; first `n_valid' valid entries with contents from the data
2433 written, then `n_invalid' invalid entries, invalidating any previously
2434 valid entries found.
2436 4.79 KVM_CREATE_DEVICE
2438 Capability: KVM_CAP_DEVICE_CTRL
2440 Parameters: struct kvm_create_device (in/out)
2441 Returns: 0 on success, -1 on error
2443 ENODEV: The device type is unknown or unsupported
2444 EEXIST: Device already created, and this type of device may not
2445 be instantiated multiple times
2447 Other error conditions may be defined by individual device types or
2448 have their standard meanings.
2450 Creates an emulated device in the kernel. The file descriptor returned
2451 in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR.
2453 If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the
2454 device type is supported (not necessarily whether it can be created
2457 Individual devices should not define flags. Attributes should be used
2458 for specifying any behavior that is not implied by the device type
2461 struct kvm_create_device {
2462 __u32 type; /* in: KVM_DEV_TYPE_xxx */
2463 __u32 fd; /* out: device handle */
2464 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */
2467 4.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR
2469 Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device
2470 Type: device ioctl, vm ioctl
2471 Parameters: struct kvm_device_attr
2472 Returns: 0 on success, -1 on error
2474 ENXIO: The group or attribute is unknown/unsupported for this device
2475 EPERM: The attribute cannot (currently) be accessed this way
2476 (e.g. read-only attribute, or attribute that only makes
2477 sense when the device is in a different state)
2479 Other error conditions may be defined by individual device types.
2481 Gets/sets a specified piece of device configuration and/or state. The
2482 semantics are device-specific. See individual device documentation in
2483 the "devices" directory. As with ONE_REG, the size of the data
2484 transferred is defined by the particular attribute.
2486 struct kvm_device_attr {
2487 __u32 flags; /* no flags currently defined */
2488 __u32 group; /* device-defined */
2489 __u64 attr; /* group-defined */
2490 __u64 addr; /* userspace address of attr data */
2493 4.81 KVM_HAS_DEVICE_ATTR
2495 Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device
2496 Type: device ioctl, vm ioctl
2497 Parameters: struct kvm_device_attr
2498 Returns: 0 on success, -1 on error
2500 ENXIO: The group or attribute is unknown/unsupported for this device
2502 Tests whether a device supports a particular attribute. A successful
2503 return indicates the attribute is implemented. It does not necessarily
2504 indicate that the attribute can be read or written in the device's
2505 current state. "addr" is ignored.
2507 4.82 KVM_ARM_VCPU_INIT
2510 Architectures: arm, arm64
2512 Parameters: struct kvm_vcpu_init (in)
2513 Returns: 0 on success; -1 on error
2515 Â EINVAL: Â Â Â the target is unknown, or the combination of features is invalid.
2516 Â ENOENT: Â Â Â a features bit specified is unknown.
2518 This tells KVM what type of CPU to present to the guest, and what
2519 optional features it should have. Â This will cause a reset of the cpu
2520 registers to their initial values. Â If this is not called, KVM_RUN will
2521 return ENOEXEC for that vcpu.
2523 Note that because some registers reflect machine topology, all vcpus
2524 should be created before this ioctl is invoked.
2526 Userspace can call this function multiple times for a given vcpu, including
2527 after the vcpu has been run. This will reset the vcpu to its initial
2528 state. All calls to this function after the initial call must use the same
2529 target and same set of feature flags, otherwise EINVAL will be returned.
2532 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state.
2533 Depends on KVM_CAP_ARM_PSCI. If not set, the CPU will be powered on
2534 and execute guest code when KVM_RUN is called.
2535 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode.
2536 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only).
2537 - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 for the CPU.
2538 Depends on KVM_CAP_ARM_PSCI_0_2.
2541 4.83 KVM_ARM_PREFERRED_TARGET
2544 Architectures: arm, arm64
2546 Parameters: struct struct kvm_vcpu_init (out)
2547 Returns: 0 on success; -1 on error
2549 ENODEV: no preferred target available for the host
2551 This queries KVM for preferred CPU target type which can be emulated
2552 by KVM on underlying host.
2554 The ioctl returns struct kvm_vcpu_init instance containing information
2555 about preferred CPU target type and recommended features for it. The
2556 kvm_vcpu_init->features bitmap returned will have feature bits set if
2557 the preferred target recommends setting these features, but this is
2560 The information returned by this ioctl can be used to prepare an instance
2561 of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in
2562 in VCPU matching underlying host.
2565 4.84 KVM_GET_REG_LIST
2568 Architectures: arm, arm64, mips
2570 Parameters: struct kvm_reg_list (in/out)
2571 Returns: 0 on success; -1 on error
2573 Â E2BIG: Â Â Â Â the reg index list is too big to fit in the array specified by
2574 Â Â Â Â Â Â Â Â Â Â Â Â the user (the number required will be written into n).
2576 struct kvm_reg_list {
2577 __u64 n; /* number of registers in reg[] */
2581 This ioctl returns the guest registers that are supported for the
2582 KVM_GET_ONE_REG/KVM_SET_ONE_REG calls.
2585 4.85 KVM_ARM_SET_DEVICE_ADDR (deprecated)
2587 Capability: KVM_CAP_ARM_SET_DEVICE_ADDR
2588 Architectures: arm, arm64
2590 Parameters: struct kvm_arm_device_address (in)
2591 Returns: 0 on success, -1 on error
2593 ENODEV: The device id is unknown
2594 ENXIO: Device not supported on current system
2595 EEXIST: Address already set
2596 E2BIG: Address outside guest physical address space
2597 EBUSY: Address overlaps with other device range
2599 struct kvm_arm_device_addr {
2604 Specify a device address in the guest's physical address space where guests
2605 can access emulated or directly exposed devices, which the host kernel needs
2606 to know about. The id field is an architecture specific identifier for a
2609 ARM/arm64 divides the id field into two parts, a device id and an
2610 address type id specific to the individual device.
2612 Â bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 |
2613 field: | 0x00000000 | device id | addr type id |
2615 ARM/arm64 currently only require this when using the in-kernel GIC
2616 support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2
2617 as the device id. When setting the base address for the guest's
2618 mapping of the VGIC virtual CPU and distributor interface, the ioctl
2619 must be called after calling KVM_CREATE_IRQCHIP, but before calling
2620 KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the
2621 base addresses will return -EEXIST.
2623 Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API
2624 should be used instead.
2627 4.86 KVM_PPC_RTAS_DEFINE_TOKEN
2629 Capability: KVM_CAP_PPC_RTAS
2632 Parameters: struct kvm_rtas_token_args
2633 Returns: 0 on success, -1 on error
2635 Defines a token value for a RTAS (Run Time Abstraction Services)
2636 service in order to allow it to be handled in the kernel. The
2637 argument struct gives the name of the service, which must be the name
2638 of a service that has a kernel-side implementation. If the token
2639 value is non-zero, it will be associated with that service, and
2640 subsequent RTAS calls by the guest specifying that token will be
2641 handled by the kernel. If the token value is 0, then any token
2642 associated with the service will be forgotten, and subsequent RTAS
2643 calls by the guest for that service will be passed to userspace to be
2646 4.87 KVM_SET_GUEST_DEBUG
2648 Capability: KVM_CAP_SET_GUEST_DEBUG
2649 Architectures: x86, s390, ppc
2651 Parameters: struct kvm_guest_debug (in)
2652 Returns: 0 on success; -1 on error
2654 struct kvm_guest_debug {
2657 struct kvm_guest_debug_arch arch;
2660 Set up the processor specific debug registers and configure vcpu for
2661 handling guest debug events. There are two parts to the structure, the
2662 first a control bitfield indicates the type of debug events to handle
2663 when running. Common control bits are:
2665 - KVM_GUESTDBG_ENABLE: guest debugging is enabled
2666 - KVM_GUESTDBG_SINGLESTEP: the next run should single-step
2668 The top 16 bits of the control field are architecture specific control
2669 flags which can include the following:
2671 - KVM_GUESTDBG_USE_SW_BP: using software breakpoints [x86]
2672 - KVM_GUESTDBG_USE_HW_BP: using hardware breakpoints [x86, s390]
2673 - KVM_GUESTDBG_INJECT_DB: inject DB type exception [x86]
2674 - KVM_GUESTDBG_INJECT_BP: inject BP type exception [x86]
2675 - KVM_GUESTDBG_EXIT_PENDING: trigger an immediate guest exit [s390]
2677 For example KVM_GUESTDBG_USE_SW_BP indicates that software breakpoints
2678 are enabled in memory so we need to ensure breakpoint exceptions are
2679 correctly trapped and the KVM run loop exits at the breakpoint and not
2680 running off into the normal guest vector. For KVM_GUESTDBG_USE_HW_BP
2681 we need to ensure the guest vCPUs architecture specific registers are
2682 updated to the correct (supplied) values.
2684 The second part of the structure is architecture specific and
2685 typically contains a set of debug registers.
2687 When debug events exit the main run loop with the reason
2688 KVM_EXIT_DEBUG with the kvm_debug_exit_arch part of the kvm_run
2689 structure containing architecture specific debug information.
2691 4.88 KVM_GET_EMULATED_CPUID
2693 Capability: KVM_CAP_EXT_EMUL_CPUID
2696 Parameters: struct kvm_cpuid2 (in/out)
2697 Returns: 0 on success, -1 on error
2702 struct kvm_cpuid_entry2 entries[0];
2705 The member 'flags' is used for passing flags from userspace.
2707 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
2708 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
2709 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
2711 struct kvm_cpuid_entry2 {
2722 This ioctl returns x86 cpuid features which are emulated by
2723 kvm.Userspace can use the information returned by this ioctl to query
2724 which features are emulated by kvm instead of being present natively.
2726 Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2
2727 structure with the 'nent' field indicating the number of entries in
2728 the variable-size array 'entries'. If the number of entries is too low
2729 to describe the cpu capabilities, an error (E2BIG) is returned. If the
2730 number is too high, the 'nent' field is adjusted and an error (ENOMEM)
2731 is returned. If the number is just right, the 'nent' field is adjusted
2732 to the number of valid entries in the 'entries' array, which is then
2735 The entries returned are the set CPUID bits of the respective features
2736 which kvm emulates, as returned by the CPUID instruction, with unknown
2737 or unsupported feature bits cleared.
2739 Features like x2apic, for example, may not be present in the host cpu
2740 but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be
2741 emulated efficiently and thus not included here.
2743 The fields in each entry are defined as follows:
2745 function: the eax value used to obtain the entry
2746 index: the ecx value used to obtain the entry (for entries that are
2748 flags: an OR of zero or more of the following:
2749 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
2750 if the index field is valid
2751 KVM_CPUID_FLAG_STATEFUL_FUNC:
2752 if cpuid for this function returns different values for successive
2753 invocations; there will be several entries with the same function,
2754 all with this flag set
2755 KVM_CPUID_FLAG_STATE_READ_NEXT:
2756 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
2757 the first entry to be read by a cpu
2758 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
2759 this function/index combination
2761 4.89 KVM_S390_MEM_OP
2763 Capability: KVM_CAP_S390_MEM_OP
2766 Parameters: struct kvm_s390_mem_op (in)
2767 Returns: = 0 on success,
2768 < 0 on generic error (e.g. -EFAULT or -ENOMEM),
2769 > 0 if an exception occurred while walking the page tables
2771 Read or write data from/to the logical (virtual) memory of a VPCU.
2773 Parameters are specified via the following structure:
2775 struct kvm_s390_mem_op {
2776 __u64 gaddr; /* the guest address */
2777 __u64 flags; /* flags */
2778 __u32 size; /* amount of bytes */
2779 __u32 op; /* type of operation */
2780 __u64 buf; /* buffer in userspace */
2781 __u8 ar; /* the access register number */
2782 __u8 reserved[31]; /* should be set to 0 */
2785 The type of operation is specified in the "op" field. It is either
2786 KVM_S390_MEMOP_LOGICAL_READ for reading from logical memory space or
2787 KVM_S390_MEMOP_LOGICAL_WRITE for writing to logical memory space. The
2788 KVM_S390_MEMOP_F_CHECK_ONLY flag can be set in the "flags" field to check
2789 whether the corresponding memory access would create an access exception
2790 (without touching the data in the memory at the destination). In case an
2791 access exception occurred while walking the MMU tables of the guest, the
2792 ioctl returns a positive error number to indicate the type of exception.
2793 This exception is also raised directly at the corresponding VCPU if the
2794 flag KVM_S390_MEMOP_F_INJECT_EXCEPTION is set in the "flags" field.
2796 The start address of the memory region has to be specified in the "gaddr"
2797 field, and the length of the region in the "size" field. "buf" is the buffer
2798 supplied by the userspace application where the read data should be written
2799 to for KVM_S390_MEMOP_LOGICAL_READ, or where the data that should be written
2800 is stored for a KVM_S390_MEMOP_LOGICAL_WRITE. "buf" is unused and can be NULL
2801 when KVM_S390_MEMOP_F_CHECK_ONLY is specified. "ar" designates the access
2802 register number to be used.
2804 The "reserved" field is meant for future extensions. It is not used by
2805 KVM with the currently defined set of flags.
2807 4.90 KVM_S390_GET_SKEYS
2809 Capability: KVM_CAP_S390_SKEYS
2812 Parameters: struct kvm_s390_skeys
2813 Returns: 0 on success, KVM_S390_GET_KEYS_NONE if guest is not using storage
2814 keys, negative value on error
2816 This ioctl is used to get guest storage key values on the s390
2817 architecture. The ioctl takes parameters via the kvm_s390_skeys struct.
2819 struct kvm_s390_skeys {
2822 __u64 skeydata_addr;
2827 The start_gfn field is the number of the first guest frame whose storage keys
2830 The count field is the number of consecutive frames (starting from start_gfn)
2831 whose storage keys to get. The count field must be at least 1 and the maximum
2832 allowed value is defined as KVM_S390_SKEYS_ALLOC_MAX. Values outside this range
2833 will cause the ioctl to return -EINVAL.
2835 The skeydata_addr field is the address to a buffer large enough to hold count
2836 bytes. This buffer will be filled with storage key data by the ioctl.
2838 4.91 KVM_S390_SET_SKEYS
2840 Capability: KVM_CAP_S390_SKEYS
2843 Parameters: struct kvm_s390_skeys
2844 Returns: 0 on success, negative value on error
2846 This ioctl is used to set guest storage key values on the s390
2847 architecture. The ioctl takes parameters via the kvm_s390_skeys struct.
2848 See section on KVM_S390_GET_SKEYS for struct definition.
2850 The start_gfn field is the number of the first guest frame whose storage keys
2853 The count field is the number of consecutive frames (starting from start_gfn)
2854 whose storage keys to get. The count field must be at least 1 and the maximum
2855 allowed value is defined as KVM_S390_SKEYS_ALLOC_MAX. Values outside this range
2856 will cause the ioctl to return -EINVAL.
2858 The skeydata_addr field is the address to a buffer containing count bytes of
2859 storage keys. Each byte in the buffer will be set as the storage key for a
2860 single frame starting at start_gfn for count frames.
2862 Note: If any architecturally invalid key value is found in the given data then
2863 the ioctl will return -EINVAL.
2867 Capability: KVM_CAP_S390_INJECT_IRQ
2870 Parameters: struct kvm_s390_irq (in)
2871 Returns: 0 on success, -1 on error
2873 EINVAL: interrupt type is invalid
2874 type is KVM_S390_SIGP_STOP and flag parameter is invalid value
2875 type is KVM_S390_INT_EXTERNAL_CALL and code is bigger
2876 than the maximum of VCPUs
2877 EBUSY: type is KVM_S390_SIGP_SET_PREFIX and vcpu is not stopped
2878 type is KVM_S390_SIGP_STOP and a stop irq is already pending
2879 type is KVM_S390_INT_EXTERNAL_CALL and an external call interrupt
2882 Allows to inject an interrupt to the guest.
2884 Using struct kvm_s390_irq as a parameter allows
2885 to inject additional payload which is not
2886 possible via KVM_S390_INTERRUPT.
2888 Interrupt parameters are passed via kvm_s390_irq:
2890 struct kvm_s390_irq {
2893 struct kvm_s390_io_info io;
2894 struct kvm_s390_ext_info ext;
2895 struct kvm_s390_pgm_info pgm;
2896 struct kvm_s390_emerg_info emerg;
2897 struct kvm_s390_extcall_info extcall;
2898 struct kvm_s390_prefix_info prefix;
2899 struct kvm_s390_stop_info stop;
2900 struct kvm_s390_mchk_info mchk;
2905 type can be one of the following:
2907 KVM_S390_SIGP_STOP - sigp stop; parameter in .stop
2908 KVM_S390_PROGRAM_INT - program check; parameters in .pgm
2909 KVM_S390_SIGP_SET_PREFIX - sigp set prefix; parameters in .prefix
2910 KVM_S390_RESTART - restart; no parameters
2911 KVM_S390_INT_CLOCK_COMP - clock comparator interrupt; no parameters
2912 KVM_S390_INT_CPU_TIMER - CPU timer interrupt; no parameters
2913 KVM_S390_INT_EMERGENCY - sigp emergency; parameters in .emerg
2914 KVM_S390_INT_EXTERNAL_CALL - sigp external call; parameters in .extcall
2915 KVM_S390_MCHK - machine check interrupt; parameters in .mchk
2918 Note that the vcpu ioctl is asynchronous to vcpu execution.
2920 4.94 KVM_S390_GET_IRQ_STATE
2922 Capability: KVM_CAP_S390_IRQ_STATE
2925 Parameters: struct kvm_s390_irq_state (out)
2926 Returns: >= number of bytes copied into buffer,
2927 -EINVAL if buffer size is 0,
2928 -ENOBUFS if buffer size is too small to fit all pending interrupts,
2929 -EFAULT if the buffer address was invalid
2931 This ioctl allows userspace to retrieve the complete state of all currently
2932 pending interrupts in a single buffer. Use cases include migration
2933 and introspection. The parameter structure contains the address of a
2934 userspace buffer and its length:
2936 struct kvm_s390_irq_state {
2943 Userspace passes in the above struct and for each pending interrupt a
2944 struct kvm_s390_irq is copied to the provided buffer.
2946 If -ENOBUFS is returned the buffer provided was too small and userspace
2947 may retry with a bigger buffer.
2949 4.95 KVM_S390_SET_IRQ_STATE
2951 Capability: KVM_CAP_S390_IRQ_STATE
2954 Parameters: struct kvm_s390_irq_state (in)
2955 Returns: 0 on success,
2956 -EFAULT if the buffer address was invalid,
2957 -EINVAL for an invalid buffer length (see below),
2958 -EBUSY if there were already interrupts pending,
2959 errors occurring when actually injecting the
2960 interrupt. See KVM_S390_IRQ.
2962 This ioctl allows userspace to set the complete state of all cpu-local
2963 interrupts currently pending for the vcpu. It is intended for restoring
2964 interrupt state after a migration. The input parameter is a userspace buffer
2965 containing a struct kvm_s390_irq_state:
2967 struct kvm_s390_irq_state {
2973 The userspace memory referenced by buf contains a struct kvm_s390_irq
2974 for each interrupt to be injected into the guest.
2975 If one of the interrupts could not be injected for some reason the
2978 len must be a multiple of sizeof(struct kvm_s390_irq). It must be > 0
2979 and it must not exceed (max_vcpus + 32) * sizeof(struct kvm_s390_irq),
2980 which is the maximum number of possibly pending cpu-local interrupts.
2982 5. The kvm_run structure
2983 ------------------------
2985 Application code obtains a pointer to the kvm_run structure by
2986 mmap()ing a vcpu fd. From that point, application code can control
2987 execution by changing fields in kvm_run prior to calling the KVM_RUN
2988 ioctl, and obtain information about the reason KVM_RUN returned by
2989 looking up structure members.
2993 __u8 request_interrupt_window;
2995 Request that KVM_RUN return when it becomes possible to inject external
2996 interrupts into the guest. Useful in conjunction with KVM_INTERRUPT.
3003 When KVM_RUN has returned successfully (return value 0), this informs
3004 application code why KVM_RUN has returned. Allowable values for this
3005 field are detailed below.
3007 __u8 ready_for_interrupt_injection;
3009 If request_interrupt_window has been specified, this field indicates
3010 an interrupt can be injected now with KVM_INTERRUPT.
3014 The value of the current interrupt flag. Only valid if in-kernel
3015 local APIC is not used.
3019 /* in (pre_kvm_run), out (post_kvm_run) */
3022 The value of the cr8 register. Only valid if in-kernel local APIC is
3023 not used. Both input and output.
3027 The value of the APIC BASE msr. Only valid if in-kernel local
3028 APIC is not used. Both input and output.
3031 /* KVM_EXIT_UNKNOWN */
3033 __u64 hardware_exit_reason;
3036 If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown
3037 reasons. Further architecture-specific information is available in
3038 hardware_exit_reason.
3040 /* KVM_EXIT_FAIL_ENTRY */
3042 __u64 hardware_entry_failure_reason;
3045 If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due
3046 to unknown reasons. Further architecture-specific information is
3047 available in hardware_entry_failure_reason.
3049 /* KVM_EXIT_EXCEPTION */
3059 #define KVM_EXIT_IO_IN 0
3060 #define KVM_EXIT_IO_OUT 1
3062 __u8 size; /* bytes */
3065 __u64 data_offset; /* relative to kvm_run start */
3068 If exit_reason is KVM_EXIT_IO, then the vcpu has
3069 executed a port I/O instruction which could not be satisfied by kvm.
3070 data_offset describes where the data is located (KVM_EXIT_IO_OUT) or
3071 where kvm expects application code to place the data for the next
3072 KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array.
3075 struct kvm_debug_exit_arch arch;
3088 If exit_reason is KVM_EXIT_MMIO, then the vcpu has
3089 executed a memory-mapped I/O instruction which could not be satisfied
3090 by kvm. The 'data' member contains the written data if 'is_write' is
3091 true, and should be filled by application code otherwise.
3093 The 'data' member contains, in its first 'len' bytes, the value as it would
3094 appear if the VCPU performed a load or store of the appropriate width directly
3097 NOTE: For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_PAPR and
3098 KVM_EXIT_EPR the corresponding
3099 operations are complete (and guest state is consistent) only after userspace
3100 has re-entered the kernel with KVM_RUN. The kernel side will first finish
3101 incomplete operations and then check for pending signals. Userspace
3102 can re-enter the guest with an unmasked signal pending to complete
3105 /* KVM_EXIT_HYPERCALL */
3114 Unused. This was once used for 'hypercall to userspace'. To implement
3115 such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390).
3116 Note KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO.
3118 /* KVM_EXIT_TPR_ACCESS */
3125 To be documented (KVM_TPR_ACCESS_REPORTING).
3127 /* KVM_EXIT_S390_SIEIC */
3130 __u64 mask; /* psw upper half */
3131 __u64 addr; /* psw lower half */
3138 /* KVM_EXIT_S390_RESET */
3139 #define KVM_S390_RESET_POR 1
3140 #define KVM_S390_RESET_CLEAR 2
3141 #define KVM_S390_RESET_SUBSYSTEM 4
3142 #define KVM_S390_RESET_CPU_INIT 8
3143 #define KVM_S390_RESET_IPL 16
3144 __u64 s390_reset_flags;
3148 /* KVM_EXIT_S390_UCONTROL */
3150 __u64 trans_exc_code;
3154 s390 specific. A page fault has occurred for a user controlled virtual
3155 machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be
3156 resolved by the kernel.
3157 The program code and the translation exception code that were placed
3158 in the cpu's lowcore are presented here as defined by the z Architecture
3159 Principles of Operation Book in the Chapter for Dynamic Address Translation
3169 Deprecated - was used for 440 KVM.
3176 MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch
3177 hypercalls and exit with this exit struct that contains all the guest gprs.
3179 If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall.
3180 Userspace can now handle the hypercall and when it's done modify the gprs as
3181 necessary. Upon guest entry all guest GPRs will then be replaced by the values
3184 /* KVM_EXIT_PAPR_HCALL */
3191 This is used on 64-bit PowerPC when emulating a pSeries partition,
3192 e.g. with the 'pseries' machine type in qemu. It occurs when the
3193 guest does a hypercall using the 'sc 1' instruction. The 'nr' field
3194 contains the hypercall number (from the guest R3), and 'args' contains
3195 the arguments (from the guest R4 - R12). Userspace should put the
3196 return code in 'ret' and any extra returned values in args[].
3197 The possible hypercalls are defined in the Power Architecture Platform
3198 Requirements (PAPR) document available from www.power.org (free
3199 developer registration required to access it).
3201 /* KVM_EXIT_S390_TSCH */
3203 __u16 subchannel_id;
3204 __u16 subchannel_nr;
3211 s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled
3212 and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O
3213 interrupt for the target subchannel has been dequeued and subchannel_id,
3214 subchannel_nr, io_int_parm and io_int_word contain the parameters for that
3215 interrupt. ipb is needed for instruction parameter decoding.
3222 On FSL BookE PowerPC chips, the interrupt controller has a fast patch
3223 interrupt acknowledge path to the core. When the core successfully
3224 delivers an interrupt, it automatically populates the EPR register with
3225 the interrupt vector number and acknowledges the interrupt inside
3226 the interrupt controller.
3228 In case the interrupt controller lives in user space, we need to do
3229 the interrupt acknowledge cycle through it to fetch the next to be
3230 delivered interrupt vector using this exit.
3232 It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an
3233 external interrupt has just been delivered into the guest. User space
3234 should put the acknowledged interrupt vector into the 'epr' field.
3236 /* KVM_EXIT_SYSTEM_EVENT */
3238 #define KVM_SYSTEM_EVENT_SHUTDOWN 1
3239 #define KVM_SYSTEM_EVENT_RESET 2
3244 If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered
3245 a system-level event using some architecture specific mechanism (hypercall
3246 or some special instruction). In case of ARM/ARM64, this is triggered using
3247 HVC instruction based PSCI call from the vcpu. The 'type' field describes
3248 the system-level event type. The 'flags' field describes architecture
3249 specific flags for the system-level event.
3251 Valid values for 'type' are:
3252 KVM_SYSTEM_EVENT_SHUTDOWN -- the guest has requested a shutdown of the
3253 VM. Userspace is not obliged to honour this, and if it does honour
3254 this does not need to destroy the VM synchronously (ie it may call
3255 KVM_RUN again before shutdown finally occurs).
3256 KVM_SYSTEM_EVENT_RESET -- the guest has requested a reset of the VM.
3257 As with SHUTDOWN, userspace can choose to ignore the request, or
3258 to schedule the reset to occur in the future and may call KVM_RUN again.
3260 /* Fix the size of the union. */
3265 * shared registers between kvm and userspace.
3266 * kvm_valid_regs specifies the register classes set by the host
3267 * kvm_dirty_regs specified the register classes dirtied by userspace
3268 * struct kvm_sync_regs is architecture specific, as well as the
3269 * bits for kvm_valid_regs and kvm_dirty_regs
3271 __u64 kvm_valid_regs;
3272 __u64 kvm_dirty_regs;
3274 struct kvm_sync_regs regs;
3278 If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access
3279 certain guest registers without having to call SET/GET_*REGS. Thus we can
3280 avoid some system call overhead if userspace has to handle the exit.
3281 Userspace can query the validity of the structure by checking
3282 kvm_valid_regs for specific bits. These bits are architecture specific
3283 and usually define the validity of a groups of registers. (e.g. one bit
3284 for general purpose registers)
3286 Please note that the kernel is allowed to use the kvm_run structure as the
3287 primary storage for certain register types. Therefore, the kernel may use the
3288 values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set.
3294 6. Capabilities that can be enabled on vCPUs
3295 --------------------------------------------
3297 There are certain capabilities that change the behavior of the virtual CPU or
3298 the virtual machine when enabled. To enable them, please see section 4.37.
3299 Below you can find a list of capabilities and what their effect on the vCPU or
3300 the virtual machine is when enabling them.
3302 The following information is provided along with the description:
3304 Architectures: which instruction set architectures provide this ioctl.
3305 x86 includes both i386 and x86_64.
3307 Target: whether this is a per-vcpu or per-vm capability.
3309 Parameters: what parameters are accepted by the capability.
3311 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
3312 are not detailed, but errors with specific meanings are.
3320 Returns: 0 on success; -1 on error
3322 This capability enables interception of OSI hypercalls that otherwise would
3323 be treated as normal system calls to be injected into the guest. OSI hypercalls
3324 were invented by Mac-on-Linux to have a standardized communication mechanism
3325 between the guest and the host.
3327 When this capability is enabled, KVM_EXIT_OSI can occur.
3330 6.2 KVM_CAP_PPC_PAPR
3335 Returns: 0 on success; -1 on error
3337 This capability enables interception of PAPR hypercalls. PAPR hypercalls are
3338 done using the hypercall instruction "sc 1".
3340 It also sets the guest privilege level to "supervisor" mode. Usually the guest
3341 runs in "hypervisor" privilege mode with a few missing features.
3343 In addition to the above, it changes the semantics of SDR1. In this mode, the
3344 HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the
3345 HTAB invisible to the guest.
3347 When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur.
3354 Parameters: args[0] is the address of a struct kvm_config_tlb
3355 Returns: 0 on success; -1 on error
3357 struct kvm_config_tlb {
3364 Configures the virtual CPU's TLB array, establishing a shared memory area
3365 between userspace and KVM. The "params" and "array" fields are userspace
3366 addresses of mmu-type-specific data structures. The "array_len" field is an
3367 safety mechanism, and should be set to the size in bytes of the memory that
3368 userspace has reserved for the array. It must be at least the size dictated
3369 by "mmu_type" and "params".
3371 While KVM_RUN is active, the shared region is under control of KVM. Its
3372 contents are undefined, and any modification by userspace results in
3373 boundedly undefined behavior.
3375 On return from KVM_RUN, the shared region will reflect the current state of
3376 the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB
3377 to tell KVM which entries have been changed, prior to calling KVM_RUN again
3380 For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV:
3381 - The "params" field is of type "struct kvm_book3e_206_tlb_params".
3382 - The "array" field points to an array of type "struct
3383 kvm_book3e_206_tlb_entry".
3384 - The array consists of all entries in the first TLB, followed by all
3385 entries in the second TLB.
3386 - Within a TLB, entries are ordered first by increasing set number. Within a
3387 set, entries are ordered by way (increasing ESEL).
3388 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1)
3389 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value.
3390 - The tsize field of mas1 shall be set to 4K on TLB0, even though the
3391 hardware ignores this value for TLB0.
3393 6.4 KVM_CAP_S390_CSS_SUPPORT
3398 Returns: 0 on success; -1 on error
3400 This capability enables support for handling of channel I/O instructions.
3402 TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are
3403 handled in-kernel, while the other I/O instructions are passed to userspace.
3405 When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST
3406 SUBCHANNEL intercepts.
3408 Note that even though this capability is enabled per-vcpu, the complete
3409 virtual machine is affected.
3415 Parameters: args[0] defines whether the proxy facility is active
3416 Returns: 0 on success; -1 on error
3418 This capability enables or disables the delivery of interrupts through the
3419 external proxy facility.
3421 When enabled (args[0] != 0), every time the guest gets an external interrupt
3422 delivered, it automatically exits into user space with a KVM_EXIT_EPR exit
3423 to receive the topmost interrupt vector.
3425 When disabled (args[0] == 0), behavior is as if this facility is unsupported.
3427 When this capability is enabled, KVM_EXIT_EPR can occur.
3429 6.6 KVM_CAP_IRQ_MPIC
3432 Parameters: args[0] is the MPIC device fd
3433 args[1] is the MPIC CPU number for this vcpu
3435 This capability connects the vcpu to an in-kernel MPIC device.
3437 6.7 KVM_CAP_IRQ_XICS
3441 Parameters: args[0] is the XICS device fd
3442 args[1] is the XICS CPU number (server ID) for this vcpu
3444 This capability connects the vcpu to an in-kernel XICS device.
3446 6.8 KVM_CAP_S390_IRQCHIP
3452 This capability enables the in-kernel irqchip for s390. Please refer to
3453 "4.24 KVM_CREATE_IRQCHIP" for details.
3455 6.9 KVM_CAP_MIPS_FPU
3459 Parameters: args[0] is reserved for future use (should be 0).
3461 This capability allows the use of the host Floating Point Unit by the guest. It
3462 allows the Config1.FP bit to be set to enable the FPU in the guest. Once this is
3463 done the KVM_REG_MIPS_FPR_* and KVM_REG_MIPS_FCR_* registers can be accessed
3464 (depending on the current guest FPU register mode), and the Status.FR,
3465 Config5.FRE bits are accessible via the KVM API and also from the guest,
3466 depending on them being supported by the FPU.
3468 6.10 KVM_CAP_MIPS_MSA
3472 Parameters: args[0] is reserved for future use (should be 0).
3474 This capability allows the use of the MIPS SIMD Architecture (MSA) by the guest.
3475 It allows the Config3.MSAP bit to be set to enable the use of MSA by the guest.
3476 Once this is done the KVM_REG_MIPS_VEC_* and KVM_REG_MIPS_MSA_* registers can be
3477 accessed, and the Config5.MSAEn bit is accessible via the KVM API and also from
3480 7. Capabilities that can be enabled on VMs
3481 ------------------------------------------
3483 There are certain capabilities that change the behavior of the virtual
3484 machine when enabled. To enable them, please see section 4.37. Below
3485 you can find a list of capabilities and what their effect on the VM
3486 is when enabling them.
3488 The following information is provided along with the description:
3490 Architectures: which instruction set architectures provide this ioctl.
3491 x86 includes both i386 and x86_64.
3493 Parameters: what parameters are accepted by the capability.
3495 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
3496 are not detailed, but errors with specific meanings are.
3499 7.1 KVM_CAP_PPC_ENABLE_HCALL
3502 Parameters: args[0] is the sPAPR hcall number
3503 args[1] is 0 to disable, 1 to enable in-kernel handling
3505 This capability controls whether individual sPAPR hypercalls (hcalls)
3506 get handled by the kernel or not. Enabling or disabling in-kernel
3507 handling of an hcall is effective across the VM. On creation, an
3508 initial set of hcalls are enabled for in-kernel handling, which
3509 consists of those hcalls for which in-kernel handlers were implemented
3510 before this capability was implemented. If disabled, the kernel will
3511 not to attempt to handle the hcall, but will always exit to userspace
3512 to handle it. Note that it may not make sense to enable some and
3513 disable others of a group of related hcalls, but KVM does not prevent
3514 userspace from doing that.
3516 If the hcall number specified is not one that has an in-kernel
3517 implementation, the KVM_ENABLE_CAP ioctl will fail with an EINVAL
3520 7.2 KVM_CAP_S390_USER_SIGP
3525 This capability controls which SIGP orders will be handled completely in user
3526 space. With this capability enabled, all fast orders will be handled completely
3532 - CONDITIONAL EMERGENCY SIGNAL
3534 All other orders will be handled completely in user space.
3536 Only privileged operation exceptions will be checked for in the kernel (or even
3537 in the hardware prior to interception). If this capability is not enabled, the
3538 old way of handling SIGP orders is used (partially in kernel and user space).
3540 7.3 KVM_CAP_S390_VECTOR_REGISTERS
3544 Returns: 0 on success, negative value on error
3546 Allows use of the vector registers introduced with z13 processor, and
3547 provides for the synchronization between host and user space. Will
3548 return -EINVAL if the machine does not support vectors.
3550 7.4 KVM_CAP_S390_USER_STSI
3555 This capability allows post-handlers for the STSI instruction. After
3556 initial handling in the kernel, KVM exits to user space with
3557 KVM_EXIT_S390_STSI to allow user space to insert further data.
3559 Before exiting to userspace, kvm handlers should fill in s390_stsi field of
3570 @addr - guest address of STSI SYSIB
3574 @ar - access register number
3576 KVM handlers should exit to userspace with rc = -EREMOTE.
3579 8. Other capabilities.
3580 ----------------------
3582 This section lists capabilities that give information about other
3583 features of the KVM implementation.
3585 8.1 KVM_CAP_PPC_HWRNG
3589 This capability, if KVM_CHECK_EXTENSION indicates that it is
3590 available, means that that the kernel has an implementation of the
3591 H_RANDOM hypercall backed by a hardware random-number generator.
3592 If present, the kernel H_RANDOM handler can be enabled for guest use
3593 with the KVM_CAP_PPC_ENABLE_HCALL capability.