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.
114 You probably want to use 0 as machine type.
116 In order to create user controlled virtual machines on S390, check
117 KVM_CAP_S390_UCONTROL and use the flag KVM_VM_S390_UCONTROL as
118 privileged user (CAP_SYS_ADMIN).
120 To use hardware assisted virtualization on MIPS (VZ ASE) rather than
121 the default trap & emulate implementation (which changes the virtual
122 memory layout to fit in user mode), check KVM_CAP_MIPS_VZ and use the
126 4.3 KVM_GET_MSR_INDEX_LIST
131 Parameters: struct kvm_msr_list (in/out)
132 Returns: 0 on success; -1 on error
134 E2BIG: the msr index list is to be to fit in the array specified by
137 struct kvm_msr_list {
138 __u32 nmsrs; /* number of msrs in entries */
142 This ioctl returns the guest msrs that are supported. The list varies
143 by kvm version and host processor, but does not change otherwise. The
144 user fills in the size of the indices array in nmsrs, and in return
145 kvm adjusts nmsrs to reflect the actual number of msrs and fills in
146 the indices array with their numbers.
148 Note: if kvm indicates supports MCE (KVM_CAP_MCE), then the MCE bank MSRs are
149 not returned in the MSR list, as different vcpus can have a different number
150 of banks, as set via the KVM_X86_SETUP_MCE ioctl.
153 4.4 KVM_CHECK_EXTENSION
155 Capability: basic, KVM_CAP_CHECK_EXTENSION_VM for vm ioctl
157 Type: system ioctl, vm ioctl
158 Parameters: extension identifier (KVM_CAP_*)
159 Returns: 0 if unsupported; 1 (or some other positive integer) if supported
161 The API allows the application to query about extensions to the core
162 kvm API. Userspace passes an extension identifier (an integer) and
163 receives an integer that describes the extension availability.
164 Generally 0 means no and 1 means yes, but some extensions may report
165 additional information in the integer return value.
167 Based on their initialization different VMs may have different capabilities.
168 It is thus encouraged to use the vm ioctl to query for capabilities (available
169 with KVM_CAP_CHECK_EXTENSION_VM on the vm fd)
171 4.5 KVM_GET_VCPU_MMAP_SIZE
177 Returns: size of vcpu mmap area, in bytes
179 The KVM_RUN ioctl (cf.) communicates with userspace via a shared
180 memory region. This ioctl returns the size of that region. See the
181 KVM_RUN documentation for details.
184 4.6 KVM_SET_MEMORY_REGION
189 Parameters: struct kvm_memory_region (in)
190 Returns: 0 on success, -1 on error
192 This ioctl is obsolete and has been removed.
200 Parameters: vcpu id (apic id on x86)
201 Returns: vcpu fd on success, -1 on error
203 This API adds a vcpu to a virtual machine. No more than max_vcpus may be added.
204 The vcpu id is an integer in the range [0, max_vcpu_id).
206 The recommended max_vcpus value can be retrieved using the KVM_CAP_NR_VCPUS of
207 the KVM_CHECK_EXTENSION ioctl() at run-time.
208 The maximum possible value for max_vcpus can be retrieved using the
209 KVM_CAP_MAX_VCPUS of the KVM_CHECK_EXTENSION ioctl() at run-time.
211 If the KVM_CAP_NR_VCPUS does not exist, you should assume that max_vcpus is 4
213 If the KVM_CAP_MAX_VCPUS does not exist, you should assume that max_vcpus is
214 same as the value returned from KVM_CAP_NR_VCPUS.
216 The maximum possible value for max_vcpu_id can be retrieved using the
217 KVM_CAP_MAX_VCPU_ID of the KVM_CHECK_EXTENSION ioctl() at run-time.
219 If the KVM_CAP_MAX_VCPU_ID does not exist, you should assume that max_vcpu_id
220 is the same as the value returned from KVM_CAP_MAX_VCPUS.
222 On powerpc using book3s_hv mode, the vcpus are mapped onto virtual
223 threads in one or more virtual CPU cores. (This is because the
224 hardware requires all the hardware threads in a CPU core to be in the
225 same partition.) The KVM_CAP_PPC_SMT capability indicates the number
226 of vcpus per virtual core (vcore). The vcore id is obtained by
227 dividing the vcpu id by the number of vcpus per vcore. The vcpus in a
228 given vcore will always be in the same physical core as each other
229 (though that might be a different physical core from time to time).
230 Userspace can control the threading (SMT) mode of the guest by its
231 allocation of vcpu ids. For example, if userspace wants
232 single-threaded guest vcpus, it should make all vcpu ids be a multiple
233 of the number of vcpus per vcore.
235 For virtual cpus that have been created with S390 user controlled virtual
236 machines, the resulting vcpu fd can be memory mapped at page offset
237 KVM_S390_SIE_PAGE_OFFSET in order to obtain a memory map of the virtual
238 cpu's hardware control block.
241 4.8 KVM_GET_DIRTY_LOG (vm ioctl)
246 Parameters: struct kvm_dirty_log (in/out)
247 Returns: 0 on success, -1 on error
249 /* for KVM_GET_DIRTY_LOG */
250 struct kvm_dirty_log {
254 void __user *dirty_bitmap; /* one bit per page */
259 Given a memory slot, return a bitmap containing any pages dirtied
260 since the last call to this ioctl. Bit 0 is the first page in the
261 memory slot. Ensure the entire structure is cleared to avoid padding
264 If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 specifies
265 the address space for which you want to return the dirty bitmap.
266 They must be less than the value that KVM_CHECK_EXTENSION returns for
267 the KVM_CAP_MULTI_ADDRESS_SPACE capability.
270 4.9 KVM_SET_MEMORY_ALIAS
275 Parameters: struct kvm_memory_alias (in)
276 Returns: 0 (success), -1 (error)
278 This ioctl is obsolete and has been removed.
287 Returns: 0 on success, -1 on error
289 EINTR: an unmasked signal is pending
291 This ioctl is used to run a guest virtual cpu. While there are no
292 explicit parameters, there is an implicit parameter block that can be
293 obtained by mmap()ing the vcpu fd at offset 0, with the size given by
294 KVM_GET_VCPU_MMAP_SIZE. The parameter block is formatted as a 'struct
295 kvm_run' (see below).
301 Architectures: all except ARM, arm64
303 Parameters: struct kvm_regs (out)
304 Returns: 0 on success, -1 on error
306 Reads the general purpose registers from the vcpu.
310 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
311 __u64 rax, rbx, rcx, rdx;
312 __u64 rsi, rdi, rsp, rbp;
313 __u64 r8, r9, r10, r11;
314 __u64 r12, r13, r14, r15;
320 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
331 Architectures: all except ARM, arm64
333 Parameters: struct kvm_regs (in)
334 Returns: 0 on success, -1 on error
336 Writes the general purpose registers into the vcpu.
338 See KVM_GET_REGS for the data structure.
344 Architectures: x86, ppc
346 Parameters: struct kvm_sregs (out)
347 Returns: 0 on success, -1 on error
349 Reads special registers from the vcpu.
353 struct kvm_segment cs, ds, es, fs, gs, ss;
354 struct kvm_segment tr, ldt;
355 struct kvm_dtable gdt, idt;
356 __u64 cr0, cr2, cr3, cr4, cr8;
359 __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64];
362 /* ppc -- see arch/powerpc/include/uapi/asm/kvm.h */
364 interrupt_bitmap is a bitmap of pending external interrupts. At most
365 one bit may be set. This interrupt has been acknowledged by the APIC
366 but not yet injected into the cpu core.
372 Architectures: x86, ppc
374 Parameters: struct kvm_sregs (in)
375 Returns: 0 on success, -1 on error
377 Writes special registers into the vcpu. See KVM_GET_SREGS for the
386 Parameters: struct kvm_translation (in/out)
387 Returns: 0 on success, -1 on error
389 Translates a virtual address according to the vcpu's current address
392 struct kvm_translation {
394 __u64 linear_address;
397 __u64 physical_address;
408 Architectures: x86, ppc, mips
410 Parameters: struct kvm_interrupt (in)
411 Returns: 0 on success, negative on failure.
413 Queues a hardware interrupt vector to be injected.
415 /* for KVM_INTERRUPT */
416 struct kvm_interrupt {
423 Returns: 0 on success,
424 -EEXIST if an interrupt is already enqueued
425 -EINVAL the the irq number is invalid
426 -ENXIO if the PIC is in the kernel
427 -EFAULT if the pointer is invalid
429 Note 'irq' is an interrupt vector, not an interrupt pin or line. This
430 ioctl is useful if the in-kernel PIC is not used.
434 Queues an external interrupt to be injected. This ioctl is overleaded
435 with 3 different irq values:
439 This injects an edge type external interrupt into the guest once it's ready
440 to receive interrupts. When injected, the interrupt is done.
442 b) KVM_INTERRUPT_UNSET
444 This unsets any pending interrupt.
446 Only available with KVM_CAP_PPC_UNSET_IRQ.
448 c) KVM_INTERRUPT_SET_LEVEL
450 This injects a level type external interrupt into the guest context. The
451 interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET
454 Only available with KVM_CAP_PPC_IRQ_LEVEL.
456 Note that any value for 'irq' other than the ones stated above is invalid
457 and incurs unexpected behavior.
461 Queues an external interrupt to be injected into the virtual CPU. A negative
462 interrupt number dequeues the interrupt.
473 Support for this has been removed. Use KVM_SET_GUEST_DEBUG instead.
481 Parameters: struct kvm_msrs (in/out)
482 Returns: 0 on success, -1 on error
484 Reads model-specific registers from the vcpu. Supported msr indices can
485 be obtained using KVM_GET_MSR_INDEX_LIST.
488 __u32 nmsrs; /* number of msrs in entries */
491 struct kvm_msr_entry entries[0];
494 struct kvm_msr_entry {
500 Application code should set the 'nmsrs' member (which indicates the
501 size of the entries array) and the 'index' member of each array entry.
502 kvm will fill in the 'data' member.
510 Parameters: struct kvm_msrs (in)
511 Returns: 0 on success, -1 on error
513 Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the
516 Application code should set the 'nmsrs' member (which indicates the
517 size of the entries array), and the 'index' and 'data' members of each
526 Parameters: struct kvm_cpuid (in)
527 Returns: 0 on success, -1 on error
529 Defines the vcpu responses to the cpuid instruction. Applications
530 should use the KVM_SET_CPUID2 ioctl if available.
533 struct kvm_cpuid_entry {
542 /* for KVM_SET_CPUID */
546 struct kvm_cpuid_entry entries[0];
550 4.21 KVM_SET_SIGNAL_MASK
555 Parameters: struct kvm_signal_mask (in)
556 Returns: 0 on success, -1 on error
558 Defines which signals are blocked during execution of KVM_RUN. This
559 signal mask temporarily overrides the threads signal mask. Any
560 unblocked signal received (except SIGKILL and SIGSTOP, which retain
561 their traditional behaviour) will cause KVM_RUN to return with -EINTR.
563 Note the signal will only be delivered if not blocked by the original
566 /* for KVM_SET_SIGNAL_MASK */
567 struct kvm_signal_mask {
578 Parameters: struct kvm_fpu (out)
579 Returns: 0 on success, -1 on error
581 Reads the floating point state from the vcpu.
583 /* for KVM_GET_FPU and KVM_SET_FPU */
588 __u8 ftwx; /* in fxsave format */
604 Parameters: struct kvm_fpu (in)
605 Returns: 0 on success, -1 on error
607 Writes the floating point state to the vcpu.
609 /* for KVM_GET_FPU and KVM_SET_FPU */
614 __u8 ftwx; /* in fxsave format */
625 4.24 KVM_CREATE_IRQCHIP
627 Capability: KVM_CAP_IRQCHIP, KVM_CAP_S390_IRQCHIP (s390)
628 Architectures: x86, ARM, arm64, s390
631 Returns: 0 on success, -1 on error
633 Creates an interrupt controller model in the kernel.
634 On x86, creates a virtual ioapic, a virtual PIC (two PICs, nested), and sets up
635 future vcpus to have a local APIC. IRQ routing for GSIs 0-15 is set to both
636 PIC and IOAPIC; GSI 16-23 only go to the IOAPIC.
637 On ARM/arm64, a GICv2 is created. Any other GIC versions require the usage of
638 KVM_CREATE_DEVICE, which also supports creating a GICv2. Using
639 KVM_CREATE_DEVICE is preferred over KVM_CREATE_IRQCHIP for GICv2.
640 On s390, a dummy irq routing table is created.
642 Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled
643 before KVM_CREATE_IRQCHIP can be used.
648 Capability: KVM_CAP_IRQCHIP
649 Architectures: x86, arm, arm64
651 Parameters: struct kvm_irq_level
652 Returns: 0 on success, -1 on error
654 Sets the level of a GSI input to the interrupt controller model in the kernel.
655 On some architectures it is required that an interrupt controller model has
656 been previously created with KVM_CREATE_IRQCHIP. Note that edge-triggered
657 interrupts require the level to be set to 1 and then back to 0.
659 On real hardware, interrupt pins can be active-low or active-high. This
660 does not matter for the level field of struct kvm_irq_level: 1 always
661 means active (asserted), 0 means inactive (deasserted).
663 x86 allows the operating system to program the interrupt polarity
664 (active-low/active-high) for level-triggered interrupts, and KVM used
665 to consider the polarity. However, due to bitrot in the handling of
666 active-low interrupts, the above convention is now valid on x86 too.
667 This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED. Userspace
668 should not present interrupts to the guest as active-low unless this
669 capability is present (or unless it is not using the in-kernel irqchip,
673 ARM/arm64 can signal an interrupt either at the CPU level, or at the
674 in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to
675 use PPIs designated for specific cpus. The irq field is interpreted
678 Â bits: | 31 ... 24 | 23 ... 16 | 15 ... 0 |
679 field: | irq_type | vcpu_index | irq_id |
681 The irq_type field has the following values:
682 - irq_type[0]: out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ
683 - irq_type[1]: in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.)
684 (the vcpu_index field is ignored)
685 - irq_type[2]: in-kernel GIC: PPI, irq_id between 16 and 31 (incl.)
687 (The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs)
689 In both cases, level is used to assert/deassert the line.
691 struct kvm_irq_level {
694 __s32 status; /* not used for KVM_IRQ_LEVEL */
696 __u32 level; /* 0 or 1 */
702 Capability: KVM_CAP_IRQCHIP
705 Parameters: struct kvm_irqchip (in/out)
706 Returns: 0 on success, -1 on error
708 Reads the state of a kernel interrupt controller created with
709 KVM_CREATE_IRQCHIP into a buffer provided by the caller.
712 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
715 char dummy[512]; /* reserving space */
716 struct kvm_pic_state pic;
717 struct kvm_ioapic_state ioapic;
724 Capability: KVM_CAP_IRQCHIP
727 Parameters: struct kvm_irqchip (in)
728 Returns: 0 on success, -1 on error
730 Sets the state of a kernel interrupt controller created with
731 KVM_CREATE_IRQCHIP from a buffer provided by the caller.
734 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
737 char dummy[512]; /* reserving space */
738 struct kvm_pic_state pic;
739 struct kvm_ioapic_state ioapic;
744 4.28 KVM_XEN_HVM_CONFIG
746 Capability: KVM_CAP_XEN_HVM
749 Parameters: struct kvm_xen_hvm_config (in)
750 Returns: 0 on success, -1 on error
752 Sets the MSR that the Xen HVM guest uses to initialize its hypercall
753 page, and provides the starting address and size of the hypercall
754 blobs in userspace. When the guest writes the MSR, kvm copies one
755 page of a blob (32- or 64-bit, depending on the vcpu mode) to guest
758 struct kvm_xen_hvm_config {
771 Capability: KVM_CAP_ADJUST_CLOCK
774 Parameters: struct kvm_clock_data (out)
775 Returns: 0 on success, -1 on error
777 Gets the current timestamp of kvmclock as seen by the current guest. In
778 conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios
781 When KVM_CAP_ADJUST_CLOCK is passed to KVM_CHECK_EXTENSION, it returns the
782 set of bits that KVM can return in struct kvm_clock_data's flag member.
784 The only flag defined now is KVM_CLOCK_TSC_STABLE. If set, the returned
785 value is the exact kvmclock value seen by all VCPUs at the instant
786 when KVM_GET_CLOCK was called. If clear, the returned value is simply
787 CLOCK_MONOTONIC plus a constant offset; the offset can be modified
788 with KVM_SET_CLOCK. KVM will try to make all VCPUs follow this clock,
789 but the exact value read by each VCPU could differ, because the host
792 struct kvm_clock_data {
793 __u64 clock; /* kvmclock current value */
801 Capability: KVM_CAP_ADJUST_CLOCK
804 Parameters: struct kvm_clock_data (in)
805 Returns: 0 on success, -1 on error
807 Sets the current timestamp of kvmclock to the value specified in its parameter.
808 In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios
811 struct kvm_clock_data {
812 __u64 clock; /* kvmclock current value */
818 4.31 KVM_GET_VCPU_EVENTS
820 Capability: KVM_CAP_VCPU_EVENTS
821 Extended by: KVM_CAP_INTR_SHADOW
824 Parameters: struct kvm_vcpu_event (out)
825 Returns: 0 on success, -1 on error
827 Gets currently pending exceptions, interrupts, and NMIs as well as related
830 struct kvm_vcpu_events {
860 Only two fields are defined in the flags field:
862 - KVM_VCPUEVENT_VALID_SHADOW may be set in the flags field to signal that
863 interrupt.shadow contains a valid state.
865 - KVM_VCPUEVENT_VALID_SMM may be set in the flags field to signal that
866 smi contains a valid state.
868 4.32 KVM_SET_VCPU_EVENTS
870 Capability: KVM_CAP_VCPU_EVENTS
871 Extended by: KVM_CAP_INTR_SHADOW
874 Parameters: struct kvm_vcpu_event (in)
875 Returns: 0 on success, -1 on error
877 Set pending exceptions, interrupts, and NMIs as well as related states of the
880 See KVM_GET_VCPU_EVENTS for the data structure.
882 Fields that may be modified asynchronously by running VCPUs can be excluded
883 from the update. These fields are nmi.pending, sipi_vector, smi.smm,
884 smi.pending. Keep the corresponding bits in the flags field cleared to
885 suppress overwriting the current in-kernel state. The bits are:
887 KVM_VCPUEVENT_VALID_NMI_PENDING - transfer nmi.pending to the kernel
888 KVM_VCPUEVENT_VALID_SIPI_VECTOR - transfer sipi_vector
889 KVM_VCPUEVENT_VALID_SMM - transfer the smi sub-struct.
891 If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in
892 the flags field to signal that interrupt.shadow contains a valid state and
893 shall be written into the VCPU.
895 KVM_VCPUEVENT_VALID_SMM can only be set if KVM_CAP_X86_SMM is available.
898 4.33 KVM_GET_DEBUGREGS
900 Capability: KVM_CAP_DEBUGREGS
903 Parameters: struct kvm_debugregs (out)
904 Returns: 0 on success, -1 on error
906 Reads debug registers from the vcpu.
908 struct kvm_debugregs {
917 4.34 KVM_SET_DEBUGREGS
919 Capability: KVM_CAP_DEBUGREGS
922 Parameters: struct kvm_debugregs (in)
923 Returns: 0 on success, -1 on error
925 Writes debug registers into the vcpu.
927 See KVM_GET_DEBUGREGS for the data structure. The flags field is unused
928 yet and must be cleared on entry.
931 4.35 KVM_SET_USER_MEMORY_REGION
933 Capability: KVM_CAP_USER_MEM
936 Parameters: struct kvm_userspace_memory_region (in)
937 Returns: 0 on success, -1 on error
939 struct kvm_userspace_memory_region {
942 __u64 guest_phys_addr;
943 __u64 memory_size; /* bytes */
944 __u64 userspace_addr; /* start of the userspace allocated memory */
947 /* for kvm_memory_region::flags */
948 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0)
949 #define KVM_MEM_READONLY (1UL << 1)
951 This ioctl allows the user to create or modify a guest physical memory
952 slot. When changing an existing slot, it may be moved in the guest
953 physical memory space, or its flags may be modified. It may not be
954 resized. Slots may not overlap in guest physical address space.
955 Bits 0-15 of "slot" specifies the slot id and this value should be
956 less than the maximum number of user memory slots supported per VM.
957 The maximum allowed slots can be queried using KVM_CAP_NR_MEMSLOTS,
958 if this capability is supported by the architecture.
960 If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of "slot"
961 specifies the address space which is being modified. They must be
962 less than the value that KVM_CHECK_EXTENSION returns for the
963 KVM_CAP_MULTI_ADDRESS_SPACE capability. Slots in separate address spaces
964 are unrelated; the restriction on overlapping slots only applies within
967 Memory for the region is taken starting at the address denoted by the
968 field userspace_addr, which must point at user addressable memory for
969 the entire memory slot size. Any object may back this memory, including
970 anonymous memory, ordinary files, and hugetlbfs.
972 It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr
973 be identical. This allows large pages in the guest to be backed by large
976 The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and
977 KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of
978 writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to
979 use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it,
980 to make a new slot read-only. In this case, writes to this memory will be
981 posted to userspace as KVM_EXIT_MMIO exits.
983 When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of
984 the memory region are automatically reflected into the guest. For example, an
985 mmap() that affects the region will be made visible immediately. Another
986 example is madvise(MADV_DROP).
988 It is recommended to use this API instead of the KVM_SET_MEMORY_REGION ioctl.
989 The KVM_SET_MEMORY_REGION does not allow fine grained control over memory
990 allocation and is deprecated.
993 4.36 KVM_SET_TSS_ADDR
995 Capability: KVM_CAP_SET_TSS_ADDR
998 Parameters: unsigned long tss_address (in)
999 Returns: 0 on success, -1 on error
1001 This ioctl defines the physical address of a three-page region in the guest
1002 physical address space. The region must be within the first 4GB of the
1003 guest physical address space and must not conflict with any memory slot
1004 or any mmio address. The guest may malfunction if it accesses this memory
1007 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
1008 because of a quirk in the virtualization implementation (see the internals
1009 documentation when it pops into existence).
1014 Capability: KVM_CAP_ENABLE_CAP, KVM_CAP_ENABLE_CAP_VM
1015 Architectures: x86 (only KVM_CAP_ENABLE_CAP_VM),
1016 mips (only KVM_CAP_ENABLE_CAP), ppc, s390
1017 Type: vcpu ioctl, vm ioctl (with KVM_CAP_ENABLE_CAP_VM)
1018 Parameters: struct kvm_enable_cap (in)
1019 Returns: 0 on success; -1 on error
1021 +Not all extensions are enabled by default. Using this ioctl the application
1022 can enable an extension, making it available to the guest.
1024 On systems that do not support this ioctl, it always fails. On systems that
1025 do support it, it only works for extensions that are supported for enablement.
1027 To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should
1030 struct kvm_enable_cap {
1034 The capability that is supposed to get enabled.
1038 A bitfield indicating future enhancements. Has to be 0 for now.
1042 Arguments for enabling a feature. If a feature needs initial values to
1043 function properly, this is the place to put them.
1048 The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl
1049 for vm-wide capabilities.
1051 4.38 KVM_GET_MP_STATE
1053 Capability: KVM_CAP_MP_STATE
1054 Architectures: x86, s390, arm, arm64
1056 Parameters: struct kvm_mp_state (out)
1057 Returns: 0 on success; -1 on error
1059 struct kvm_mp_state {
1063 Returns the vcpu's current "multiprocessing state" (though also valid on
1064 uniprocessor guests).
1066 Possible values are:
1068 - KVM_MP_STATE_RUNNABLE: the vcpu is currently running [x86,arm/arm64]
1069 - KVM_MP_STATE_UNINITIALIZED: the vcpu is an application processor (AP)
1070 which has not yet received an INIT signal [x86]
1071 - KVM_MP_STATE_INIT_RECEIVED: the vcpu has received an INIT signal, and is
1072 now ready for a SIPI [x86]
1073 - KVM_MP_STATE_HALTED: the vcpu has executed a HLT instruction and
1074 is waiting for an interrupt [x86]
1075 - KVM_MP_STATE_SIPI_RECEIVED: the vcpu has just received a SIPI (vector
1076 accessible via KVM_GET_VCPU_EVENTS) [x86]
1077 - KVM_MP_STATE_STOPPED: the vcpu is stopped [s390,arm/arm64]
1078 - KVM_MP_STATE_CHECK_STOP: the vcpu is in a special error state [s390]
1079 - KVM_MP_STATE_OPERATING: the vcpu is operating (running or halted)
1081 - KVM_MP_STATE_LOAD: the vcpu is in a special load/startup state
1084 On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an
1085 in-kernel irqchip, the multiprocessing state must be maintained by userspace on
1086 these architectures.
1090 The only states that are valid are KVM_MP_STATE_STOPPED and
1091 KVM_MP_STATE_RUNNABLE which reflect if the vcpu is paused or not.
1093 4.39 KVM_SET_MP_STATE
1095 Capability: KVM_CAP_MP_STATE
1096 Architectures: x86, s390, arm, arm64
1098 Parameters: struct kvm_mp_state (in)
1099 Returns: 0 on success; -1 on error
1101 Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for
1104 On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an
1105 in-kernel irqchip, the multiprocessing state must be maintained by userspace on
1106 these architectures.
1110 The only states that are valid are KVM_MP_STATE_STOPPED and
1111 KVM_MP_STATE_RUNNABLE which reflect if the vcpu should be paused or not.
1113 4.40 KVM_SET_IDENTITY_MAP_ADDR
1115 Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR
1118 Parameters: unsigned long identity (in)
1119 Returns: 0 on success, -1 on error
1121 This ioctl defines the physical address of a one-page region in the guest
1122 physical address space. The region must be within the first 4GB of the
1123 guest physical address space and must not conflict with any memory slot
1124 or any mmio address. The guest may malfunction if it accesses this memory
1127 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
1128 because of a quirk in the virtualization implementation (see the internals
1129 documentation when it pops into existence).
1132 4.41 KVM_SET_BOOT_CPU_ID
1134 Capability: KVM_CAP_SET_BOOT_CPU_ID
1137 Parameters: unsigned long vcpu_id
1138 Returns: 0 on success, -1 on error
1140 Define which vcpu is the Bootstrap Processor (BSP). Values are the same
1141 as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default
1147 Capability: KVM_CAP_XSAVE
1150 Parameters: struct kvm_xsave (out)
1151 Returns: 0 on success, -1 on error
1157 This ioctl would copy current vcpu's xsave struct to the userspace.
1162 Capability: KVM_CAP_XSAVE
1165 Parameters: struct kvm_xsave (in)
1166 Returns: 0 on success, -1 on error
1172 This ioctl would copy userspace's xsave struct to the kernel.
1177 Capability: KVM_CAP_XCRS
1180 Parameters: struct kvm_xcrs (out)
1181 Returns: 0 on success, -1 on error
1192 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1196 This ioctl would copy current vcpu's xcrs to the userspace.
1201 Capability: KVM_CAP_XCRS
1204 Parameters: struct kvm_xcrs (in)
1205 Returns: 0 on success, -1 on error
1216 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1220 This ioctl would set vcpu's xcr to the value userspace specified.
1223 4.46 KVM_GET_SUPPORTED_CPUID
1225 Capability: KVM_CAP_EXT_CPUID
1228 Parameters: struct kvm_cpuid2 (in/out)
1229 Returns: 0 on success, -1 on error
1234 struct kvm_cpuid_entry2 entries[0];
1237 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
1238 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
1239 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
1241 struct kvm_cpuid_entry2 {
1252 This ioctl returns x86 cpuid features which are supported by both the hardware
1253 and kvm. Userspace can use the information returned by this ioctl to
1254 construct cpuid information (for KVM_SET_CPUID2) that is consistent with
1255 hardware, kernel, and userspace capabilities, and with user requirements (for
1256 example, the user may wish to constrain cpuid to emulate older hardware,
1257 or for feature consistency across a cluster).
1259 Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure
1260 with the 'nent' field indicating the number of entries in the variable-size
1261 array 'entries'. If the number of entries is too low to describe the cpu
1262 capabilities, an error (E2BIG) is returned. If the number is too high,
1263 the 'nent' field is adjusted and an error (ENOMEM) is returned. If the
1264 number is just right, the 'nent' field is adjusted to the number of valid
1265 entries in the 'entries' array, which is then filled.
1267 The entries returned are the host cpuid as returned by the cpuid instruction,
1268 with unknown or unsupported features masked out. Some features (for example,
1269 x2apic), may not be present in the host cpu, but are exposed by kvm if it can
1270 emulate them efficiently. The fields in each entry are defined as follows:
1272 function: the eax value used to obtain the entry
1273 index: the ecx value used to obtain the entry (for entries that are
1275 flags: an OR of zero or more of the following:
1276 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
1277 if the index field is valid
1278 KVM_CPUID_FLAG_STATEFUL_FUNC:
1279 if cpuid for this function returns different values for successive
1280 invocations; there will be several entries with the same function,
1281 all with this flag set
1282 KVM_CPUID_FLAG_STATE_READ_NEXT:
1283 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
1284 the first entry to be read by a cpu
1285 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
1286 this function/index combination
1288 The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned
1289 as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC
1290 support. Instead it is reported via
1292 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER)
1294 if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the
1295 feature in userspace, then you can enable the feature for KVM_SET_CPUID2.
1298 4.47 KVM_PPC_GET_PVINFO
1300 Capability: KVM_CAP_PPC_GET_PVINFO
1303 Parameters: struct kvm_ppc_pvinfo (out)
1304 Returns: 0 on success, !0 on error
1306 struct kvm_ppc_pvinfo {
1312 This ioctl fetches PV specific information that need to be passed to the guest
1313 using the device tree or other means from vm context.
1315 The hcall array defines 4 instructions that make up a hypercall.
1317 If any additional field gets added to this structure later on, a bit for that
1318 additional piece of information will be set in the flags bitmap.
1320 The flags bitmap is defined as:
1322 /* the host supports the ePAPR idle hcall
1323 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0)
1325 4.52 KVM_SET_GSI_ROUTING
1327 Capability: KVM_CAP_IRQ_ROUTING
1328 Architectures: x86 s390 arm arm64
1330 Parameters: struct kvm_irq_routing (in)
1331 Returns: 0 on success, -1 on error
1333 Sets the GSI routing table entries, overwriting any previously set entries.
1335 On arm/arm64, GSI routing has the following limitation:
1336 - GSI routing does not apply to KVM_IRQ_LINE but only to KVM_IRQFD.
1338 struct kvm_irq_routing {
1341 struct kvm_irq_routing_entry entries[0];
1344 No flags are specified so far, the corresponding field must be set to zero.
1346 struct kvm_irq_routing_entry {
1352 struct kvm_irq_routing_irqchip irqchip;
1353 struct kvm_irq_routing_msi msi;
1354 struct kvm_irq_routing_s390_adapter adapter;
1355 struct kvm_irq_routing_hv_sint hv_sint;
1360 /* gsi routing entry types */
1361 #define KVM_IRQ_ROUTING_IRQCHIP 1
1362 #define KVM_IRQ_ROUTING_MSI 2
1363 #define KVM_IRQ_ROUTING_S390_ADAPTER 3
1364 #define KVM_IRQ_ROUTING_HV_SINT 4
1367 - KVM_MSI_VALID_DEVID: used along with KVM_IRQ_ROUTING_MSI routing entry
1368 type, specifies that the devid field contains a valid value. The per-VM
1369 KVM_CAP_MSI_DEVID capability advertises the requirement to provide
1370 the device ID. If this capability is not available, userspace should
1371 never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail.
1374 struct kvm_irq_routing_irqchip {
1379 struct kvm_irq_routing_msi {
1389 If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier
1390 for the device that wrote the MSI message. For PCI, this is usually a
1391 BFD identifier in the lower 16 bits.
1393 On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS
1394 feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled,
1395 address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of
1396 address_hi must be zero.
1398 struct kvm_irq_routing_s390_adapter {
1402 __u32 summary_offset;
1406 struct kvm_irq_routing_hv_sint {
1412 4.55 KVM_SET_TSC_KHZ
1414 Capability: KVM_CAP_TSC_CONTROL
1417 Parameters: virtual tsc_khz
1418 Returns: 0 on success, -1 on error
1420 Specifies the tsc frequency for the virtual machine. The unit of the
1424 4.56 KVM_GET_TSC_KHZ
1426 Capability: KVM_CAP_GET_TSC_KHZ
1430 Returns: virtual tsc-khz on success, negative value on error
1432 Returns the tsc frequency of the guest. The unit of the return value is
1433 KHz. If the host has unstable tsc this ioctl returns -EIO instead as an
1439 Capability: KVM_CAP_IRQCHIP
1442 Parameters: struct kvm_lapic_state (out)
1443 Returns: 0 on success, -1 on error
1445 #define KVM_APIC_REG_SIZE 0x400
1446 struct kvm_lapic_state {
1447 char regs[KVM_APIC_REG_SIZE];
1450 Reads the Local APIC registers and copies them into the input argument. The
1451 data format and layout are the same as documented in the architecture manual.
1453 If KVM_X2APIC_API_USE_32BIT_IDS feature of KVM_CAP_X2APIC_API is
1454 enabled, then the format of APIC_ID register depends on the APIC mode
1455 (reported by MSR_IA32_APICBASE) of its VCPU. x2APIC stores APIC ID in
1456 the APIC_ID register (bytes 32-35). xAPIC only allows an 8-bit APIC ID
1457 which is stored in bits 31-24 of the APIC register, or equivalently in
1458 byte 35 of struct kvm_lapic_state's regs field. KVM_GET_LAPIC must then
1459 be called after MSR_IA32_APICBASE has been set with KVM_SET_MSR.
1461 If KVM_X2APIC_API_USE_32BIT_IDS feature is disabled, struct kvm_lapic_state
1462 always uses xAPIC format.
1467 Capability: KVM_CAP_IRQCHIP
1470 Parameters: struct kvm_lapic_state (in)
1471 Returns: 0 on success, -1 on error
1473 #define KVM_APIC_REG_SIZE 0x400
1474 struct kvm_lapic_state {
1475 char regs[KVM_APIC_REG_SIZE];
1478 Copies the input argument into the Local APIC registers. The data format
1479 and layout are the same as documented in the architecture manual.
1481 The format of the APIC ID register (bytes 32-35 of struct kvm_lapic_state's
1482 regs field) depends on the state of the KVM_CAP_X2APIC_API capability.
1483 See the note in KVM_GET_LAPIC.
1488 Capability: KVM_CAP_IOEVENTFD
1491 Parameters: struct kvm_ioeventfd (in)
1492 Returns: 0 on success, !0 on error
1494 This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address
1495 within the guest. A guest write in the registered address will signal the
1496 provided event instead of triggering an exit.
1498 struct kvm_ioeventfd {
1500 __u64 addr; /* legal pio/mmio address */
1501 __u32 len; /* 0, 1, 2, 4, or 8 bytes */
1507 For the special case of virtio-ccw devices on s390, the ioevent is matched
1508 to a subchannel/virtqueue tuple instead.
1510 The following flags are defined:
1512 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch)
1513 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio)
1514 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign)
1515 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \
1516 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify)
1518 If datamatch flag is set, the event will be signaled only if the written value
1519 to the registered address is equal to datamatch in struct kvm_ioeventfd.
1521 For virtio-ccw devices, addr contains the subchannel id and datamatch the
1524 With KVM_CAP_IOEVENTFD_ANY_LENGTH, a zero length ioeventfd is allowed, and
1525 the kernel will ignore the length of guest write and may get a faster vmexit.
1526 The speedup may only apply to specific architectures, but the ioeventfd will
1531 Capability: KVM_CAP_SW_TLB
1534 Parameters: struct kvm_dirty_tlb (in)
1535 Returns: 0 on success, -1 on error
1537 struct kvm_dirty_tlb {
1542 This must be called whenever userspace has changed an entry in the shared
1543 TLB, prior to calling KVM_RUN on the associated vcpu.
1545 The "bitmap" field is the userspace address of an array. This array
1546 consists of a number of bits, equal to the total number of TLB entries as
1547 determined by the last successful call to KVM_CONFIG_TLB, rounded up to the
1548 nearest multiple of 64.
1550 Each bit corresponds to one TLB entry, ordered the same as in the shared TLB
1553 The array is little-endian: the bit 0 is the least significant bit of the
1554 first byte, bit 8 is the least significant bit of the second byte, etc.
1555 This avoids any complications with differing word sizes.
1557 The "num_dirty" field is a performance hint for KVM to determine whether it
1558 should skip processing the bitmap and just invalidate everything. It must
1559 be set to the number of set bits in the bitmap.
1562 4.62 KVM_CREATE_SPAPR_TCE
1564 Capability: KVM_CAP_SPAPR_TCE
1565 Architectures: powerpc
1567 Parameters: struct kvm_create_spapr_tce (in)
1568 Returns: file descriptor for manipulating the created TCE table
1570 This creates a virtual TCE (translation control entry) table, which
1571 is an IOMMU for PAPR-style virtual I/O. It is used to translate
1572 logical addresses used in virtual I/O into guest physical addresses,
1573 and provides a scatter/gather capability for PAPR virtual I/O.
1575 /* for KVM_CAP_SPAPR_TCE */
1576 struct kvm_create_spapr_tce {
1581 The liobn field gives the logical IO bus number for which to create a
1582 TCE table. The window_size field specifies the size of the DMA window
1583 which this TCE table will translate - the table will contain one 64
1584 bit TCE entry for every 4kiB of the DMA window.
1586 When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE
1587 table has been created using this ioctl(), the kernel will handle it
1588 in real mode, updating the TCE table. H_PUT_TCE calls for other
1589 liobns will cause a vm exit and must be handled by userspace.
1591 The return value is a file descriptor which can be passed to mmap(2)
1592 to map the created TCE table into userspace. This lets userspace read
1593 the entries written by kernel-handled H_PUT_TCE calls, and also lets
1594 userspace update the TCE table directly which is useful in some
1598 4.63 KVM_ALLOCATE_RMA
1600 Capability: KVM_CAP_PPC_RMA
1601 Architectures: powerpc
1603 Parameters: struct kvm_allocate_rma (out)
1604 Returns: file descriptor for mapping the allocated RMA
1606 This allocates a Real Mode Area (RMA) from the pool allocated at boot
1607 time by the kernel. An RMA is a physically-contiguous, aligned region
1608 of memory used on older POWER processors to provide the memory which
1609 will be accessed by real-mode (MMU off) accesses in a KVM guest.
1610 POWER processors support a set of sizes for the RMA that usually
1611 includes 64MB, 128MB, 256MB and some larger powers of two.
1613 /* for KVM_ALLOCATE_RMA */
1614 struct kvm_allocate_rma {
1618 The return value is a file descriptor which can be passed to mmap(2)
1619 to map the allocated RMA into userspace. The mapped area can then be
1620 passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the
1621 RMA for a virtual machine. The size of the RMA in bytes (which is
1622 fixed at host kernel boot time) is returned in the rma_size field of
1623 the argument structure.
1625 The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl
1626 is supported; 2 if the processor requires all virtual machines to have
1627 an RMA, or 1 if the processor can use an RMA but doesn't require it,
1628 because it supports the Virtual RMA (VRMA) facility.
1633 Capability: KVM_CAP_USER_NMI
1637 Returns: 0 on success, -1 on error
1639 Queues an NMI on the thread's vcpu. Note this is well defined only
1640 when KVM_CREATE_IRQCHIP has not been called, since this is an interface
1641 between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP
1642 has been called, this interface is completely emulated within the kernel.
1644 To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the
1645 following algorithm:
1648 - read the local APIC's state (KVM_GET_LAPIC)
1649 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1)
1650 - if so, issue KVM_NMI
1653 Some guests configure the LINT1 NMI input to cause a panic, aiding in
1657 4.65 KVM_S390_UCAS_MAP
1659 Capability: KVM_CAP_S390_UCONTROL
1662 Parameters: struct kvm_s390_ucas_mapping (in)
1663 Returns: 0 in case of success
1665 The parameter is defined like this:
1666 struct kvm_s390_ucas_mapping {
1672 This ioctl maps the memory at "user_addr" with the length "length" to
1673 the vcpu's address space starting at "vcpu_addr". All parameters need to
1674 be aligned by 1 megabyte.
1677 4.66 KVM_S390_UCAS_UNMAP
1679 Capability: KVM_CAP_S390_UCONTROL
1682 Parameters: struct kvm_s390_ucas_mapping (in)
1683 Returns: 0 in case of success
1685 The parameter is defined like this:
1686 struct kvm_s390_ucas_mapping {
1692 This ioctl unmaps the memory in the vcpu's address space starting at
1693 "vcpu_addr" with the length "length". The field "user_addr" is ignored.
1694 All parameters need to be aligned by 1 megabyte.
1697 4.67 KVM_S390_VCPU_FAULT
1699 Capability: KVM_CAP_S390_UCONTROL
1702 Parameters: vcpu absolute address (in)
1703 Returns: 0 in case of success
1705 This call creates a page table entry on the virtual cpu's address space
1706 (for user controlled virtual machines) or the virtual machine's address
1707 space (for regular virtual machines). This only works for minor faults,
1708 thus it's recommended to access subject memory page via the user page
1709 table upfront. This is useful to handle validity intercepts for user
1710 controlled virtual machines to fault in the virtual cpu's lowcore pages
1711 prior to calling the KVM_RUN ioctl.
1714 4.68 KVM_SET_ONE_REG
1716 Capability: KVM_CAP_ONE_REG
1719 Parameters: struct kvm_one_reg (in)
1720 Returns: 0 on success, negative value on failure
1722 struct kvm_one_reg {
1727 Using this ioctl, a single vcpu register can be set to a specific value
1728 defined by user space with the passed in struct kvm_one_reg, where id
1729 refers to the register identifier as described below and addr is a pointer
1730 to a variable with the respective size. There can be architecture agnostic
1731 and architecture specific registers. Each have their own range of operation
1732 and their own constants and width. To keep track of the implemented
1733 registers, find a list below:
1735 Arch | Register | Width (bits)
1737 PPC | KVM_REG_PPC_HIOR | 64
1738 PPC | KVM_REG_PPC_IAC1 | 64
1739 PPC | KVM_REG_PPC_IAC2 | 64
1740 PPC | KVM_REG_PPC_IAC3 | 64
1741 PPC | KVM_REG_PPC_IAC4 | 64
1742 PPC | KVM_REG_PPC_DAC1 | 64
1743 PPC | KVM_REG_PPC_DAC2 | 64
1744 PPC | KVM_REG_PPC_DABR | 64
1745 PPC | KVM_REG_PPC_DSCR | 64
1746 PPC | KVM_REG_PPC_PURR | 64
1747 PPC | KVM_REG_PPC_SPURR | 64
1748 PPC | KVM_REG_PPC_DAR | 64
1749 PPC | KVM_REG_PPC_DSISR | 32
1750 PPC | KVM_REG_PPC_AMR | 64
1751 PPC | KVM_REG_PPC_UAMOR | 64
1752 PPC | KVM_REG_PPC_MMCR0 | 64
1753 PPC | KVM_REG_PPC_MMCR1 | 64
1754 PPC | KVM_REG_PPC_MMCRA | 64
1755 PPC | KVM_REG_PPC_MMCR2 | 64
1756 PPC | KVM_REG_PPC_MMCRS | 64
1757 PPC | KVM_REG_PPC_SIAR | 64
1758 PPC | KVM_REG_PPC_SDAR | 64
1759 PPC | KVM_REG_PPC_SIER | 64
1760 PPC | KVM_REG_PPC_PMC1 | 32
1761 PPC | KVM_REG_PPC_PMC2 | 32
1762 PPC | KVM_REG_PPC_PMC3 | 32
1763 PPC | KVM_REG_PPC_PMC4 | 32
1764 PPC | KVM_REG_PPC_PMC5 | 32
1765 PPC | KVM_REG_PPC_PMC6 | 32
1766 PPC | KVM_REG_PPC_PMC7 | 32
1767 PPC | KVM_REG_PPC_PMC8 | 32
1768 PPC | KVM_REG_PPC_FPR0 | 64
1770 PPC | KVM_REG_PPC_FPR31 | 64
1771 PPC | KVM_REG_PPC_VR0 | 128
1773 PPC | KVM_REG_PPC_VR31 | 128
1774 PPC | KVM_REG_PPC_VSR0 | 128
1776 PPC | KVM_REG_PPC_VSR31 | 128
1777 PPC | KVM_REG_PPC_FPSCR | 64
1778 PPC | KVM_REG_PPC_VSCR | 32
1779 PPC | KVM_REG_PPC_VPA_ADDR | 64
1780 PPC | KVM_REG_PPC_VPA_SLB | 128
1781 PPC | KVM_REG_PPC_VPA_DTL | 128
1782 PPC | KVM_REG_PPC_EPCR | 32
1783 PPC | KVM_REG_PPC_EPR | 32
1784 PPC | KVM_REG_PPC_TCR | 32
1785 PPC | KVM_REG_PPC_TSR | 32
1786 PPC | KVM_REG_PPC_OR_TSR | 32
1787 PPC | KVM_REG_PPC_CLEAR_TSR | 32
1788 PPC | KVM_REG_PPC_MAS0 | 32
1789 PPC | KVM_REG_PPC_MAS1 | 32
1790 PPC | KVM_REG_PPC_MAS2 | 64
1791 PPC | KVM_REG_PPC_MAS7_3 | 64
1792 PPC | KVM_REG_PPC_MAS4 | 32
1793 PPC | KVM_REG_PPC_MAS6 | 32
1794 PPC | KVM_REG_PPC_MMUCFG | 32
1795 PPC | KVM_REG_PPC_TLB0CFG | 32
1796 PPC | KVM_REG_PPC_TLB1CFG | 32
1797 PPC | KVM_REG_PPC_TLB2CFG | 32
1798 PPC | KVM_REG_PPC_TLB3CFG | 32
1799 PPC | KVM_REG_PPC_TLB0PS | 32
1800 PPC | KVM_REG_PPC_TLB1PS | 32
1801 PPC | KVM_REG_PPC_TLB2PS | 32
1802 PPC | KVM_REG_PPC_TLB3PS | 32
1803 PPC | KVM_REG_PPC_EPTCFG | 32
1804 PPC | KVM_REG_PPC_ICP_STATE | 64
1805 PPC | KVM_REG_PPC_TB_OFFSET | 64
1806 PPC | KVM_REG_PPC_SPMC1 | 32
1807 PPC | KVM_REG_PPC_SPMC2 | 32
1808 PPC | KVM_REG_PPC_IAMR | 64
1809 PPC | KVM_REG_PPC_TFHAR | 64
1810 PPC | KVM_REG_PPC_TFIAR | 64
1811 PPC | KVM_REG_PPC_TEXASR | 64
1812 PPC | KVM_REG_PPC_FSCR | 64
1813 PPC | KVM_REG_PPC_PSPB | 32
1814 PPC | KVM_REG_PPC_EBBHR | 64
1815 PPC | KVM_REG_PPC_EBBRR | 64
1816 PPC | KVM_REG_PPC_BESCR | 64
1817 PPC | KVM_REG_PPC_TAR | 64
1818 PPC | KVM_REG_PPC_DPDES | 64
1819 PPC | KVM_REG_PPC_DAWR | 64
1820 PPC | KVM_REG_PPC_DAWRX | 64
1821 PPC | KVM_REG_PPC_CIABR | 64
1822 PPC | KVM_REG_PPC_IC | 64
1823 PPC | KVM_REG_PPC_VTB | 64
1824 PPC | KVM_REG_PPC_CSIGR | 64
1825 PPC | KVM_REG_PPC_TACR | 64
1826 PPC | KVM_REG_PPC_TCSCR | 64
1827 PPC | KVM_REG_PPC_PID | 64
1828 PPC | KVM_REG_PPC_ACOP | 64
1829 PPC | KVM_REG_PPC_VRSAVE | 32
1830 PPC | KVM_REG_PPC_LPCR | 32
1831 PPC | KVM_REG_PPC_LPCR_64 | 64
1832 PPC | KVM_REG_PPC_PPR | 64
1833 PPC | KVM_REG_PPC_ARCH_COMPAT | 32
1834 PPC | KVM_REG_PPC_DABRX | 32
1835 PPC | KVM_REG_PPC_WORT | 64
1836 PPC | KVM_REG_PPC_SPRG9 | 64
1837 PPC | KVM_REG_PPC_DBSR | 32
1838 PPC | KVM_REG_PPC_TIDR | 64
1839 PPC | KVM_REG_PPC_PSSCR | 64
1840 PPC | KVM_REG_PPC_TM_GPR0 | 64
1842 PPC | KVM_REG_PPC_TM_GPR31 | 64
1843 PPC | KVM_REG_PPC_TM_VSR0 | 128
1845 PPC | KVM_REG_PPC_TM_VSR63 | 128
1846 PPC | KVM_REG_PPC_TM_CR | 64
1847 PPC | KVM_REG_PPC_TM_LR | 64
1848 PPC | KVM_REG_PPC_TM_CTR | 64
1849 PPC | KVM_REG_PPC_TM_FPSCR | 64
1850 PPC | KVM_REG_PPC_TM_AMR | 64
1851 PPC | KVM_REG_PPC_TM_PPR | 64
1852 PPC | KVM_REG_PPC_TM_VRSAVE | 64
1853 PPC | KVM_REG_PPC_TM_VSCR | 32
1854 PPC | KVM_REG_PPC_TM_DSCR | 64
1855 PPC | KVM_REG_PPC_TM_TAR | 64
1856 PPC | KVM_REG_PPC_TM_XER | 64
1858 MIPS | KVM_REG_MIPS_R0 | 64
1860 MIPS | KVM_REG_MIPS_R31 | 64
1861 MIPS | KVM_REG_MIPS_HI | 64
1862 MIPS | KVM_REG_MIPS_LO | 64
1863 MIPS | KVM_REG_MIPS_PC | 64
1864 MIPS | KVM_REG_MIPS_CP0_INDEX | 32
1865 MIPS | KVM_REG_MIPS_CP0_ENTRYLO0 | 64
1866 MIPS | KVM_REG_MIPS_CP0_ENTRYLO1 | 64
1867 MIPS | KVM_REG_MIPS_CP0_CONTEXT | 64
1868 MIPS | KVM_REG_MIPS_CP0_CONTEXTCONFIG| 32
1869 MIPS | KVM_REG_MIPS_CP0_USERLOCAL | 64
1870 MIPS | KVM_REG_MIPS_CP0_XCONTEXTCONFIG| 64
1871 MIPS | KVM_REG_MIPS_CP0_PAGEMASK | 32
1872 MIPS | KVM_REG_MIPS_CP0_PAGEGRAIN | 32
1873 MIPS | KVM_REG_MIPS_CP0_SEGCTL0 | 64
1874 MIPS | KVM_REG_MIPS_CP0_SEGCTL1 | 64
1875 MIPS | KVM_REG_MIPS_CP0_SEGCTL2 | 64
1876 MIPS | KVM_REG_MIPS_CP0_PWBASE | 64
1877 MIPS | KVM_REG_MIPS_CP0_PWFIELD | 64
1878 MIPS | KVM_REG_MIPS_CP0_PWSIZE | 64
1879 MIPS | KVM_REG_MIPS_CP0_WIRED | 32
1880 MIPS | KVM_REG_MIPS_CP0_PWCTL | 32
1881 MIPS | KVM_REG_MIPS_CP0_HWRENA | 32
1882 MIPS | KVM_REG_MIPS_CP0_BADVADDR | 64
1883 MIPS | KVM_REG_MIPS_CP0_BADINSTR | 32
1884 MIPS | KVM_REG_MIPS_CP0_BADINSTRP | 32
1885 MIPS | KVM_REG_MIPS_CP0_COUNT | 32
1886 MIPS | KVM_REG_MIPS_CP0_ENTRYHI | 64
1887 MIPS | KVM_REG_MIPS_CP0_COMPARE | 32
1888 MIPS | KVM_REG_MIPS_CP0_STATUS | 32
1889 MIPS | KVM_REG_MIPS_CP0_INTCTL | 32
1890 MIPS | KVM_REG_MIPS_CP0_CAUSE | 32
1891 MIPS | KVM_REG_MIPS_CP0_EPC | 64
1892 MIPS | KVM_REG_MIPS_CP0_PRID | 32
1893 MIPS | KVM_REG_MIPS_CP0_EBASE | 64
1894 MIPS | KVM_REG_MIPS_CP0_CONFIG | 32
1895 MIPS | KVM_REG_MIPS_CP0_CONFIG1 | 32
1896 MIPS | KVM_REG_MIPS_CP0_CONFIG2 | 32
1897 MIPS | KVM_REG_MIPS_CP0_CONFIG3 | 32
1898 MIPS | KVM_REG_MIPS_CP0_CONFIG4 | 32
1899 MIPS | KVM_REG_MIPS_CP0_CONFIG5 | 32
1900 MIPS | KVM_REG_MIPS_CP0_CONFIG7 | 32
1901 MIPS | KVM_REG_MIPS_CP0_XCONTEXT | 64
1902 MIPS | KVM_REG_MIPS_CP0_ERROREPC | 64
1903 MIPS | KVM_REG_MIPS_CP0_KSCRATCH1 | 64
1904 MIPS | KVM_REG_MIPS_CP0_KSCRATCH2 | 64
1905 MIPS | KVM_REG_MIPS_CP0_KSCRATCH3 | 64
1906 MIPS | KVM_REG_MIPS_CP0_KSCRATCH4 | 64
1907 MIPS | KVM_REG_MIPS_CP0_KSCRATCH5 | 64
1908 MIPS | KVM_REG_MIPS_CP0_KSCRATCH6 | 64
1909 MIPS | KVM_REG_MIPS_CP0_MAAR(0..63) | 64
1910 MIPS | KVM_REG_MIPS_COUNT_CTL | 64
1911 MIPS | KVM_REG_MIPS_COUNT_RESUME | 64
1912 MIPS | KVM_REG_MIPS_COUNT_HZ | 64
1913 MIPS | KVM_REG_MIPS_FPR_32(0..31) | 32
1914 MIPS | KVM_REG_MIPS_FPR_64(0..31) | 64
1915 MIPS | KVM_REG_MIPS_VEC_128(0..31) | 128
1916 MIPS | KVM_REG_MIPS_FCR_IR | 32
1917 MIPS | KVM_REG_MIPS_FCR_CSR | 32
1918 MIPS | KVM_REG_MIPS_MSA_IR | 32
1919 MIPS | KVM_REG_MIPS_MSA_CSR | 32
1921 ARM registers are mapped using the lower 32 bits. The upper 16 of that
1922 is the register group type, or coprocessor number:
1924 ARM core registers have the following id bit patterns:
1925 0x4020 0000 0010 <index into the kvm_regs struct:16>
1927 ARM 32-bit CP15 registers have the following id bit patterns:
1928 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3>
1930 ARM 64-bit CP15 registers have the following id bit patterns:
1931 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3>
1933 ARM CCSIDR registers are demultiplexed by CSSELR value:
1934 0x4020 0000 0011 00 <csselr:8>
1936 ARM 32-bit VFP control registers have the following id bit patterns:
1937 0x4020 0000 0012 1 <regno:12>
1939 ARM 64-bit FP registers have the following id bit patterns:
1940 0x4030 0000 0012 0 <regno:12>
1943 arm64 registers are mapped using the lower 32 bits. The upper 16 of
1944 that is the register group type, or coprocessor number:
1946 arm64 core/FP-SIMD registers have the following id bit patterns. Note
1947 that the size of the access is variable, as the kvm_regs structure
1948 contains elements ranging from 32 to 128 bits. The index is a 32bit
1949 value in the kvm_regs structure seen as a 32bit array.
1950 0x60x0 0000 0010 <index into the kvm_regs struct:16>
1952 arm64 CCSIDR registers are demultiplexed by CSSELR value:
1953 0x6020 0000 0011 00 <csselr:8>
1955 arm64 system registers have the following id bit patterns:
1956 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3>
1959 MIPS registers are mapped using the lower 32 bits. The upper 16 of that is
1960 the register group type:
1962 MIPS core registers (see above) have the following id bit patterns:
1963 0x7030 0000 0000 <reg:16>
1965 MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit
1966 patterns depending on whether they're 32-bit or 64-bit registers:
1967 0x7020 0000 0001 00 <reg:5> <sel:3> (32-bit)
1968 0x7030 0000 0001 00 <reg:5> <sel:3> (64-bit)
1970 Note: KVM_REG_MIPS_CP0_ENTRYLO0 and KVM_REG_MIPS_CP0_ENTRYLO1 are the MIPS64
1971 versions of the EntryLo registers regardless of the word size of the host
1972 hardware, host kernel, guest, and whether XPA is present in the guest, i.e.
1973 with the RI and XI bits (if they exist) in bits 63 and 62 respectively, and
1974 the PFNX field starting at bit 30.
1976 MIPS MAARs (see KVM_REG_MIPS_CP0_MAAR(*) above) have the following id bit
1978 0x7030 0000 0001 01 <reg:8>
1980 MIPS KVM control registers (see above) have the following id bit patterns:
1981 0x7030 0000 0002 <reg:16>
1983 MIPS FPU registers (see KVM_REG_MIPS_FPR_{32,64}() above) have the following
1984 id bit patterns depending on the size of the register being accessed. They are
1985 always accessed according to the current guest FPU mode (Status.FR and
1986 Config5.FRE), i.e. as the guest would see them, and they become unpredictable
1987 if the guest FPU mode is changed. MIPS SIMD Architecture (MSA) vector
1988 registers (see KVM_REG_MIPS_VEC_128() above) have similar patterns as they
1989 overlap the FPU registers:
1990 0x7020 0000 0003 00 <0:3> <reg:5> (32-bit FPU registers)
1991 0x7030 0000 0003 00 <0:3> <reg:5> (64-bit FPU registers)
1992 0x7040 0000 0003 00 <0:3> <reg:5> (128-bit MSA vector registers)
1994 MIPS FPU control registers (see KVM_REG_MIPS_FCR_{IR,CSR} above) have the
1995 following id bit patterns:
1996 0x7020 0000 0003 01 <0:3> <reg:5>
1998 MIPS MSA control registers (see KVM_REG_MIPS_MSA_{IR,CSR} above) have the
1999 following id bit patterns:
2000 0x7020 0000 0003 02 <0:3> <reg:5>
2003 4.69 KVM_GET_ONE_REG
2005 Capability: KVM_CAP_ONE_REG
2008 Parameters: struct kvm_one_reg (in and out)
2009 Returns: 0 on success, negative value on failure
2011 This ioctl allows to receive the value of a single register implemented
2012 in a vcpu. The register to read is indicated by the "id" field of the
2013 kvm_one_reg struct passed in. On success, the register value can be found
2014 at the memory location pointed to by "addr".
2016 The list of registers accessible using this interface is identical to the
2020 4.70 KVM_KVMCLOCK_CTRL
2022 Capability: KVM_CAP_KVMCLOCK_CTRL
2023 Architectures: Any that implement pvclocks (currently x86 only)
2026 Returns: 0 on success, -1 on error
2028 This signals to the host kernel that the specified guest is being paused by
2029 userspace. The host will set a flag in the pvclock structure that is checked
2030 from the soft lockup watchdog. The flag is part of the pvclock structure that
2031 is shared between guest and host, specifically the second bit of the flags
2032 field of the pvclock_vcpu_time_info structure. It will be set exclusively by
2033 the host and read/cleared exclusively by the guest. The guest operation of
2034 checking and clearing the flag must an atomic operation so
2035 load-link/store-conditional, or equivalent must be used. There are two cases
2036 where the guest will clear the flag: when the soft lockup watchdog timer resets
2037 itself or when a soft lockup is detected. This ioctl can be called any time
2038 after pausing the vcpu, but before it is resumed.
2043 Capability: KVM_CAP_SIGNAL_MSI
2044 Architectures: x86 arm arm64
2046 Parameters: struct kvm_msi (in)
2047 Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error
2049 Directly inject a MSI message. Only valid with in-kernel irqchip that handles
2061 flags: KVM_MSI_VALID_DEVID: devid contains a valid value. The per-VM
2062 KVM_CAP_MSI_DEVID capability advertises the requirement to provide
2063 the device ID. If this capability is not available, userspace
2064 should never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail.
2066 If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier
2067 for the device that wrote the MSI message. For PCI, this is usually a
2068 BFD identifier in the lower 16 bits.
2070 On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS
2071 feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled,
2072 address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of
2073 address_hi must be zero.
2076 4.71 KVM_CREATE_PIT2
2078 Capability: KVM_CAP_PIT2
2081 Parameters: struct kvm_pit_config (in)
2082 Returns: 0 on success, -1 on error
2084 Creates an in-kernel device model for the i8254 PIT. This call is only valid
2085 after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following
2086 parameters have to be passed:
2088 struct kvm_pit_config {
2095 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */
2097 PIT timer interrupts may use a per-VM kernel thread for injection. If it
2098 exists, this thread will have a name of the following pattern:
2100 kvm-pit/<owner-process-pid>
2102 When running a guest with elevated priorities, the scheduling parameters of
2103 this thread may have to be adjusted accordingly.
2105 This IOCTL replaces the obsolete KVM_CREATE_PIT.
2110 Capability: KVM_CAP_PIT_STATE2
2113 Parameters: struct kvm_pit_state2 (out)
2114 Returns: 0 on success, -1 on error
2116 Retrieves the state of the in-kernel PIT model. Only valid after
2117 KVM_CREATE_PIT2. The state is returned in the following structure:
2119 struct kvm_pit_state2 {
2120 struct kvm_pit_channel_state channels[3];
2127 /* disable PIT in HPET legacy mode */
2128 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001
2130 This IOCTL replaces the obsolete KVM_GET_PIT.
2135 Capability: KVM_CAP_PIT_STATE2
2138 Parameters: struct kvm_pit_state2 (in)
2139 Returns: 0 on success, -1 on error
2141 Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2.
2142 See KVM_GET_PIT2 for details on struct kvm_pit_state2.
2144 This IOCTL replaces the obsolete KVM_SET_PIT.
2147 4.74 KVM_PPC_GET_SMMU_INFO
2149 Capability: KVM_CAP_PPC_GET_SMMU_INFO
2150 Architectures: powerpc
2153 Returns: 0 on success, -1 on error
2155 This populates and returns a structure describing the features of
2156 the "Server" class MMU emulation supported by KVM.
2157 This can in turn be used by userspace to generate the appropriate
2158 device-tree properties for the guest operating system.
2160 The structure contains some global information, followed by an
2161 array of supported segment page sizes:
2163 struct kvm_ppc_smmu_info {
2167 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ];
2170 The supported flags are:
2172 - KVM_PPC_PAGE_SIZES_REAL:
2173 When that flag is set, guest page sizes must "fit" the backing
2174 store page sizes. When not set, any page size in the list can
2175 be used regardless of how they are backed by userspace.
2177 - KVM_PPC_1T_SEGMENTS
2178 The emulated MMU supports 1T segments in addition to the
2181 The "slb_size" field indicates how many SLB entries are supported
2183 The "sps" array contains 8 entries indicating the supported base
2184 page sizes for a segment in increasing order. Each entry is defined
2187 struct kvm_ppc_one_seg_page_size {
2188 __u32 page_shift; /* Base page shift of segment (or 0) */
2189 __u32 slb_enc; /* SLB encoding for BookS */
2190 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ];
2193 An entry with a "page_shift" of 0 is unused. Because the array is
2194 organized in increasing order, a lookup can stop when encoutering
2197 The "slb_enc" field provides the encoding to use in the SLB for the
2198 page size. The bits are in positions such as the value can directly
2199 be OR'ed into the "vsid" argument of the slbmte instruction.
2201 The "enc" array is a list which for each of those segment base page
2202 size provides the list of supported actual page sizes (which can be
2203 only larger or equal to the base page size), along with the
2204 corresponding encoding in the hash PTE. Similarly, the array is
2205 8 entries sorted by increasing sizes and an entry with a "0" shift
2206 is an empty entry and a terminator:
2208 struct kvm_ppc_one_page_size {
2209 __u32 page_shift; /* Page shift (or 0) */
2210 __u32 pte_enc; /* Encoding in the HPTE (>>12) */
2213 The "pte_enc" field provides a value that can OR'ed into the hash
2214 PTE's RPN field (ie, it needs to be shifted left by 12 to OR it
2215 into the hash PTE second double word).
2219 Capability: KVM_CAP_IRQFD
2220 Architectures: x86 s390 arm arm64
2222 Parameters: struct kvm_irqfd (in)
2223 Returns: 0 on success, -1 on error
2225 Allows setting an eventfd to directly trigger a guest interrupt.
2226 kvm_irqfd.fd specifies the file descriptor to use as the eventfd and
2227 kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When
2228 an event is triggered on the eventfd, an interrupt is injected into
2229 the guest using the specified gsi pin. The irqfd is removed using
2230 the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd
2233 With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify
2234 mechanism allowing emulation of level-triggered, irqfd-based
2235 interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an
2236 additional eventfd in the kvm_irqfd.resamplefd field. When operating
2237 in resample mode, posting of an interrupt through kvm_irq.fd asserts
2238 the specified gsi in the irqchip. When the irqchip is resampled, such
2239 as from an EOI, the gsi is de-asserted and the user is notified via
2240 kvm_irqfd.resamplefd. It is the user's responsibility to re-queue
2241 the interrupt if the device making use of it still requires service.
2242 Note that closing the resamplefd is not sufficient to disable the
2243 irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment
2244 and need not be specified with KVM_IRQFD_FLAG_DEASSIGN.
2246 On arm/arm64, gsi routing being supported, the following can happen:
2247 - in case no routing entry is associated to this gsi, injection fails
2248 - in case the gsi is associated to an irqchip routing entry,
2249 irqchip.pin + 32 corresponds to the injected SPI ID.
2250 - in case the gsi is associated to an MSI routing entry, the MSI
2251 message and device ID are translated into an LPI (support restricted
2252 to GICv3 ITS in-kernel emulation).
2254 4.76 KVM_PPC_ALLOCATE_HTAB
2256 Capability: KVM_CAP_PPC_ALLOC_HTAB
2257 Architectures: powerpc
2259 Parameters: Pointer to u32 containing hash table order (in/out)
2260 Returns: 0 on success, -1 on error
2262 This requests the host kernel to allocate an MMU hash table for a
2263 guest using the PAPR paravirtualization interface. This only does
2264 anything if the kernel is configured to use the Book 3S HV style of
2265 virtualization. Otherwise the capability doesn't exist and the ioctl
2266 returns an ENOTTY error. The rest of this description assumes Book 3S
2269 There must be no vcpus running when this ioctl is called; if there
2270 are, it will do nothing and return an EBUSY error.
2272 The parameter is a pointer to a 32-bit unsigned integer variable
2273 containing the order (log base 2) of the desired size of the hash
2274 table, which must be between 18 and 46. On successful return from the
2275 ioctl, the value will not be changed by the kernel.
2277 If no hash table has been allocated when any vcpu is asked to run
2278 (with the KVM_RUN ioctl), the host kernel will allocate a
2279 default-sized hash table (16 MB).
2281 If this ioctl is called when a hash table has already been allocated,
2282 with a different order from the existing hash table, the existing hash
2283 table will be freed and a new one allocated. If this is ioctl is
2284 called when a hash table has already been allocated of the same order
2285 as specified, the kernel will clear out the existing hash table (zero
2286 all HPTEs). In either case, if the guest is using the virtualized
2287 real-mode area (VRMA) facility, the kernel will re-create the VMRA
2288 HPTEs on the next KVM_RUN of any vcpu.
2290 4.77 KVM_S390_INTERRUPT
2294 Type: vm ioctl, vcpu ioctl
2295 Parameters: struct kvm_s390_interrupt (in)
2296 Returns: 0 on success, -1 on error
2298 Allows to inject an interrupt to the guest. Interrupts can be floating
2299 (vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type.
2301 Interrupt parameters are passed via kvm_s390_interrupt:
2303 struct kvm_s390_interrupt {
2309 type can be one of the following:
2311 KVM_S390_SIGP_STOP (vcpu) - sigp stop; optional flags in parm
2312 KVM_S390_PROGRAM_INT (vcpu) - program check; code in parm
2313 KVM_S390_SIGP_SET_PREFIX (vcpu) - sigp set prefix; prefix address in parm
2314 KVM_S390_RESTART (vcpu) - restart
2315 KVM_S390_INT_CLOCK_COMP (vcpu) - clock comparator interrupt
2316 KVM_S390_INT_CPU_TIMER (vcpu) - CPU timer interrupt
2317 KVM_S390_INT_VIRTIO (vm) - virtio external interrupt; external interrupt
2318 parameters in parm and parm64
2319 KVM_S390_INT_SERVICE (vm) - sclp external interrupt; sclp parameter in parm
2320 KVM_S390_INT_EMERGENCY (vcpu) - sigp emergency; source cpu in parm
2321 KVM_S390_INT_EXTERNAL_CALL (vcpu) - sigp external call; source cpu in parm
2322 KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) - compound value to indicate an
2323 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel);
2324 I/O interruption parameters in parm (subchannel) and parm64 (intparm,
2325 interruption subclass)
2326 KVM_S390_MCHK (vm, vcpu) - machine check interrupt; cr 14 bits in parm,
2327 machine check interrupt code in parm64 (note that
2328 machine checks needing further payload are not
2329 supported by this ioctl)
2331 Note that the vcpu ioctl is asynchronous to vcpu execution.
2333 4.78 KVM_PPC_GET_HTAB_FD
2335 Capability: KVM_CAP_PPC_HTAB_FD
2336 Architectures: powerpc
2338 Parameters: Pointer to struct kvm_get_htab_fd (in)
2339 Returns: file descriptor number (>= 0) on success, -1 on error
2341 This returns a file descriptor that can be used either to read out the
2342 entries in the guest's hashed page table (HPT), or to write entries to
2343 initialize the HPT. The returned fd can only be written to if the
2344 KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and
2345 can only be read if that bit is clear. The argument struct looks like
2348 /* For KVM_PPC_GET_HTAB_FD */
2349 struct kvm_get_htab_fd {
2355 /* Values for kvm_get_htab_fd.flags */
2356 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1)
2357 #define KVM_GET_HTAB_WRITE ((__u64)0x2)
2359 The `start_index' field gives the index in the HPT of the entry at
2360 which to start reading. It is ignored when writing.
2362 Reads on the fd will initially supply information about all
2363 "interesting" HPT entries. Interesting entries are those with the
2364 bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise
2365 all entries. When the end of the HPT is reached, the read() will
2366 return. If read() is called again on the fd, it will start again from
2367 the beginning of the HPT, but will only return HPT entries that have
2368 changed since they were last read.
2370 Data read or written is structured as a header (8 bytes) followed by a
2371 series of valid HPT entries (16 bytes) each. The header indicates how
2372 many valid HPT entries there are and how many invalid entries follow
2373 the valid entries. The invalid entries are not represented explicitly
2374 in the stream. The header format is:
2376 struct kvm_get_htab_header {
2382 Writes to the fd create HPT entries starting at the index given in the
2383 header; first `n_valid' valid entries with contents from the data
2384 written, then `n_invalid' invalid entries, invalidating any previously
2385 valid entries found.
2387 4.79 KVM_CREATE_DEVICE
2389 Capability: KVM_CAP_DEVICE_CTRL
2391 Parameters: struct kvm_create_device (in/out)
2392 Returns: 0 on success, -1 on error
2394 ENODEV: The device type is unknown or unsupported
2395 EEXIST: Device already created, and this type of device may not
2396 be instantiated multiple times
2398 Other error conditions may be defined by individual device types or
2399 have their standard meanings.
2401 Creates an emulated device in the kernel. The file descriptor returned
2402 in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR.
2404 If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the
2405 device type is supported (not necessarily whether it can be created
2408 Individual devices should not define flags. Attributes should be used
2409 for specifying any behavior that is not implied by the device type
2412 struct kvm_create_device {
2413 __u32 type; /* in: KVM_DEV_TYPE_xxx */
2414 __u32 fd; /* out: device handle */
2415 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */
2418 4.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR
2420 Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device,
2421 KVM_CAP_VCPU_ATTRIBUTES for vcpu device
2422 Type: device ioctl, vm ioctl, vcpu ioctl
2423 Parameters: struct kvm_device_attr
2424 Returns: 0 on success, -1 on error
2426 ENXIO: The group or attribute is unknown/unsupported for this device
2427 or hardware support is missing.
2428 EPERM: The attribute cannot (currently) be accessed this way
2429 (e.g. read-only attribute, or attribute that only makes
2430 sense when the device is in a different state)
2432 Other error conditions may be defined by individual device types.
2434 Gets/sets a specified piece of device configuration and/or state. The
2435 semantics are device-specific. See individual device documentation in
2436 the "devices" directory. As with ONE_REG, the size of the data
2437 transferred is defined by the particular attribute.
2439 struct kvm_device_attr {
2440 __u32 flags; /* no flags currently defined */
2441 __u32 group; /* device-defined */
2442 __u64 attr; /* group-defined */
2443 __u64 addr; /* userspace address of attr data */
2446 4.81 KVM_HAS_DEVICE_ATTR
2448 Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device,
2449 KVM_CAP_VCPU_ATTRIBUTES for vcpu device
2450 Type: device ioctl, vm ioctl, vcpu ioctl
2451 Parameters: struct kvm_device_attr
2452 Returns: 0 on success, -1 on error
2454 ENXIO: The group or attribute is unknown/unsupported for this device
2455 or hardware support is missing.
2457 Tests whether a device supports a particular attribute. A successful
2458 return indicates the attribute is implemented. It does not necessarily
2459 indicate that the attribute can be read or written in the device's
2460 current state. "addr" is ignored.
2462 4.82 KVM_ARM_VCPU_INIT
2465 Architectures: arm, arm64
2467 Parameters: struct kvm_vcpu_init (in)
2468 Returns: 0 on success; -1 on error
2470 Â EINVAL: Â Â Â the target is unknown, or the combination of features is invalid.
2471 Â ENOENT: Â Â Â a features bit specified is unknown.
2473 This tells KVM what type of CPU to present to the guest, and what
2474 optional features it should have. Â This will cause a reset of the cpu
2475 registers to their initial values. Â If this is not called, KVM_RUN will
2476 return ENOEXEC for that vcpu.
2478 Note that because some registers reflect machine topology, all vcpus
2479 should be created before this ioctl is invoked.
2481 Userspace can call this function multiple times for a given vcpu, including
2482 after the vcpu has been run. This will reset the vcpu to its initial
2483 state. All calls to this function after the initial call must use the same
2484 target and same set of feature flags, otherwise EINVAL will be returned.
2487 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state.
2488 Depends on KVM_CAP_ARM_PSCI. If not set, the CPU will be powered on
2489 and execute guest code when KVM_RUN is called.
2490 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode.
2491 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only).
2492 - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 for the CPU.
2493 Depends on KVM_CAP_ARM_PSCI_0_2.
2494 - KVM_ARM_VCPU_PMU_V3: Emulate PMUv3 for the CPU.
2495 Depends on KVM_CAP_ARM_PMU_V3.
2498 4.83 KVM_ARM_PREFERRED_TARGET
2501 Architectures: arm, arm64
2503 Parameters: struct struct kvm_vcpu_init (out)
2504 Returns: 0 on success; -1 on error
2506 ENODEV: no preferred target available for the host
2508 This queries KVM for preferred CPU target type which can be emulated
2509 by KVM on underlying host.
2511 The ioctl returns struct kvm_vcpu_init instance containing information
2512 about preferred CPU target type and recommended features for it. The
2513 kvm_vcpu_init->features bitmap returned will have feature bits set if
2514 the preferred target recommends setting these features, but this is
2517 The information returned by this ioctl can be used to prepare an instance
2518 of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in
2519 in VCPU matching underlying host.
2522 4.84 KVM_GET_REG_LIST
2525 Architectures: arm, arm64, mips
2527 Parameters: struct kvm_reg_list (in/out)
2528 Returns: 0 on success; -1 on error
2530 Â E2BIG: Â Â Â Â the reg index list is too big to fit in the array specified by
2531 Â Â Â Â Â Â Â Â Â Â Â Â the user (the number required will be written into n).
2533 struct kvm_reg_list {
2534 __u64 n; /* number of registers in reg[] */
2538 This ioctl returns the guest registers that are supported for the
2539 KVM_GET_ONE_REG/KVM_SET_ONE_REG calls.
2542 4.85 KVM_ARM_SET_DEVICE_ADDR (deprecated)
2544 Capability: KVM_CAP_ARM_SET_DEVICE_ADDR
2545 Architectures: arm, arm64
2547 Parameters: struct kvm_arm_device_address (in)
2548 Returns: 0 on success, -1 on error
2550 ENODEV: The device id is unknown
2551 ENXIO: Device not supported on current system
2552 EEXIST: Address already set
2553 E2BIG: Address outside guest physical address space
2554 EBUSY: Address overlaps with other device range
2556 struct kvm_arm_device_addr {
2561 Specify a device address in the guest's physical address space where guests
2562 can access emulated or directly exposed devices, which the host kernel needs
2563 to know about. The id field is an architecture specific identifier for a
2566 ARM/arm64 divides the id field into two parts, a device id and an
2567 address type id specific to the individual device.
2569 Â bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 |
2570 field: | 0x00000000 | device id | addr type id |
2572 ARM/arm64 currently only require this when using the in-kernel GIC
2573 support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2
2574 as the device id. When setting the base address for the guest's
2575 mapping of the VGIC virtual CPU and distributor interface, the ioctl
2576 must be called after calling KVM_CREATE_IRQCHIP, but before calling
2577 KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the
2578 base addresses will return -EEXIST.
2580 Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API
2581 should be used instead.
2584 4.86 KVM_PPC_RTAS_DEFINE_TOKEN
2586 Capability: KVM_CAP_PPC_RTAS
2589 Parameters: struct kvm_rtas_token_args
2590 Returns: 0 on success, -1 on error
2592 Defines a token value for a RTAS (Run Time Abstraction Services)
2593 service in order to allow it to be handled in the kernel. The
2594 argument struct gives the name of the service, which must be the name
2595 of a service that has a kernel-side implementation. If the token
2596 value is non-zero, it will be associated with that service, and
2597 subsequent RTAS calls by the guest specifying that token will be
2598 handled by the kernel. If the token value is 0, then any token
2599 associated with the service will be forgotten, and subsequent RTAS
2600 calls by the guest for that service will be passed to userspace to be
2603 4.87 KVM_SET_GUEST_DEBUG
2605 Capability: KVM_CAP_SET_GUEST_DEBUG
2606 Architectures: x86, s390, ppc, arm64
2608 Parameters: struct kvm_guest_debug (in)
2609 Returns: 0 on success; -1 on error
2611 struct kvm_guest_debug {
2614 struct kvm_guest_debug_arch arch;
2617 Set up the processor specific debug registers and configure vcpu for
2618 handling guest debug events. There are two parts to the structure, the
2619 first a control bitfield indicates the type of debug events to handle
2620 when running. Common control bits are:
2622 - KVM_GUESTDBG_ENABLE: guest debugging is enabled
2623 - KVM_GUESTDBG_SINGLESTEP: the next run should single-step
2625 The top 16 bits of the control field are architecture specific control
2626 flags which can include the following:
2628 - KVM_GUESTDBG_USE_SW_BP: using software breakpoints [x86, arm64]
2629 - KVM_GUESTDBG_USE_HW_BP: using hardware breakpoints [x86, s390, arm64]
2630 - KVM_GUESTDBG_INJECT_DB: inject DB type exception [x86]
2631 - KVM_GUESTDBG_INJECT_BP: inject BP type exception [x86]
2632 - KVM_GUESTDBG_EXIT_PENDING: trigger an immediate guest exit [s390]
2634 For example KVM_GUESTDBG_USE_SW_BP indicates that software breakpoints
2635 are enabled in memory so we need to ensure breakpoint exceptions are
2636 correctly trapped and the KVM run loop exits at the breakpoint and not
2637 running off into the normal guest vector. For KVM_GUESTDBG_USE_HW_BP
2638 we need to ensure the guest vCPUs architecture specific registers are
2639 updated to the correct (supplied) values.
2641 The second part of the structure is architecture specific and
2642 typically contains a set of debug registers.
2644 For arm64 the number of debug registers is implementation defined and
2645 can be determined by querying the KVM_CAP_GUEST_DEBUG_HW_BPS and
2646 KVM_CAP_GUEST_DEBUG_HW_WPS capabilities which return a positive number
2647 indicating the number of supported registers.
2649 When debug events exit the main run loop with the reason
2650 KVM_EXIT_DEBUG with the kvm_debug_exit_arch part of the kvm_run
2651 structure containing architecture specific debug information.
2653 4.88 KVM_GET_EMULATED_CPUID
2655 Capability: KVM_CAP_EXT_EMUL_CPUID
2658 Parameters: struct kvm_cpuid2 (in/out)
2659 Returns: 0 on success, -1 on error
2664 struct kvm_cpuid_entry2 entries[0];
2667 The member 'flags' is used for passing flags from userspace.
2669 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
2670 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
2671 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
2673 struct kvm_cpuid_entry2 {
2684 This ioctl returns x86 cpuid features which are emulated by
2685 kvm.Userspace can use the information returned by this ioctl to query
2686 which features are emulated by kvm instead of being present natively.
2688 Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2
2689 structure with the 'nent' field indicating the number of entries in
2690 the variable-size array 'entries'. If the number of entries is too low
2691 to describe the cpu capabilities, an error (E2BIG) is returned. If the
2692 number is too high, the 'nent' field is adjusted and an error (ENOMEM)
2693 is returned. If the number is just right, the 'nent' field is adjusted
2694 to the number of valid entries in the 'entries' array, which is then
2697 The entries returned are the set CPUID bits of the respective features
2698 which kvm emulates, as returned by the CPUID instruction, with unknown
2699 or unsupported feature bits cleared.
2701 Features like x2apic, for example, may not be present in the host cpu
2702 but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be
2703 emulated efficiently and thus not included here.
2705 The fields in each entry are defined as follows:
2707 function: the eax value used to obtain the entry
2708 index: the ecx value used to obtain the entry (for entries that are
2710 flags: an OR of zero or more of the following:
2711 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
2712 if the index field is valid
2713 KVM_CPUID_FLAG_STATEFUL_FUNC:
2714 if cpuid for this function returns different values for successive
2715 invocations; there will be several entries with the same function,
2716 all with this flag set
2717 KVM_CPUID_FLAG_STATE_READ_NEXT:
2718 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
2719 the first entry to be read by a cpu
2720 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
2721 this function/index combination
2723 4.89 KVM_S390_MEM_OP
2725 Capability: KVM_CAP_S390_MEM_OP
2728 Parameters: struct kvm_s390_mem_op (in)
2729 Returns: = 0 on success,
2730 < 0 on generic error (e.g. -EFAULT or -ENOMEM),
2731 > 0 if an exception occurred while walking the page tables
2733 Read or write data from/to the logical (virtual) memory of a VCPU.
2735 Parameters are specified via the following structure:
2737 struct kvm_s390_mem_op {
2738 __u64 gaddr; /* the guest address */
2739 __u64 flags; /* flags */
2740 __u32 size; /* amount of bytes */
2741 __u32 op; /* type of operation */
2742 __u64 buf; /* buffer in userspace */
2743 __u8 ar; /* the access register number */
2744 __u8 reserved[31]; /* should be set to 0 */
2747 The type of operation is specified in the "op" field. It is either
2748 KVM_S390_MEMOP_LOGICAL_READ for reading from logical memory space or
2749 KVM_S390_MEMOP_LOGICAL_WRITE for writing to logical memory space. The
2750 KVM_S390_MEMOP_F_CHECK_ONLY flag can be set in the "flags" field to check
2751 whether the corresponding memory access would create an access exception
2752 (without touching the data in the memory at the destination). In case an
2753 access exception occurred while walking the MMU tables of the guest, the
2754 ioctl returns a positive error number to indicate the type of exception.
2755 This exception is also raised directly at the corresponding VCPU if the
2756 flag KVM_S390_MEMOP_F_INJECT_EXCEPTION is set in the "flags" field.
2758 The start address of the memory region has to be specified in the "gaddr"
2759 field, and the length of the region in the "size" field. "buf" is the buffer
2760 supplied by the userspace application where the read data should be written
2761 to for KVM_S390_MEMOP_LOGICAL_READ, or where the data that should be written
2762 is stored for a KVM_S390_MEMOP_LOGICAL_WRITE. "buf" is unused and can be NULL
2763 when KVM_S390_MEMOP_F_CHECK_ONLY is specified. "ar" designates the access
2764 register number to be used.
2766 The "reserved" field is meant for future extensions. It is not used by
2767 KVM with the currently defined set of flags.
2769 4.90 KVM_S390_GET_SKEYS
2771 Capability: KVM_CAP_S390_SKEYS
2774 Parameters: struct kvm_s390_skeys
2775 Returns: 0 on success, KVM_S390_GET_KEYS_NONE if guest is not using storage
2776 keys, negative value on error
2778 This ioctl is used to get guest storage key values on the s390
2779 architecture. The ioctl takes parameters via the kvm_s390_skeys struct.
2781 struct kvm_s390_skeys {
2784 __u64 skeydata_addr;
2789 The start_gfn field is the number of the first guest frame whose storage keys
2792 The count field is the number of consecutive frames (starting from start_gfn)
2793 whose storage keys to get. The count field must be at least 1 and the maximum
2794 allowed value is defined as KVM_S390_SKEYS_ALLOC_MAX. Values outside this range
2795 will cause the ioctl to return -EINVAL.
2797 The skeydata_addr field is the address to a buffer large enough to hold count
2798 bytes. This buffer will be filled with storage key data by the ioctl.
2800 4.91 KVM_S390_SET_SKEYS
2802 Capability: KVM_CAP_S390_SKEYS
2805 Parameters: struct kvm_s390_skeys
2806 Returns: 0 on success, negative value on error
2808 This ioctl is used to set guest storage key values on the s390
2809 architecture. The ioctl takes parameters via the kvm_s390_skeys struct.
2810 See section on KVM_S390_GET_SKEYS for struct definition.
2812 The start_gfn field is the number of the first guest frame whose storage keys
2815 The count field is the number of consecutive frames (starting from start_gfn)
2816 whose storage keys to get. The count field must be at least 1 and the maximum
2817 allowed value is defined as KVM_S390_SKEYS_ALLOC_MAX. Values outside this range
2818 will cause the ioctl to return -EINVAL.
2820 The skeydata_addr field is the address to a buffer containing count bytes of
2821 storage keys. Each byte in the buffer will be set as the storage key for a
2822 single frame starting at start_gfn for count frames.
2824 Note: If any architecturally invalid key value is found in the given data then
2825 the ioctl will return -EINVAL.
2829 Capability: KVM_CAP_S390_INJECT_IRQ
2832 Parameters: struct kvm_s390_irq (in)
2833 Returns: 0 on success, -1 on error
2835 EINVAL: interrupt type is invalid
2836 type is KVM_S390_SIGP_STOP and flag parameter is invalid value
2837 type is KVM_S390_INT_EXTERNAL_CALL and code is bigger
2838 than the maximum of VCPUs
2839 EBUSY: type is KVM_S390_SIGP_SET_PREFIX and vcpu is not stopped
2840 type is KVM_S390_SIGP_STOP and a stop irq is already pending
2841 type is KVM_S390_INT_EXTERNAL_CALL and an external call interrupt
2844 Allows to inject an interrupt to the guest.
2846 Using struct kvm_s390_irq as a parameter allows
2847 to inject additional payload which is not
2848 possible via KVM_S390_INTERRUPT.
2850 Interrupt parameters are passed via kvm_s390_irq:
2852 struct kvm_s390_irq {
2855 struct kvm_s390_io_info io;
2856 struct kvm_s390_ext_info ext;
2857 struct kvm_s390_pgm_info pgm;
2858 struct kvm_s390_emerg_info emerg;
2859 struct kvm_s390_extcall_info extcall;
2860 struct kvm_s390_prefix_info prefix;
2861 struct kvm_s390_stop_info stop;
2862 struct kvm_s390_mchk_info mchk;
2867 type can be one of the following:
2869 KVM_S390_SIGP_STOP - sigp stop; parameter in .stop
2870 KVM_S390_PROGRAM_INT - program check; parameters in .pgm
2871 KVM_S390_SIGP_SET_PREFIX - sigp set prefix; parameters in .prefix
2872 KVM_S390_RESTART - restart; no parameters
2873 KVM_S390_INT_CLOCK_COMP - clock comparator interrupt; no parameters
2874 KVM_S390_INT_CPU_TIMER - CPU timer interrupt; no parameters
2875 KVM_S390_INT_EMERGENCY - sigp emergency; parameters in .emerg
2876 KVM_S390_INT_EXTERNAL_CALL - sigp external call; parameters in .extcall
2877 KVM_S390_MCHK - machine check interrupt; parameters in .mchk
2880 Note that the vcpu ioctl is asynchronous to vcpu execution.
2882 4.94 KVM_S390_GET_IRQ_STATE
2884 Capability: KVM_CAP_S390_IRQ_STATE
2887 Parameters: struct kvm_s390_irq_state (out)
2888 Returns: >= number of bytes copied into buffer,
2889 -EINVAL if buffer size is 0,
2890 -ENOBUFS if buffer size is too small to fit all pending interrupts,
2891 -EFAULT if the buffer address was invalid
2893 This ioctl allows userspace to retrieve the complete state of all currently
2894 pending interrupts in a single buffer. Use cases include migration
2895 and introspection. The parameter structure contains the address of a
2896 userspace buffer and its length:
2898 struct kvm_s390_irq_state {
2905 Userspace passes in the above struct and for each pending interrupt a
2906 struct kvm_s390_irq is copied to the provided buffer.
2908 If -ENOBUFS is returned the buffer provided was too small and userspace
2909 may retry with a bigger buffer.
2911 4.95 KVM_S390_SET_IRQ_STATE
2913 Capability: KVM_CAP_S390_IRQ_STATE
2916 Parameters: struct kvm_s390_irq_state (in)
2917 Returns: 0 on success,
2918 -EFAULT if the buffer address was invalid,
2919 -EINVAL for an invalid buffer length (see below),
2920 -EBUSY if there were already interrupts pending,
2921 errors occurring when actually injecting the
2922 interrupt. See KVM_S390_IRQ.
2924 This ioctl allows userspace to set the complete state of all cpu-local
2925 interrupts currently pending for the vcpu. It is intended for restoring
2926 interrupt state after a migration. The input parameter is a userspace buffer
2927 containing a struct kvm_s390_irq_state:
2929 struct kvm_s390_irq_state {
2935 The userspace memory referenced by buf contains a struct kvm_s390_irq
2936 for each interrupt to be injected into the guest.
2937 If one of the interrupts could not be injected for some reason the
2940 len must be a multiple of sizeof(struct kvm_s390_irq). It must be > 0
2941 and it must not exceed (max_vcpus + 32) * sizeof(struct kvm_s390_irq),
2942 which is the maximum number of possibly pending cpu-local interrupts.
2946 Capability: KVM_CAP_X86_SMM
2950 Returns: 0 on success, -1 on error
2952 Queues an SMI on the thread's vcpu.
2954 4.97 KVM_CAP_PPC_MULTITCE
2956 Capability: KVM_CAP_PPC_MULTITCE
2960 This capability means the kernel is capable of handling hypercalls
2961 H_PUT_TCE_INDIRECT and H_STUFF_TCE without passing those into the user
2962 space. This significantly accelerates DMA operations for PPC KVM guests.
2963 User space should expect that its handlers for these hypercalls
2964 are not going to be called if user space previously registered LIOBN
2965 in KVM (via KVM_CREATE_SPAPR_TCE or similar calls).
2967 In order to enable H_PUT_TCE_INDIRECT and H_STUFF_TCE use in the guest,
2968 user space might have to advertise it for the guest. For example,
2969 IBM pSeries (sPAPR) guest starts using them if "hcall-multi-tce" is
2970 present in the "ibm,hypertas-functions" device-tree property.
2972 The hypercalls mentioned above may or may not be processed successfully
2973 in the kernel based fast path. If they can not be handled by the kernel,
2974 they will get passed on to user space. So user space still has to have
2975 an implementation for these despite the in kernel acceleration.
2977 This capability is always enabled.
2979 4.98 KVM_CREATE_SPAPR_TCE_64
2981 Capability: KVM_CAP_SPAPR_TCE_64
2982 Architectures: powerpc
2984 Parameters: struct kvm_create_spapr_tce_64 (in)
2985 Returns: file descriptor for manipulating the created TCE table
2987 This is an extension for KVM_CAP_SPAPR_TCE which only supports 32bit
2988 windows, described in 4.62 KVM_CREATE_SPAPR_TCE
2990 This capability uses extended struct in ioctl interface:
2992 /* for KVM_CAP_SPAPR_TCE_64 */
2993 struct kvm_create_spapr_tce_64 {
2997 __u64 offset; /* in pages */
2998 __u64 size; /* in pages */
3001 The aim of extension is to support an additional bigger DMA window with
3002 a variable page size.
3003 KVM_CREATE_SPAPR_TCE_64 receives a 64bit window size, an IOMMU page shift and
3004 a bus offset of the corresponding DMA window, @size and @offset are numbers
3007 @flags are not used at the moment.
3009 The rest of functionality is identical to KVM_CREATE_SPAPR_TCE.
3011 4.99 KVM_REINJECT_CONTROL
3013 Capability: KVM_CAP_REINJECT_CONTROL
3016 Parameters: struct kvm_reinject_control (in)
3017 Returns: 0 on success,
3018 -EFAULT if struct kvm_reinject_control cannot be read,
3019 -ENXIO if KVM_CREATE_PIT or KVM_CREATE_PIT2 didn't succeed earlier.
3021 i8254 (PIT) has two modes, reinject and !reinject. The default is reinject,
3022 where KVM queues elapsed i8254 ticks and monitors completion of interrupt from
3023 vector(s) that i8254 injects. Reinject mode dequeues a tick and injects its
3024 interrupt whenever there isn't a pending interrupt from i8254.
3025 !reinject mode injects an interrupt as soon as a tick arrives.
3027 struct kvm_reinject_control {
3032 pit_reinject = 0 (!reinject mode) is recommended, unless running an old
3033 operating system that uses the PIT for timing (e.g. Linux 2.4.x).
3035 4.100 KVM_PPC_CONFIGURE_V3_MMU
3037 Capability: KVM_CAP_PPC_RADIX_MMU or KVM_CAP_PPC_HASH_MMU_V3
3040 Parameters: struct kvm_ppc_mmuv3_cfg (in)
3041 Returns: 0 on success,
3042 -EFAULT if struct kvm_ppc_mmuv3_cfg cannot be read,
3043 -EINVAL if the configuration is invalid
3045 This ioctl controls whether the guest will use radix or HPT (hashed
3046 page table) translation, and sets the pointer to the process table for
3049 struct kvm_ppc_mmuv3_cfg {
3051 __u64 process_table;
3054 There are two bits that can be set in flags; KVM_PPC_MMUV3_RADIX and
3055 KVM_PPC_MMUV3_GTSE. KVM_PPC_MMUV3_RADIX, if set, configures the guest
3056 to use radix tree translation, and if clear, to use HPT translation.
3057 KVM_PPC_MMUV3_GTSE, if set and if KVM permits it, configures the guest
3058 to be able to use the global TLB and SLB invalidation instructions;
3059 if clear, the guest may not use these instructions.
3061 The process_table field specifies the address and size of the guest
3062 process table, which is in the guest's space. This field is formatted
3063 as the second doubleword of the partition table entry, as defined in
3064 the Power ISA V3.00, Book III section 5.7.6.1.
3066 4.101 KVM_PPC_GET_RMMU_INFO
3068 Capability: KVM_CAP_PPC_RADIX_MMU
3071 Parameters: struct kvm_ppc_rmmu_info (out)
3072 Returns: 0 on success,
3073 -EFAULT if struct kvm_ppc_rmmu_info cannot be written,
3074 -EINVAL if no useful information can be returned
3076 This ioctl returns a structure containing two things: (a) a list
3077 containing supported radix tree geometries, and (b) a list that maps
3078 page sizes to put in the "AP" (actual page size) field for the tlbie
3079 (TLB invalidate entry) instruction.
3081 struct kvm_ppc_rmmu_info {
3082 struct kvm_ppc_radix_geom {
3087 __u32 ap_encodings[8];
3090 The geometries[] field gives up to 8 supported geometries for the
3091 radix page table, in terms of the log base 2 of the smallest page
3092 size, and the number of bits indexed at each level of the tree, from
3093 the PTE level up to the PGD level in that order. Any unused entries
3094 will have 0 in the page_shift field.
3096 The ap_encodings gives the supported page sizes and their AP field
3097 encodings, encoded with the AP value in the top 3 bits and the log
3098 base 2 of the page size in the bottom 6 bits.
3100 4.102 KVM_PPC_RESIZE_HPT_PREPARE
3102 Capability: KVM_CAP_SPAPR_RESIZE_HPT
3103 Architectures: powerpc
3105 Parameters: struct kvm_ppc_resize_hpt (in)
3106 Returns: 0 on successful completion,
3107 >0 if a new HPT is being prepared, the value is an estimated
3108 number of milliseconds until preparation is complete
3109 -EFAULT if struct kvm_reinject_control cannot be read,
3110 -EINVAL if the supplied shift or flags are invalid
3111 -ENOMEM if unable to allocate the new HPT
3112 -ENOSPC if there was a hash collision when moving existing
3113 HPT entries to the new HPT
3114 -EIO on other error conditions
3116 Used to implement the PAPR extension for runtime resizing of a guest's
3117 Hashed Page Table (HPT). Specifically this starts, stops or monitors
3118 the preparation of a new potential HPT for the guest, essentially
3119 implementing the H_RESIZE_HPT_PREPARE hypercall.
3121 If called with shift > 0 when there is no pending HPT for the guest,
3122 this begins preparation of a new pending HPT of size 2^(shift) bytes.
3123 It then returns a positive integer with the estimated number of
3124 milliseconds until preparation is complete.
3126 If called when there is a pending HPT whose size does not match that
3127 requested in the parameters, discards the existing pending HPT and
3128 creates a new one as above.
3130 If called when there is a pending HPT of the size requested, will:
3131 * If preparation of the pending HPT is already complete, return 0
3132 * If preparation of the pending HPT has failed, return an error
3133 code, then discard the pending HPT.
3134 * If preparation of the pending HPT is still in progress, return an
3135 estimated number of milliseconds until preparation is complete.
3137 If called with shift == 0, discards any currently pending HPT and
3138 returns 0 (i.e. cancels any in-progress preparation).
3140 flags is reserved for future expansion, currently setting any bits in
3141 flags will result in an -EINVAL.
3143 Normally this will be called repeatedly with the same parameters until
3144 it returns <= 0. The first call will initiate preparation, subsequent
3145 ones will monitor preparation until it completes or fails.
3147 struct kvm_ppc_resize_hpt {
3153 4.103 KVM_PPC_RESIZE_HPT_COMMIT
3155 Capability: KVM_CAP_SPAPR_RESIZE_HPT
3156 Architectures: powerpc
3158 Parameters: struct kvm_ppc_resize_hpt (in)
3159 Returns: 0 on successful completion,
3160 -EFAULT if struct kvm_reinject_control cannot be read,
3161 -EINVAL if the supplied shift or flags are invalid
3162 -ENXIO is there is no pending HPT, or the pending HPT doesn't
3163 have the requested size
3164 -EBUSY if the pending HPT is not fully prepared
3165 -ENOSPC if there was a hash collision when moving existing
3166 HPT entries to the new HPT
3167 -EIO on other error conditions
3169 Used to implement the PAPR extension for runtime resizing of a guest's
3170 Hashed Page Table (HPT). Specifically this requests that the guest be
3171 transferred to working with the new HPT, essentially implementing the
3172 H_RESIZE_HPT_COMMIT hypercall.
3174 This should only be called after KVM_PPC_RESIZE_HPT_PREPARE has
3175 returned 0 with the same parameters. In other cases
3176 KVM_PPC_RESIZE_HPT_COMMIT will return an error (usually -ENXIO or
3177 -EBUSY, though others may be possible if the preparation was started,
3180 This will have undefined effects on the guest if it has not already
3181 placed itself in a quiescent state where no vcpu will make MMU enabled
3184 On succsful completion, the pending HPT will become the guest's active
3185 HPT and the previous HPT will be discarded.
3187 On failure, the guest will still be operating on its previous HPT.
3189 struct kvm_ppc_resize_hpt {
3195 4.104 KVM_X86_GET_MCE_CAP_SUPPORTED
3197 Capability: KVM_CAP_MCE
3200 Parameters: u64 mce_cap (out)
3201 Returns: 0 on success, -1 on error
3203 Returns supported MCE capabilities. The u64 mce_cap parameter
3204 has the same format as the MSR_IA32_MCG_CAP register. Supported
3205 capabilities will have the corresponding bits set.
3207 4.105 KVM_X86_SETUP_MCE
3209 Capability: KVM_CAP_MCE
3212 Parameters: u64 mcg_cap (in)
3213 Returns: 0 on success,
3214 -EFAULT if u64 mcg_cap cannot be read,
3215 -EINVAL if the requested number of banks is invalid,
3216 -EINVAL if requested MCE capability is not supported.
3218 Initializes MCE support for use. The u64 mcg_cap parameter
3219 has the same format as the MSR_IA32_MCG_CAP register and
3220 specifies which capabilities should be enabled. The maximum
3221 supported number of error-reporting banks can be retrieved when
3222 checking for KVM_CAP_MCE. The supported capabilities can be
3223 retrieved with KVM_X86_GET_MCE_CAP_SUPPORTED.
3225 4.106 KVM_X86_SET_MCE
3227 Capability: KVM_CAP_MCE
3230 Parameters: struct kvm_x86_mce (in)
3231 Returns: 0 on success,
3232 -EFAULT if struct kvm_x86_mce cannot be read,
3233 -EINVAL if the bank number is invalid,
3234 -EINVAL if VAL bit is not set in status field.
3236 Inject a machine check error (MCE) into the guest. The input
3239 struct kvm_x86_mce {
3249 If the MCE being reported is an uncorrected error, KVM will
3250 inject it as an MCE exception into the guest. If the guest
3251 MCG_STATUS register reports that an MCE is in progress, KVM
3252 causes an KVM_EXIT_SHUTDOWN vmexit.
3254 Otherwise, if the MCE is a corrected error, KVM will just
3255 store it in the corresponding bank (provided this bank is
3256 not holding a previously reported uncorrected error).
3258 4.107 KVM_S390_GET_CMMA_BITS
3260 Capability: KVM_CAP_S390_CMMA_MIGRATION
3263 Parameters: struct kvm_s390_cmma_log (in, out)
3264 Returns: 0 on success, a negative value on error
3266 This ioctl is used to get the values of the CMMA bits on the s390
3267 architecture. It is meant to be used in two scenarios:
3268 - During live migration to save the CMMA values. Live migration needs
3269 to be enabled via the KVM_REQ_START_MIGRATION VM property.
3270 - To non-destructively peek at the CMMA values, with the flag
3271 KVM_S390_CMMA_PEEK set.
3273 The ioctl takes parameters via the kvm_s390_cmma_log struct. The desired
3274 values are written to a buffer whose location is indicated via the "values"
3275 member in the kvm_s390_cmma_log struct. The values in the input struct are
3276 also updated as needed.
3277 Each CMMA value takes up one byte.
3279 struct kvm_s390_cmma_log {
3290 start_gfn is the number of the first guest frame whose CMMA values are
3293 count is the length of the buffer in bytes,
3295 values points to the buffer where the result will be written to.
3297 If count is greater than KVM_S390_SKEYS_MAX, then it is considered to be
3298 KVM_S390_SKEYS_MAX. KVM_S390_SKEYS_MAX is re-used for consistency with
3301 The result is written in the buffer pointed to by the field values, and
3302 the values of the input parameter are updated as follows.
3304 Depending on the flags, different actions are performed. The only
3305 supported flag so far is KVM_S390_CMMA_PEEK.
3307 The default behaviour if KVM_S390_CMMA_PEEK is not set is:
3308 start_gfn will indicate the first page frame whose CMMA bits were dirty.
3309 It is not necessarily the same as the one passed as input, as clean pages
3312 count will indicate the number of bytes actually written in the buffer.
3313 It can (and very often will) be smaller than the input value, since the
3314 buffer is only filled until 16 bytes of clean values are found (which
3315 are then not copied in the buffer). Since a CMMA migration block needs
3316 the base address and the length, for a total of 16 bytes, we will send
3317 back some clean data if there is some dirty data afterwards, as long as
3318 the size of the clean data does not exceed the size of the header. This
3319 allows to minimize the amount of data to be saved or transferred over
3320 the network at the expense of more roundtrips to userspace. The next
3321 invocation of the ioctl will skip over all the clean values, saving
3322 potentially more than just the 16 bytes we found.
3324 If KVM_S390_CMMA_PEEK is set:
3325 the existing storage attributes are read even when not in migration
3326 mode, and no other action is performed;
3328 the output start_gfn will be equal to the input start_gfn,
3330 the output count will be equal to the input count, except if the end of
3331 memory has been reached.
3334 the field "remaining" will indicate the total number of dirty CMMA values
3335 still remaining, or 0 if KVM_S390_CMMA_PEEK is set and migration mode is
3340 values points to the userspace buffer where the result will be stored.
3342 This ioctl can fail with -ENOMEM if not enough memory can be allocated to
3343 complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if
3344 KVM_S390_CMMA_PEEK is not set but migration mode was not enabled, with
3345 -EFAULT if the userspace address is invalid or if no page table is
3346 present for the addresses (e.g. when using hugepages).
3348 4.108 KVM_S390_SET_CMMA_BITS
3350 Capability: KVM_CAP_S390_CMMA_MIGRATION
3353 Parameters: struct kvm_s390_cmma_log (in)
3354 Returns: 0 on success, a negative value on error
3356 This ioctl is used to set the values of the CMMA bits on the s390
3357 architecture. It is meant to be used during live migration to restore
3358 the CMMA values, but there are no restrictions on its use.
3359 The ioctl takes parameters via the kvm_s390_cmma_values struct.
3360 Each CMMA value takes up one byte.
3362 struct kvm_s390_cmma_log {
3373 start_gfn indicates the starting guest frame number,
3375 count indicates how many values are to be considered in the buffer,
3377 flags is not used and must be 0.
3379 mask indicates which PGSTE bits are to be considered.
3381 remaining is not used.
3383 values points to the buffer in userspace where to store the values.
3385 This ioctl can fail with -ENOMEM if not enough memory can be allocated to
3386 complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if
3387 the count field is too large (e.g. more than KVM_S390_CMMA_SIZE_MAX) or
3388 if the flags field was not 0, with -EFAULT if the userspace address is
3389 invalid, if invalid pages are written to (e.g. after the end of memory)
3390 or if no page table is present for the addresses (e.g. when using
3393 5. The kvm_run structure
3394 ------------------------
3396 Application code obtains a pointer to the kvm_run structure by
3397 mmap()ing a vcpu fd. From that point, application code can control
3398 execution by changing fields in kvm_run prior to calling the KVM_RUN
3399 ioctl, and obtain information about the reason KVM_RUN returned by
3400 looking up structure members.
3404 __u8 request_interrupt_window;
3406 Request that KVM_RUN return when it becomes possible to inject external
3407 interrupts into the guest. Useful in conjunction with KVM_INTERRUPT.
3409 __u8 immediate_exit;
3411 This field is polled once when KVM_RUN starts; if non-zero, KVM_RUN
3412 exits immediately, returning -EINTR. In the common scenario where a
3413 signal is used to "kick" a VCPU out of KVM_RUN, this field can be used
3414 to avoid usage of KVM_SET_SIGNAL_MASK, which has worse scalability.
3415 Rather than blocking the signal outside KVM_RUN, userspace can set up
3416 a signal handler that sets run->immediate_exit to a non-zero value.
3418 This field is ignored if KVM_CAP_IMMEDIATE_EXIT is not available.
3425 When KVM_RUN has returned successfully (return value 0), this informs
3426 application code why KVM_RUN has returned. Allowable values for this
3427 field are detailed below.
3429 __u8 ready_for_interrupt_injection;
3431 If request_interrupt_window has been specified, this field indicates
3432 an interrupt can be injected now with KVM_INTERRUPT.
3436 The value of the current interrupt flag. Only valid if in-kernel
3437 local APIC is not used.
3441 More architecture-specific flags detailing state of the VCPU that may
3442 affect the device's behavior. The only currently defined flag is
3443 KVM_RUN_X86_SMM, which is valid on x86 machines and is set if the
3444 VCPU is in system management mode.
3446 /* in (pre_kvm_run), out (post_kvm_run) */
3449 The value of the cr8 register. Only valid if in-kernel local APIC is
3450 not used. Both input and output.
3454 The value of the APIC BASE msr. Only valid if in-kernel local
3455 APIC is not used. Both input and output.
3458 /* KVM_EXIT_UNKNOWN */
3460 __u64 hardware_exit_reason;
3463 If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown
3464 reasons. Further architecture-specific information is available in
3465 hardware_exit_reason.
3467 /* KVM_EXIT_FAIL_ENTRY */
3469 __u64 hardware_entry_failure_reason;
3472 If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due
3473 to unknown reasons. Further architecture-specific information is
3474 available in hardware_entry_failure_reason.
3476 /* KVM_EXIT_EXCEPTION */
3486 #define KVM_EXIT_IO_IN 0
3487 #define KVM_EXIT_IO_OUT 1
3489 __u8 size; /* bytes */
3492 __u64 data_offset; /* relative to kvm_run start */
3495 If exit_reason is KVM_EXIT_IO, then the vcpu has
3496 executed a port I/O instruction which could not be satisfied by kvm.
3497 data_offset describes where the data is located (KVM_EXIT_IO_OUT) or
3498 where kvm expects application code to place the data for the next
3499 KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array.
3501 /* KVM_EXIT_DEBUG */
3503 struct kvm_debug_exit_arch arch;
3506 If the exit_reason is KVM_EXIT_DEBUG, then a vcpu is processing a debug event
3507 for which architecture specific information is returned.
3517 If exit_reason is KVM_EXIT_MMIO, then the vcpu has
3518 executed a memory-mapped I/O instruction which could not be satisfied
3519 by kvm. The 'data' member contains the written data if 'is_write' is
3520 true, and should be filled by application code otherwise.
3522 The 'data' member contains, in its first 'len' bytes, the value as it would
3523 appear if the VCPU performed a load or store of the appropriate width directly
3526 NOTE: For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_PAPR and
3527 KVM_EXIT_EPR the corresponding
3528 operations are complete (and guest state is consistent) only after userspace
3529 has re-entered the kernel with KVM_RUN. The kernel side will first finish
3530 incomplete operations and then check for pending signals. Userspace
3531 can re-enter the guest with an unmasked signal pending to complete
3534 /* KVM_EXIT_HYPERCALL */
3543 Unused. This was once used for 'hypercall to userspace'. To implement
3544 such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390).
3545 Note KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO.
3547 /* KVM_EXIT_TPR_ACCESS */
3554 To be documented (KVM_TPR_ACCESS_REPORTING).
3556 /* KVM_EXIT_S390_SIEIC */
3559 __u64 mask; /* psw upper half */
3560 __u64 addr; /* psw lower half */
3567 /* KVM_EXIT_S390_RESET */
3568 #define KVM_S390_RESET_POR 1
3569 #define KVM_S390_RESET_CLEAR 2
3570 #define KVM_S390_RESET_SUBSYSTEM 4
3571 #define KVM_S390_RESET_CPU_INIT 8
3572 #define KVM_S390_RESET_IPL 16
3573 __u64 s390_reset_flags;
3577 /* KVM_EXIT_S390_UCONTROL */
3579 __u64 trans_exc_code;
3583 s390 specific. A page fault has occurred for a user controlled virtual
3584 machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be
3585 resolved by the kernel.
3586 The program code and the translation exception code that were placed
3587 in the cpu's lowcore are presented here as defined by the z Architecture
3588 Principles of Operation Book in the Chapter for Dynamic Address Translation
3598 Deprecated - was used for 440 KVM.
3605 MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch
3606 hypercalls and exit with this exit struct that contains all the guest gprs.
3608 If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall.
3609 Userspace can now handle the hypercall and when it's done modify the gprs as
3610 necessary. Upon guest entry all guest GPRs will then be replaced by the values
3613 /* KVM_EXIT_PAPR_HCALL */
3620 This is used on 64-bit PowerPC when emulating a pSeries partition,
3621 e.g. with the 'pseries' machine type in qemu. It occurs when the
3622 guest does a hypercall using the 'sc 1' instruction. The 'nr' field
3623 contains the hypercall number (from the guest R3), and 'args' contains
3624 the arguments (from the guest R4 - R12). Userspace should put the
3625 return code in 'ret' and any extra returned values in args[].
3626 The possible hypercalls are defined in the Power Architecture Platform
3627 Requirements (PAPR) document available from www.power.org (free
3628 developer registration required to access it).
3630 /* KVM_EXIT_S390_TSCH */
3632 __u16 subchannel_id;
3633 __u16 subchannel_nr;
3640 s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled
3641 and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O
3642 interrupt for the target subchannel has been dequeued and subchannel_id,
3643 subchannel_nr, io_int_parm and io_int_word contain the parameters for that
3644 interrupt. ipb is needed for instruction parameter decoding.
3651 On FSL BookE PowerPC chips, the interrupt controller has a fast patch
3652 interrupt acknowledge path to the core. When the core successfully
3653 delivers an interrupt, it automatically populates the EPR register with
3654 the interrupt vector number and acknowledges the interrupt inside
3655 the interrupt controller.
3657 In case the interrupt controller lives in user space, we need to do
3658 the interrupt acknowledge cycle through it to fetch the next to be
3659 delivered interrupt vector using this exit.
3661 It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an
3662 external interrupt has just been delivered into the guest. User space
3663 should put the acknowledged interrupt vector into the 'epr' field.
3665 /* KVM_EXIT_SYSTEM_EVENT */
3667 #define KVM_SYSTEM_EVENT_SHUTDOWN 1
3668 #define KVM_SYSTEM_EVENT_RESET 2
3669 #define KVM_SYSTEM_EVENT_CRASH 3
3674 If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered
3675 a system-level event using some architecture specific mechanism (hypercall
3676 or some special instruction). In case of ARM/ARM64, this is triggered using
3677 HVC instruction based PSCI call from the vcpu. The 'type' field describes
3678 the system-level event type. The 'flags' field describes architecture
3679 specific flags for the system-level event.
3681 Valid values for 'type' are:
3682 KVM_SYSTEM_EVENT_SHUTDOWN -- the guest has requested a shutdown of the
3683 VM. Userspace is not obliged to honour this, and if it does honour
3684 this does not need to destroy the VM synchronously (ie it may call
3685 KVM_RUN again before shutdown finally occurs).
3686 KVM_SYSTEM_EVENT_RESET -- the guest has requested a reset of the VM.
3687 As with SHUTDOWN, userspace can choose to ignore the request, or
3688 to schedule the reset to occur in the future and may call KVM_RUN again.
3689 KVM_SYSTEM_EVENT_CRASH -- the guest crash occurred and the guest
3690 has requested a crash condition maintenance. Userspace can choose
3691 to ignore the request, or to gather VM memory core dump and/or
3692 reset/shutdown of the VM.
3694 /* KVM_EXIT_IOAPIC_EOI */
3699 Indicates that the VCPU's in-kernel local APIC received an EOI for a
3700 level-triggered IOAPIC interrupt. This exit only triggers when the
3701 IOAPIC is implemented in userspace (i.e. KVM_CAP_SPLIT_IRQCHIP is enabled);
3702 the userspace IOAPIC should process the EOI and retrigger the interrupt if
3703 it is still asserted. Vector is the LAPIC interrupt vector for which the
3706 struct kvm_hyperv_exit {
3707 #define KVM_EXIT_HYPERV_SYNIC 1
3708 #define KVM_EXIT_HYPERV_HCALL 2
3724 /* KVM_EXIT_HYPERV */
3725 struct kvm_hyperv_exit hyperv;
3726 Indicates that the VCPU exits into userspace to process some tasks
3727 related to Hyper-V emulation.
3728 Valid values for 'type' are:
3729 KVM_EXIT_HYPERV_SYNIC -- synchronously notify user-space about
3730 Hyper-V SynIC state change. Notification is used to remap SynIC
3731 event/message pages and to enable/disable SynIC messages/events processing
3734 /* Fix the size of the union. */
3739 * shared registers between kvm and userspace.
3740 * kvm_valid_regs specifies the register classes set by the host
3741 * kvm_dirty_regs specified the register classes dirtied by userspace
3742 * struct kvm_sync_regs is architecture specific, as well as the
3743 * bits for kvm_valid_regs and kvm_dirty_regs
3745 __u64 kvm_valid_regs;
3746 __u64 kvm_dirty_regs;
3748 struct kvm_sync_regs regs;
3752 If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access
3753 certain guest registers without having to call SET/GET_*REGS. Thus we can
3754 avoid some system call overhead if userspace has to handle the exit.
3755 Userspace can query the validity of the structure by checking
3756 kvm_valid_regs for specific bits. These bits are architecture specific
3757 and usually define the validity of a groups of registers. (e.g. one bit
3758 for general purpose registers)
3760 Please note that the kernel is allowed to use the kvm_run structure as the
3761 primary storage for certain register types. Therefore, the kernel may use the
3762 values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set.
3768 6. Capabilities that can be enabled on vCPUs
3769 --------------------------------------------
3771 There are certain capabilities that change the behavior of the virtual CPU or
3772 the virtual machine when enabled. To enable them, please see section 4.37.
3773 Below you can find a list of capabilities and what their effect on the vCPU or
3774 the virtual machine is when enabling them.
3776 The following information is provided along with the description:
3778 Architectures: which instruction set architectures provide this ioctl.
3779 x86 includes both i386 and x86_64.
3781 Target: whether this is a per-vcpu or per-vm capability.
3783 Parameters: what parameters are accepted by the capability.
3785 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
3786 are not detailed, but errors with specific meanings are.
3794 Returns: 0 on success; -1 on error
3796 This capability enables interception of OSI hypercalls that otherwise would
3797 be treated as normal system calls to be injected into the guest. OSI hypercalls
3798 were invented by Mac-on-Linux to have a standardized communication mechanism
3799 between the guest and the host.
3801 When this capability is enabled, KVM_EXIT_OSI can occur.
3804 6.2 KVM_CAP_PPC_PAPR
3809 Returns: 0 on success; -1 on error
3811 This capability enables interception of PAPR hypercalls. PAPR hypercalls are
3812 done using the hypercall instruction "sc 1".
3814 It also sets the guest privilege level to "supervisor" mode. Usually the guest
3815 runs in "hypervisor" privilege mode with a few missing features.
3817 In addition to the above, it changes the semantics of SDR1. In this mode, the
3818 HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the
3819 HTAB invisible to the guest.
3821 When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur.
3828 Parameters: args[0] is the address of a struct kvm_config_tlb
3829 Returns: 0 on success; -1 on error
3831 struct kvm_config_tlb {
3838 Configures the virtual CPU's TLB array, establishing a shared memory area
3839 between userspace and KVM. The "params" and "array" fields are userspace
3840 addresses of mmu-type-specific data structures. The "array_len" field is an
3841 safety mechanism, and should be set to the size in bytes of the memory that
3842 userspace has reserved for the array. It must be at least the size dictated
3843 by "mmu_type" and "params".
3845 While KVM_RUN is active, the shared region is under control of KVM. Its
3846 contents are undefined, and any modification by userspace results in
3847 boundedly undefined behavior.
3849 On return from KVM_RUN, the shared region will reflect the current state of
3850 the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB
3851 to tell KVM which entries have been changed, prior to calling KVM_RUN again
3854 For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV:
3855 - The "params" field is of type "struct kvm_book3e_206_tlb_params".
3856 - The "array" field points to an array of type "struct
3857 kvm_book3e_206_tlb_entry".
3858 - The array consists of all entries in the first TLB, followed by all
3859 entries in the second TLB.
3860 - Within a TLB, entries are ordered first by increasing set number. Within a
3861 set, entries are ordered by way (increasing ESEL).
3862 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1)
3863 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value.
3864 - The tsize field of mas1 shall be set to 4K on TLB0, even though the
3865 hardware ignores this value for TLB0.
3867 6.4 KVM_CAP_S390_CSS_SUPPORT
3872 Returns: 0 on success; -1 on error
3874 This capability enables support for handling of channel I/O instructions.
3876 TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are
3877 handled in-kernel, while the other I/O instructions are passed to userspace.
3879 When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST
3880 SUBCHANNEL intercepts.
3882 Note that even though this capability is enabled per-vcpu, the complete
3883 virtual machine is affected.
3889 Parameters: args[0] defines whether the proxy facility is active
3890 Returns: 0 on success; -1 on error
3892 This capability enables or disables the delivery of interrupts through the
3893 external proxy facility.
3895 When enabled (args[0] != 0), every time the guest gets an external interrupt
3896 delivered, it automatically exits into user space with a KVM_EXIT_EPR exit
3897 to receive the topmost interrupt vector.
3899 When disabled (args[0] == 0), behavior is as if this facility is unsupported.
3901 When this capability is enabled, KVM_EXIT_EPR can occur.
3903 6.6 KVM_CAP_IRQ_MPIC
3906 Parameters: args[0] is the MPIC device fd
3907 args[1] is the MPIC CPU number for this vcpu
3909 This capability connects the vcpu to an in-kernel MPIC device.
3911 6.7 KVM_CAP_IRQ_XICS
3915 Parameters: args[0] is the XICS device fd
3916 args[1] is the XICS CPU number (server ID) for this vcpu
3918 This capability connects the vcpu to an in-kernel XICS device.
3920 6.8 KVM_CAP_S390_IRQCHIP
3926 This capability enables the in-kernel irqchip for s390. Please refer to
3927 "4.24 KVM_CREATE_IRQCHIP" for details.
3929 6.9 KVM_CAP_MIPS_FPU
3933 Parameters: args[0] is reserved for future use (should be 0).
3935 This capability allows the use of the host Floating Point Unit by the guest. It
3936 allows the Config1.FP bit to be set to enable the FPU in the guest. Once this is
3937 done the KVM_REG_MIPS_FPR_* and KVM_REG_MIPS_FCR_* registers can be accessed
3938 (depending on the current guest FPU register mode), and the Status.FR,
3939 Config5.FRE bits are accessible via the KVM API and also from the guest,
3940 depending on them being supported by the FPU.
3942 6.10 KVM_CAP_MIPS_MSA
3946 Parameters: args[0] is reserved for future use (should be 0).
3948 This capability allows the use of the MIPS SIMD Architecture (MSA) by the guest.
3949 It allows the Config3.MSAP bit to be set to enable the use of MSA by the guest.
3950 Once this is done the KVM_REG_MIPS_VEC_* and KVM_REG_MIPS_MSA_* registers can be
3951 accessed, and the Config5.MSAEn bit is accessible via the KVM API and also from
3954 7. Capabilities that can be enabled on VMs
3955 ------------------------------------------
3957 There are certain capabilities that change the behavior of the virtual
3958 machine when enabled. To enable them, please see section 4.37. Below
3959 you can find a list of capabilities and what their effect on the VM
3960 is when enabling them.
3962 The following information is provided along with the description:
3964 Architectures: which instruction set architectures provide this ioctl.
3965 x86 includes both i386 and x86_64.
3967 Parameters: what parameters are accepted by the capability.
3969 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
3970 are not detailed, but errors with specific meanings are.
3973 7.1 KVM_CAP_PPC_ENABLE_HCALL
3976 Parameters: args[0] is the sPAPR hcall number
3977 args[1] is 0 to disable, 1 to enable in-kernel handling
3979 This capability controls whether individual sPAPR hypercalls (hcalls)
3980 get handled by the kernel or not. Enabling or disabling in-kernel
3981 handling of an hcall is effective across the VM. On creation, an
3982 initial set of hcalls are enabled for in-kernel handling, which
3983 consists of those hcalls for which in-kernel handlers were implemented
3984 before this capability was implemented. If disabled, the kernel will
3985 not to attempt to handle the hcall, but will always exit to userspace
3986 to handle it. Note that it may not make sense to enable some and
3987 disable others of a group of related hcalls, but KVM does not prevent
3988 userspace from doing that.
3990 If the hcall number specified is not one that has an in-kernel
3991 implementation, the KVM_ENABLE_CAP ioctl will fail with an EINVAL
3994 7.2 KVM_CAP_S390_USER_SIGP
3999 This capability controls which SIGP orders will be handled completely in user
4000 space. With this capability enabled, all fast orders will be handled completely
4006 - CONDITIONAL EMERGENCY SIGNAL
4008 All other orders will be handled completely in user space.
4010 Only privileged operation exceptions will be checked for in the kernel (or even
4011 in the hardware prior to interception). If this capability is not enabled, the
4012 old way of handling SIGP orders is used (partially in kernel and user space).
4014 7.3 KVM_CAP_S390_VECTOR_REGISTERS
4018 Returns: 0 on success, negative value on error
4020 Allows use of the vector registers introduced with z13 processor, and
4021 provides for the synchronization between host and user space. Will
4022 return -EINVAL if the machine does not support vectors.
4024 7.4 KVM_CAP_S390_USER_STSI
4029 This capability allows post-handlers for the STSI instruction. After
4030 initial handling in the kernel, KVM exits to user space with
4031 KVM_EXIT_S390_STSI to allow user space to insert further data.
4033 Before exiting to userspace, kvm handlers should fill in s390_stsi field of
4044 @addr - guest address of STSI SYSIB
4048 @ar - access register number
4050 KVM handlers should exit to userspace with rc = -EREMOTE.
4052 7.5 KVM_CAP_SPLIT_IRQCHIP
4055 Parameters: args[0] - number of routes reserved for userspace IOAPICs
4056 Returns: 0 on success, -1 on error
4058 Create a local apic for each processor in the kernel. This can be used
4059 instead of KVM_CREATE_IRQCHIP if the userspace VMM wishes to emulate the
4060 IOAPIC and PIC (and also the PIT, even though this has to be enabled
4063 This capability also enables in kernel routing of interrupt requests;
4064 when KVM_CAP_SPLIT_IRQCHIP only routes of KVM_IRQ_ROUTING_MSI type are
4065 used in the IRQ routing table. The first args[0] MSI routes are reserved
4066 for the IOAPIC pins. Whenever the LAPIC receives an EOI for these routes,
4067 a KVM_EXIT_IOAPIC_EOI vmexit will be reported to userspace.
4069 Fails if VCPU has already been created, or if the irqchip is already in the
4070 kernel (i.e. KVM_CREATE_IRQCHIP has already been called).
4077 Allows use of runtime-instrumentation introduced with zEC12 processor.
4078 Will return -EINVAL if the machine does not support runtime-instrumentation.
4079 Will return -EBUSY if a VCPU has already been created.
4081 7.7 KVM_CAP_X2APIC_API
4084 Parameters: args[0] - features that should be enabled
4085 Returns: 0 on success, -EINVAL when args[0] contains invalid features
4087 Valid feature flags in args[0] are
4089 #define KVM_X2APIC_API_USE_32BIT_IDS (1ULL << 0)
4090 #define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK (1ULL << 1)
4092 Enabling KVM_X2APIC_API_USE_32BIT_IDS changes the behavior of
4093 KVM_SET_GSI_ROUTING, KVM_SIGNAL_MSI, KVM_SET_LAPIC, and KVM_GET_LAPIC,
4094 allowing the use of 32-bit APIC IDs. See KVM_CAP_X2APIC_API in their
4095 respective sections.
4097 KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK must be enabled for x2APIC to work
4098 in logical mode or with more than 255 VCPUs. Otherwise, KVM treats 0xff
4099 as a broadcast even in x2APIC mode in order to support physical x2APIC
4100 without interrupt remapping. This is undesirable in logical mode,
4101 where 0xff represents CPUs 0-7 in cluster 0.
4103 7.8 KVM_CAP_S390_USER_INSTR0
4108 With this capability enabled, all illegal instructions 0x0000 (2 bytes) will
4109 be intercepted and forwarded to user space. User space can use this
4110 mechanism e.g. to realize 2-byte software breakpoints. The kernel will
4111 not inject an operating exception for these instructions, user space has
4112 to take care of that.
4114 This capability can be enabled dynamically even if VCPUs were already
4115 created and are running.
4121 Returns: 0 on success; -EINVAL if the machine does not support
4122 guarded storage; -EBUSY if a VCPU has already been created.
4124 Allows use of guarded storage for the KVM guest.
4126 7.10 KVM_CAP_S390_AIS
4131 Allow use of adapter-interruption suppression.
4132 Returns: 0 on success; -EBUSY if a VCPU has already been created.
4134 7.11 KVM_CAP_PPC_SMT
4137 Parameters: vsmt_mode, flags
4139 Enabling this capability on a VM provides userspace with a way to set
4140 the desired virtual SMT mode (i.e. the number of virtual CPUs per
4141 virtual core). The virtual SMT mode, vsmt_mode, must be a power of 2
4142 between 1 and 8. On POWER8, vsmt_mode must also be no greater than
4143 the number of threads per subcore for the host. Currently flags must
4144 be 0. A successful call to enable this capability will result in
4145 vsmt_mode being returned when the KVM_CAP_PPC_SMT capability is
4146 subsequently queried for the VM. This capability is only supported by
4147 HV KVM, and can only be set before any VCPUs have been created.
4148 The KVM_CAP_PPC_SMT_POSSIBLE capability indicates which virtual SMT
4149 modes are available.
4151 7.12 KVM_CAP_PPC_FWNMI
4156 With this capability a machine check exception in the guest address
4157 space will cause KVM to exit the guest with NMI exit reason. This
4158 enables QEMU to build error log and branch to guest kernel registered
4159 machine check handling routine. Without this capability KVM will
4160 branch to guests' 0x200 interrupt vector.
4162 8. Other capabilities.
4163 ----------------------
4165 This section lists capabilities that give information about other
4166 features of the KVM implementation.
4168 8.1 KVM_CAP_PPC_HWRNG
4172 This capability, if KVM_CHECK_EXTENSION indicates that it is
4173 available, means that that the kernel has an implementation of the
4174 H_RANDOM hypercall backed by a hardware random-number generator.
4175 If present, the kernel H_RANDOM handler can be enabled for guest use
4176 with the KVM_CAP_PPC_ENABLE_HCALL capability.
4178 8.2 KVM_CAP_HYPERV_SYNIC
4181 This capability, if KVM_CHECK_EXTENSION indicates that it is
4182 available, means that that the kernel has an implementation of the
4183 Hyper-V Synthetic interrupt controller(SynIC). Hyper-V SynIC is
4184 used to support Windows Hyper-V based guest paravirt drivers(VMBus).
4186 In order to use SynIC, it has to be activated by setting this
4187 capability via KVM_ENABLE_CAP ioctl on the vcpu fd. Note that this
4188 will disable the use of APIC hardware virtualization even if supported
4189 by the CPU, as it's incompatible with SynIC auto-EOI behavior.
4191 8.3 KVM_CAP_PPC_RADIX_MMU
4195 This capability, if KVM_CHECK_EXTENSION indicates that it is
4196 available, means that that the kernel can support guests using the
4197 radix MMU defined in Power ISA V3.00 (as implemented in the POWER9
4200 8.4 KVM_CAP_PPC_HASH_MMU_V3
4204 This capability, if KVM_CHECK_EXTENSION indicates that it is
4205 available, means that that the kernel can support guests using the
4206 hashed page table MMU defined in Power ISA V3.00 (as implemented in
4207 the POWER9 processor), including in-memory segment tables.
4213 This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that
4214 it is available, means that full hardware assisted virtualization capabilities
4215 of the hardware are available for use through KVM. An appropriate
4216 KVM_VM_MIPS_* type must be passed to KVM_CREATE_VM to create a VM which
4219 If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is
4220 available, it means that the VM is using full hardware assisted virtualization
4221 capabilities of the hardware. This is useful to check after creating a VM with
4222 KVM_VM_MIPS_DEFAULT.
4224 The value returned by KVM_CHECK_EXTENSION should be compared against known
4225 values (see below). All other values are reserved. This is to allow for the
4226 possibility of other hardware assisted virtualization implementations which
4227 may be incompatible with the MIPS VZ ASE.
4229 0: The trap & emulate implementation is in use to run guest code in user
4230 mode. Guest virtual memory segments are rearranged to fit the guest in the
4231 user mode address space.
4233 1: The MIPS VZ ASE is in use, providing full hardware assisted
4234 virtualization, including standard guest virtual memory segments.
4240 This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that
4241 it is available, means that the trap & emulate implementation is available to
4242 run guest code in user mode, even if KVM_CAP_MIPS_VZ indicates that hardware
4243 assisted virtualisation is also available. KVM_VM_MIPS_TE (0) must be passed
4244 to KVM_CREATE_VM to create a VM which utilises it.
4246 If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is
4247 available, it means that the VM is using trap & emulate.
4249 8.7 KVM_CAP_MIPS_64BIT
4253 This capability indicates the supported architecture type of the guest, i.e. the
4254 supported register and address width.
4256 The values returned when this capability is checked by KVM_CHECK_EXTENSION on a
4257 kvm VM handle correspond roughly to the CP0_Config.AT register field, and should
4258 be checked specifically against known values (see below). All other values are
4261 0: MIPS32 or microMIPS32.
4262 Both registers and addresses are 32-bits wide.
4263 It will only be possible to run 32-bit guest code.
4265 1: MIPS64 or microMIPS64 with access only to 32-bit compatibility segments.
4266 Registers are 64-bits wide, but addresses are 32-bits wide.
4267 64-bit guest code may run but cannot access MIPS64 memory segments.
4268 It will also be possible to run 32-bit guest code.
4270 2: MIPS64 or microMIPS64 with access to all address segments.
4271 Both registers and addresses are 64-bits wide.
4272 It will be possible to run 64-bit or 32-bit guest code.
4274 8.8 KVM_CAP_X86_GUEST_MWAIT
4278 This capability indicates that guest using memory monotoring instructions
4279 (MWAIT/MWAITX) to stop the virtual CPU will not cause a VM exit. As such time
4280 spent while virtual CPU is halted in this way will then be accounted for as
4281 guest running time on the host (as opposed to e.g. HLT).
4283 8.9 KVM_CAP_ARM_USER_IRQ
4285 Architectures: arm, arm64
4286 This capability, if KVM_CHECK_EXTENSION indicates that it is available, means
4287 that if userspace creates a VM without an in-kernel interrupt controller, it
4288 will be notified of changes to the output level of in-kernel emulated devices,
4289 which can generate virtual interrupts, presented to the VM.
4290 For such VMs, on every return to userspace, the kernel
4291 updates the vcpu's run->s.regs.device_irq_level field to represent the actual
4292 output level of the device.
4294 Whenever kvm detects a change in the device output level, kvm guarantees at
4295 least one return to userspace before running the VM. This exit could either
4296 be a KVM_EXIT_INTR or any other exit event, like KVM_EXIT_MMIO. This way,
4297 userspace can always sample the device output level and re-compute the state of
4298 the userspace interrupt controller. Userspace should always check the state
4299 of run->s.regs.device_irq_level on every kvm exit.
4300 The value in run->s.regs.device_irq_level can represent both level and edge
4301 triggered interrupt signals, depending on the device. Edge triggered interrupt
4302 signals will exit to userspace with the bit in run->s.regs.device_irq_level
4303 set exactly once per edge signal.
4305 The field run->s.regs.device_irq_level is available independent of
4306 run->kvm_valid_regs or run->kvm_dirty_regs bits.
4308 If KVM_CAP_ARM_USER_IRQ is supported, the KVM_CHECK_EXTENSION ioctl returns a
4309 number larger than 0 indicating the version of this capability is implemented
4310 and thereby which bits in in run->s.regs.device_irq_level can signal values.
4312 Currently the following bits are defined for the device_irq_level bitmap:
4314 KVM_CAP_ARM_USER_IRQ >= 1:
4316 KVM_ARM_DEV_EL1_VTIMER - EL1 virtual timer
4317 KVM_ARM_DEV_EL1_PTIMER - EL1 physical timer
4318 KVM_ARM_DEV_PMU - ARM PMU overflow interrupt signal
4320 Future versions of kvm may implement additional events. These will get
4321 indicated by returning a higher number from KVM_CHECK_EXTENSION and will be
4324 8.10 KVM_CAP_PPC_SMT_POSSIBLE
4328 Querying this capability returns a bitmap indicating the possible
4329 virtual SMT modes that can be set using KVM_CAP_PPC_SMT. If bit N
4330 (counting from the right) is set, then a virtual SMT mode of 2^N is
4333 8.11 KVM_CAP_HYPERV_SYNIC2
4337 This capability enables a newer version of Hyper-V Synthetic interrupt
4338 controller (SynIC). The only difference with KVM_CAP_HYPERV_SYNIC is that KVM
4339 doesn't clear SynIC message and event flags pages when they are enabled by
4340 writing to the respective MSRs.
4342 8.12 KVM_CAP_HYPERV_VP_INDEX
4346 This capability indicates that userspace can load HV_X64_MSR_VP_INDEX msr. Its
4347 value is used to denote the target vcpu for a SynIC interrupt. For
4348 compatibilty, KVM initializes this msr to KVM's internal vcpu index. When this
4349 capability is absent, userspace can still query this msr's value.