1 The Definitive KVM (Kernel-based Virtual Machine) API Documentation
2 ===================================================================
7 The kvm API is a set of ioctls that are issued to control various aspects
8 of a virtual machine. The ioctls belong to three classes
10 - System ioctls: These query and set global attributes which affect the
11 whole kvm subsystem. In addition a system ioctl is used to create
14 - VM ioctls: These query and set attributes that affect an entire virtual
15 machine, for example memory layout. In addition a VM ioctl is used to
16 create virtual cpus (vcpus).
18 Only run VM ioctls from the same process (address space) that was used
21 - vcpu ioctls: These query and set attributes that control the operation
22 of a single virtual cpu.
24 Only run vcpu ioctls from the same thread that was used to create the
31 The kvm API is centered around file descriptors. An initial
32 open("/dev/kvm") obtains a handle to the kvm subsystem; this handle
33 can be used to issue system ioctls. A KVM_CREATE_VM ioctl on this
34 handle will create a VM file descriptor which can be used to issue VM
35 ioctls. A KVM_CREATE_VCPU ioctl on a VM fd will create a virtual cpu
36 and return a file descriptor pointing to it. Finally, ioctls on a vcpu
37 fd can be used to control the vcpu, including the important task of
38 actually running guest code.
40 In general file descriptors can be migrated among processes by means
41 of fork() and the SCM_RIGHTS facility of unix domain socket. These
42 kinds of tricks are explicitly not supported by kvm. While they will
43 not cause harm to the host, their actual behavior is not guaranteed by
44 the API. The only supported use is one virtual machine per process,
45 and one vcpu per thread.
51 As of Linux 2.6.22, the KVM ABI has been stabilized: no backward
52 incompatible change are allowed. However, there is an extension
53 facility that allows backward-compatible extensions to the API to be
56 The extension mechanism is not based on the Linux version number.
57 Instead, kvm defines extension identifiers and a facility to query
58 whether a particular extension identifier is available. If it is, a
59 set of ioctls is available for application use.
65 This section describes ioctls that can be used to control kvm guests.
66 For each ioctl, the following information is provided along with a
69 Capability: which KVM extension provides this ioctl. Can be 'basic',
70 which means that is will be provided by any kernel that supports
71 API version 12 (see section 4.1), a KVM_CAP_xyz constant, which
72 means availability needs to be checked with KVM_CHECK_EXTENSION
73 (see section 4.4), or 'none' which means that while not all kernels
74 support this ioctl, there's no capability bit to check its
75 availability: for kernels that don't support the ioctl,
76 the ioctl returns -ENOTTY.
78 Architectures: which instruction set architectures provide this ioctl.
79 x86 includes both i386 and x86_64.
81 Type: system, vm, or vcpu.
83 Parameters: what parameters are accepted by the ioctl.
85 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
86 are not detailed, but errors with specific meanings are.
89 4.1 KVM_GET_API_VERSION
95 Returns: the constant KVM_API_VERSION (=12)
97 This identifies the API version as the stable kvm API. It is not
98 expected that this number will change. However, Linux 2.6.20 and
99 2.6.21 report earlier versions; these are not documented and not
100 supported. Applications should refuse to run if KVM_GET_API_VERSION
101 returns a value other than 12. If this check passes, all ioctls
102 described as 'basic' will be available.
110 Parameters: machine type identifier (KVM_VM_*)
111 Returns: a VM fd that can be used to control the new virtual machine.
113 The new VM has no virtual cpus and no memory. An mmap() of a VM fd
114 will access the virtual machine's physical address space; offset zero
115 corresponds to guest physical address zero. Use of mmap() on a VM fd
116 is discouraged if userspace memory allocation (KVM_CAP_USER_MEMORY) is
118 You most certainly want to use 0 as machine type.
120 In order to create user controlled virtual machines on S390, check
121 KVM_CAP_S390_UCONTROL and use the flag KVM_VM_S390_UCONTROL as
122 privileged user (CAP_SYS_ADMIN).
125 4.3 KVM_GET_MSR_INDEX_LIST
130 Parameters: struct kvm_msr_list (in/out)
131 Returns: 0 on success; -1 on error
133 E2BIG: the msr index list is to be to fit in the array specified by
136 struct kvm_msr_list {
137 __u32 nmsrs; /* number of msrs in entries */
141 This ioctl returns the guest msrs that are supported. The list varies
142 by kvm version and host processor, but does not change otherwise. The
143 user fills in the size of the indices array in nmsrs, and in return
144 kvm adjusts nmsrs to reflect the actual number of msrs and fills in
145 the indices array with their numbers.
147 Note: if kvm indicates supports MCE (KVM_CAP_MCE), then the MCE bank MSRs are
148 not returned in the MSR list, as different vcpus can have a different number
149 of banks, as set via the KVM_X86_SETUP_MCE ioctl.
152 4.4 KVM_CHECK_EXTENSION
154 Capability: basic, KVM_CAP_CHECK_EXTENSION_VM for vm ioctl
156 Type: system ioctl, vm ioctl
157 Parameters: extension identifier (KVM_CAP_*)
158 Returns: 0 if unsupported; 1 (or some other positive integer) if supported
160 The API allows the application to query about extensions to the core
161 kvm API. Userspace passes an extension identifier (an integer) and
162 receives an integer that describes the extension availability.
163 Generally 0 means no and 1 means yes, but some extensions may report
164 additional information in the integer return value.
166 Based on their initialization different VMs may have different capabilities.
167 It is thus encouraged to use the vm ioctl to query for capabilities (available
168 with KVM_CAP_CHECK_EXTENSION_VM on the vm fd)
170 4.5 KVM_GET_VCPU_MMAP_SIZE
176 Returns: size of vcpu mmap area, in bytes
178 The KVM_RUN ioctl (cf.) communicates with userspace via a shared
179 memory region. This ioctl returns the size of that region. See the
180 KVM_RUN documentation for details.
183 4.6 KVM_SET_MEMORY_REGION
188 Parameters: struct kvm_memory_region (in)
189 Returns: 0 on success, -1 on error
191 This ioctl is obsolete and has been removed.
199 Parameters: vcpu id (apic id on x86)
200 Returns: vcpu fd on success, -1 on error
202 This API adds a vcpu to a virtual machine. No more than max_vcpus may be added.
203 The vcpu id is an integer in the range [0, max_vcpu_id).
205 The recommended max_vcpus value can be retrieved using the KVM_CAP_NR_VCPUS of
206 the KVM_CHECK_EXTENSION ioctl() at run-time.
207 The maximum possible value for max_vcpus can be retrieved using the
208 KVM_CAP_MAX_VCPUS of the KVM_CHECK_EXTENSION ioctl() at run-time.
210 If the KVM_CAP_NR_VCPUS does not exist, you should assume that max_vcpus is 4
212 If the KVM_CAP_MAX_VCPUS does not exist, you should assume that max_vcpus is
213 same as the value returned from KVM_CAP_NR_VCPUS.
215 The maximum possible value for max_vcpu_id can be retrieved using the
216 KVM_CAP_MAX_VCPU_ID of the KVM_CHECK_EXTENSION ioctl() at run-time.
218 If the KVM_CAP_MAX_VCPU_ID does not exist, you should assume that max_vcpu_id
219 is the same as the value returned from KVM_CAP_MAX_VCPUS.
221 On powerpc using book3s_hv mode, the vcpus are mapped onto virtual
222 threads in one or more virtual CPU cores. (This is because the
223 hardware requires all the hardware threads in a CPU core to be in the
224 same partition.) The KVM_CAP_PPC_SMT capability indicates the number
225 of vcpus per virtual core (vcore). The vcore id is obtained by
226 dividing the vcpu id by the number of vcpus per vcore. The vcpus in a
227 given vcore will always be in the same physical core as each other
228 (though that might be a different physical core from time to time).
229 Userspace can control the threading (SMT) mode of the guest by its
230 allocation of vcpu ids. For example, if userspace wants
231 single-threaded guest vcpus, it should make all vcpu ids be a multiple
232 of the number of vcpus per vcore.
234 For virtual cpus that have been created with S390 user controlled virtual
235 machines, the resulting vcpu fd can be memory mapped at page offset
236 KVM_S390_SIE_PAGE_OFFSET in order to obtain a memory map of the virtual
237 cpu's hardware control block.
240 4.8 KVM_GET_DIRTY_LOG (vm ioctl)
245 Parameters: struct kvm_dirty_log (in/out)
246 Returns: 0 on success, -1 on error
248 /* for KVM_GET_DIRTY_LOG */
249 struct kvm_dirty_log {
253 void __user *dirty_bitmap; /* one bit per page */
258 Given a memory slot, return a bitmap containing any pages dirtied
259 since the last call to this ioctl. Bit 0 is the first page in the
260 memory slot. Ensure the entire structure is cleared to avoid padding
263 If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 specifies
264 the address space for which you want to return the dirty bitmap.
265 They must be less than the value that KVM_CHECK_EXTENSION returns for
266 the KVM_CAP_MULTI_ADDRESS_SPACE capability.
269 4.9 KVM_SET_MEMORY_ALIAS
274 Parameters: struct kvm_memory_alias (in)
275 Returns: 0 (success), -1 (error)
277 This ioctl is obsolete and has been removed.
286 Returns: 0 on success, -1 on error
288 EINTR: an unmasked signal is pending
290 This ioctl is used to run a guest virtual cpu. While there are no
291 explicit parameters, there is an implicit parameter block that can be
292 obtained by mmap()ing the vcpu fd at offset 0, with the size given by
293 KVM_GET_VCPU_MMAP_SIZE. The parameter block is formatted as a 'struct
294 kvm_run' (see below).
300 Architectures: all except ARM, arm64
302 Parameters: struct kvm_regs (out)
303 Returns: 0 on success, -1 on error
305 Reads the general purpose registers from the vcpu.
309 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
310 __u64 rax, rbx, rcx, rdx;
311 __u64 rsi, rdi, rsp, rbp;
312 __u64 r8, r9, r10, r11;
313 __u64 r12, r13, r14, r15;
319 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
330 Architectures: all except ARM, arm64
332 Parameters: struct kvm_regs (in)
333 Returns: 0 on success, -1 on error
335 Writes the general purpose registers into the vcpu.
337 See KVM_GET_REGS for the data structure.
343 Architectures: x86, ppc
345 Parameters: struct kvm_sregs (out)
346 Returns: 0 on success, -1 on error
348 Reads special registers from the vcpu.
352 struct kvm_segment cs, ds, es, fs, gs, ss;
353 struct kvm_segment tr, ldt;
354 struct kvm_dtable gdt, idt;
355 __u64 cr0, cr2, cr3, cr4, cr8;
358 __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64];
361 /* ppc -- see arch/powerpc/include/uapi/asm/kvm.h */
363 interrupt_bitmap is a bitmap of pending external interrupts. At most
364 one bit may be set. This interrupt has been acknowledged by the APIC
365 but not yet injected into the cpu core.
371 Architectures: x86, ppc
373 Parameters: struct kvm_sregs (in)
374 Returns: 0 on success, -1 on error
376 Writes special registers into the vcpu. See KVM_GET_SREGS for the
385 Parameters: struct kvm_translation (in/out)
386 Returns: 0 on success, -1 on error
388 Translates a virtual address according to the vcpu's current address
391 struct kvm_translation {
393 __u64 linear_address;
396 __u64 physical_address;
407 Architectures: x86, ppc, mips
409 Parameters: struct kvm_interrupt (in)
410 Returns: 0 on success, negative on failure.
412 Queues a hardware interrupt vector to be injected.
414 /* for KVM_INTERRUPT */
415 struct kvm_interrupt {
422 Returns: 0 on success,
423 -EEXIST if an interrupt is already enqueued
424 -EINVAL the the irq number is invalid
425 -ENXIO if the PIC is in the kernel
426 -EFAULT if the pointer is invalid
428 Note 'irq' is an interrupt vector, not an interrupt pin or line. This
429 ioctl is useful if the in-kernel PIC is not used.
433 Queues an external interrupt to be injected. This ioctl is overleaded
434 with 3 different irq values:
438 This injects an edge type external interrupt into the guest once it's ready
439 to receive interrupts. When injected, the interrupt is done.
441 b) KVM_INTERRUPT_UNSET
443 This unsets any pending interrupt.
445 Only available with KVM_CAP_PPC_UNSET_IRQ.
447 c) KVM_INTERRUPT_SET_LEVEL
449 This injects a level type external interrupt into the guest context. The
450 interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET
453 Only available with KVM_CAP_PPC_IRQ_LEVEL.
455 Note that any value for 'irq' other than the ones stated above is invalid
456 and incurs unexpected behavior.
460 Queues an external interrupt to be injected into the virtual CPU. A negative
461 interrupt number dequeues the interrupt.
472 Support for this has been removed. Use KVM_SET_GUEST_DEBUG instead.
480 Parameters: struct kvm_msrs (in/out)
481 Returns: 0 on success, -1 on error
483 Reads model-specific registers from the vcpu. Supported msr indices can
484 be obtained using KVM_GET_MSR_INDEX_LIST.
487 __u32 nmsrs; /* number of msrs in entries */
490 struct kvm_msr_entry entries[0];
493 struct kvm_msr_entry {
499 Application code should set the 'nmsrs' member (which indicates the
500 size of the entries array) and the 'index' member of each array entry.
501 kvm will fill in the 'data' member.
509 Parameters: struct kvm_msrs (in)
510 Returns: 0 on success, -1 on error
512 Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the
515 Application code should set the 'nmsrs' member (which indicates the
516 size of the entries array), and the 'index' and 'data' members of each
525 Parameters: struct kvm_cpuid (in)
526 Returns: 0 on success, -1 on error
528 Defines the vcpu responses to the cpuid instruction. Applications
529 should use the KVM_SET_CPUID2 ioctl if available.
532 struct kvm_cpuid_entry {
541 /* for KVM_SET_CPUID */
545 struct kvm_cpuid_entry entries[0];
549 4.21 KVM_SET_SIGNAL_MASK
554 Parameters: struct kvm_signal_mask (in)
555 Returns: 0 on success, -1 on error
557 Defines which signals are blocked during execution of KVM_RUN. This
558 signal mask temporarily overrides the threads signal mask. Any
559 unblocked signal received (except SIGKILL and SIGSTOP, which retain
560 their traditional behaviour) will cause KVM_RUN to return with -EINTR.
562 Note the signal will only be delivered if not blocked by the original
565 /* for KVM_SET_SIGNAL_MASK */
566 struct kvm_signal_mask {
577 Parameters: struct kvm_fpu (out)
578 Returns: 0 on success, -1 on error
580 Reads the floating point state from the vcpu.
582 /* for KVM_GET_FPU and KVM_SET_FPU */
587 __u8 ftwx; /* in fxsave format */
603 Parameters: struct kvm_fpu (in)
604 Returns: 0 on success, -1 on error
606 Writes the floating point state to the vcpu.
608 /* for KVM_GET_FPU and KVM_SET_FPU */
613 __u8 ftwx; /* in fxsave format */
624 4.24 KVM_CREATE_IRQCHIP
626 Capability: KVM_CAP_IRQCHIP, KVM_CAP_S390_IRQCHIP (s390)
627 Architectures: x86, ARM, arm64, s390
630 Returns: 0 on success, -1 on error
632 Creates an interrupt controller model in the kernel.
633 On x86, creates a virtual ioapic, a virtual PIC (two PICs, nested), and sets up
634 future vcpus to have a local APIC. IRQ routing for GSIs 0-15 is set to both
635 PIC and IOAPIC; GSI 16-23 only go to the IOAPIC.
636 On ARM/arm64, a GICv2 is created. Any other GIC versions require the usage of
637 KVM_CREATE_DEVICE, which also supports creating a GICv2. Using
638 KVM_CREATE_DEVICE is preferred over KVM_CREATE_IRQCHIP for GICv2.
639 On s390, a dummy irq routing table is created.
641 Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled
642 before KVM_CREATE_IRQCHIP can be used.
647 Capability: KVM_CAP_IRQCHIP
648 Architectures: x86, arm, arm64
650 Parameters: struct kvm_irq_level
651 Returns: 0 on success, -1 on error
653 Sets the level of a GSI input to the interrupt controller model in the kernel.
654 On some architectures it is required that an interrupt controller model has
655 been previously created with KVM_CREATE_IRQCHIP. Note that edge-triggered
656 interrupts require the level to be set to 1 and then back to 0.
658 On real hardware, interrupt pins can be active-low or active-high. This
659 does not matter for the level field of struct kvm_irq_level: 1 always
660 means active (asserted), 0 means inactive (deasserted).
662 x86 allows the operating system to program the interrupt polarity
663 (active-low/active-high) for level-triggered interrupts, and KVM used
664 to consider the polarity. However, due to bitrot in the handling of
665 active-low interrupts, the above convention is now valid on x86 too.
666 This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED. Userspace
667 should not present interrupts to the guest as active-low unless this
668 capability is present (or unless it is not using the in-kernel irqchip,
672 ARM/arm64 can signal an interrupt either at the CPU level, or at the
673 in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to
674 use PPIs designated for specific cpus. The irq field is interpreted
677 Â bits: | 31 ... 24 | 23 ... 16 | 15 ... 0 |
678 field: | irq_type | vcpu_index | irq_id |
680 The irq_type field has the following values:
681 - irq_type[0]: out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ
682 - irq_type[1]: in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.)
683 (the vcpu_index field is ignored)
684 - irq_type[2]: in-kernel GIC: PPI, irq_id between 16 and 31 (incl.)
686 (The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs)
688 In both cases, level is used to assert/deassert the line.
690 struct kvm_irq_level {
693 __s32 status; /* not used for KVM_IRQ_LEVEL */
695 __u32 level; /* 0 or 1 */
701 Capability: KVM_CAP_IRQCHIP
704 Parameters: struct kvm_irqchip (in/out)
705 Returns: 0 on success, -1 on error
707 Reads the state of a kernel interrupt controller created with
708 KVM_CREATE_IRQCHIP into a buffer provided by the caller.
711 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
714 char dummy[512]; /* reserving space */
715 struct kvm_pic_state pic;
716 struct kvm_ioapic_state ioapic;
723 Capability: KVM_CAP_IRQCHIP
726 Parameters: struct kvm_irqchip (in)
727 Returns: 0 on success, -1 on error
729 Sets the state of a kernel interrupt controller created with
730 KVM_CREATE_IRQCHIP from a buffer provided by the caller.
733 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
736 char dummy[512]; /* reserving space */
737 struct kvm_pic_state pic;
738 struct kvm_ioapic_state ioapic;
743 4.28 KVM_XEN_HVM_CONFIG
745 Capability: KVM_CAP_XEN_HVM
748 Parameters: struct kvm_xen_hvm_config (in)
749 Returns: 0 on success, -1 on error
751 Sets the MSR that the Xen HVM guest uses to initialize its hypercall
752 page, and provides the starting address and size of the hypercall
753 blobs in userspace. When the guest writes the MSR, kvm copies one
754 page of a blob (32- or 64-bit, depending on the vcpu mode) to guest
757 struct kvm_xen_hvm_config {
770 Capability: KVM_CAP_ADJUST_CLOCK
773 Parameters: struct kvm_clock_data (out)
774 Returns: 0 on success, -1 on error
776 Gets the current timestamp of kvmclock as seen by the current guest. In
777 conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios
780 When KVM_CAP_ADJUST_CLOCK is passed to KVM_CHECK_EXTENSION, it returns the
781 set of bits that KVM can return in struct kvm_clock_data's flag member.
783 The only flag defined now is KVM_CLOCK_TSC_STABLE. If set, the returned
784 value is the exact kvmclock value seen by all VCPUs at the instant
785 when KVM_GET_CLOCK was called. If clear, the returned value is simply
786 CLOCK_MONOTONIC plus a constant offset; the offset can be modified
787 with KVM_SET_CLOCK. KVM will try to make all VCPUs follow this clock,
788 but the exact value read by each VCPU could differ, because the host
791 struct kvm_clock_data {
792 __u64 clock; /* kvmclock current value */
800 Capability: KVM_CAP_ADJUST_CLOCK
803 Parameters: struct kvm_clock_data (in)
804 Returns: 0 on success, -1 on error
806 Sets the current timestamp of kvmclock to the value specified in its parameter.
807 In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios
810 struct kvm_clock_data {
811 __u64 clock; /* kvmclock current value */
817 4.31 KVM_GET_VCPU_EVENTS
819 Capability: KVM_CAP_VCPU_EVENTS
820 Extended by: KVM_CAP_INTR_SHADOW
823 Parameters: struct kvm_vcpu_event (out)
824 Returns: 0 on success, -1 on error
826 Gets currently pending exceptions, interrupts, and NMIs as well as related
829 struct kvm_vcpu_events {
859 Only two fields are defined in the flags field:
861 - KVM_VCPUEVENT_VALID_SHADOW may be set in the flags field to signal that
862 interrupt.shadow contains a valid state.
864 - KVM_VCPUEVENT_VALID_SMM may be set in the flags field to signal that
865 smi contains a valid state.
867 4.32 KVM_SET_VCPU_EVENTS
869 Capability: KVM_CAP_VCPU_EVENTS
870 Extended by: KVM_CAP_INTR_SHADOW
873 Parameters: struct kvm_vcpu_event (in)
874 Returns: 0 on success, -1 on error
876 Set pending exceptions, interrupts, and NMIs as well as related states of the
879 See KVM_GET_VCPU_EVENTS for the data structure.
881 Fields that may be modified asynchronously by running VCPUs can be excluded
882 from the update. These fields are nmi.pending, sipi_vector, smi.smm,
883 smi.pending. Keep the corresponding bits in the flags field cleared to
884 suppress overwriting the current in-kernel state. The bits are:
886 KVM_VCPUEVENT_VALID_NMI_PENDING - transfer nmi.pending to the kernel
887 KVM_VCPUEVENT_VALID_SIPI_VECTOR - transfer sipi_vector
888 KVM_VCPUEVENT_VALID_SMM - transfer the smi sub-struct.
890 If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in
891 the flags field to signal that interrupt.shadow contains a valid state and
892 shall be written into the VCPU.
894 KVM_VCPUEVENT_VALID_SMM can only be set if KVM_CAP_X86_SMM is available.
897 4.33 KVM_GET_DEBUGREGS
899 Capability: KVM_CAP_DEBUGREGS
902 Parameters: struct kvm_debugregs (out)
903 Returns: 0 on success, -1 on error
905 Reads debug registers from the vcpu.
907 struct kvm_debugregs {
916 4.34 KVM_SET_DEBUGREGS
918 Capability: KVM_CAP_DEBUGREGS
921 Parameters: struct kvm_debugregs (in)
922 Returns: 0 on success, -1 on error
924 Writes debug registers into the vcpu.
926 See KVM_GET_DEBUGREGS for the data structure. The flags field is unused
927 yet and must be cleared on entry.
930 4.35 KVM_SET_USER_MEMORY_REGION
932 Capability: KVM_CAP_USER_MEM
935 Parameters: struct kvm_userspace_memory_region (in)
936 Returns: 0 on success, -1 on error
938 struct kvm_userspace_memory_region {
941 __u64 guest_phys_addr;
942 __u64 memory_size; /* bytes */
943 __u64 userspace_addr; /* start of the userspace allocated memory */
946 /* for kvm_memory_region::flags */
947 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0)
948 #define KVM_MEM_READONLY (1UL << 1)
950 This ioctl allows the user to create or modify a guest physical memory
951 slot. When changing an existing slot, it may be moved in the guest
952 physical memory space, or its flags may be modified. It may not be
953 resized. Slots may not overlap in guest physical address space.
954 Bits 0-15 of "slot" specifies the slot id and this value should be
955 less than the maximum number of user memory slots supported per VM.
956 The maximum allowed slots can be queried using KVM_CAP_NR_MEMSLOTS,
957 if this capability is supported by the architecture.
959 If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of "slot"
960 specifies the address space which is being modified. They must be
961 less than the value that KVM_CHECK_EXTENSION returns for the
962 KVM_CAP_MULTI_ADDRESS_SPACE capability. Slots in separate address spaces
963 are unrelated; the restriction on overlapping slots only applies within
966 Memory for the region is taken starting at the address denoted by the
967 field userspace_addr, which must point at user addressable memory for
968 the entire memory slot size. Any object may back this memory, including
969 anonymous memory, ordinary files, and hugetlbfs.
971 It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr
972 be identical. This allows large pages in the guest to be backed by large
975 The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and
976 KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of
977 writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to
978 use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it,
979 to make a new slot read-only. In this case, writes to this memory will be
980 posted to userspace as KVM_EXIT_MMIO exits.
982 When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of
983 the memory region are automatically reflected into the guest. For example, an
984 mmap() that affects the region will be made visible immediately. Another
985 example is madvise(MADV_DROP).
987 It is recommended to use this API instead of the KVM_SET_MEMORY_REGION ioctl.
988 The KVM_SET_MEMORY_REGION does not allow fine grained control over memory
989 allocation and is deprecated.
992 4.36 KVM_SET_TSS_ADDR
994 Capability: KVM_CAP_SET_TSS_ADDR
997 Parameters: unsigned long tss_address (in)
998 Returns: 0 on success, -1 on error
1000 This ioctl defines the physical address of a three-page region in the guest
1001 physical address space. The region must be within the first 4GB of the
1002 guest physical address space and must not conflict with any memory slot
1003 or any mmio address. The guest may malfunction if it accesses this memory
1006 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
1007 because of a quirk in the virtualization implementation (see the internals
1008 documentation when it pops into existence).
1013 Capability: KVM_CAP_ENABLE_CAP, KVM_CAP_ENABLE_CAP_VM
1014 Architectures: x86 (only KVM_CAP_ENABLE_CAP_VM),
1015 mips (only KVM_CAP_ENABLE_CAP), ppc, s390
1016 Type: vcpu ioctl, vm ioctl (with KVM_CAP_ENABLE_CAP_VM)
1017 Parameters: struct kvm_enable_cap (in)
1018 Returns: 0 on success; -1 on error
1020 +Not all extensions are enabled by default. Using this ioctl the application
1021 can enable an extension, making it available to the guest.
1023 On systems that do not support this ioctl, it always fails. On systems that
1024 do support it, it only works for extensions that are supported for enablement.
1026 To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should
1029 struct kvm_enable_cap {
1033 The capability that is supposed to get enabled.
1037 A bitfield indicating future enhancements. Has to be 0 for now.
1041 Arguments for enabling a feature. If a feature needs initial values to
1042 function properly, this is the place to put them.
1047 The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl
1048 for vm-wide capabilities.
1050 4.38 KVM_GET_MP_STATE
1052 Capability: KVM_CAP_MP_STATE
1053 Architectures: x86, s390, arm, arm64
1055 Parameters: struct kvm_mp_state (out)
1056 Returns: 0 on success; -1 on error
1058 struct kvm_mp_state {
1062 Returns the vcpu's current "multiprocessing state" (though also valid on
1063 uniprocessor guests).
1065 Possible values are:
1067 - KVM_MP_STATE_RUNNABLE: the vcpu is currently running [x86,arm/arm64]
1068 - KVM_MP_STATE_UNINITIALIZED: the vcpu is an application processor (AP)
1069 which has not yet received an INIT signal [x86]
1070 - KVM_MP_STATE_INIT_RECEIVED: the vcpu has received an INIT signal, and is
1071 now ready for a SIPI [x86]
1072 - KVM_MP_STATE_HALTED: the vcpu has executed a HLT instruction and
1073 is waiting for an interrupt [x86]
1074 - KVM_MP_STATE_SIPI_RECEIVED: the vcpu has just received a SIPI (vector
1075 accessible via KVM_GET_VCPU_EVENTS) [x86]
1076 - KVM_MP_STATE_STOPPED: the vcpu is stopped [s390,arm/arm64]
1077 - KVM_MP_STATE_CHECK_STOP: the vcpu is in a special error state [s390]
1078 - KVM_MP_STATE_OPERATING: the vcpu is operating (running or halted)
1080 - KVM_MP_STATE_LOAD: the vcpu is in a special load/startup state
1083 On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an
1084 in-kernel irqchip, the multiprocessing state must be maintained by userspace on
1085 these architectures.
1089 The only states that are valid are KVM_MP_STATE_STOPPED and
1090 KVM_MP_STATE_RUNNABLE which reflect if the vcpu is paused or not.
1092 4.39 KVM_SET_MP_STATE
1094 Capability: KVM_CAP_MP_STATE
1095 Architectures: x86, s390, arm, arm64
1097 Parameters: struct kvm_mp_state (in)
1098 Returns: 0 on success; -1 on error
1100 Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for
1103 On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an
1104 in-kernel irqchip, the multiprocessing state must be maintained by userspace on
1105 these architectures.
1109 The only states that are valid are KVM_MP_STATE_STOPPED and
1110 KVM_MP_STATE_RUNNABLE which reflect if the vcpu should be paused or not.
1112 4.40 KVM_SET_IDENTITY_MAP_ADDR
1114 Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR
1117 Parameters: unsigned long identity (in)
1118 Returns: 0 on success, -1 on error
1120 This ioctl defines the physical address of a one-page region in the guest
1121 physical address space. The region must be within the first 4GB of the
1122 guest physical address space and must not conflict with any memory slot
1123 or any mmio address. The guest may malfunction if it accesses this memory
1126 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
1127 because of a quirk in the virtualization implementation (see the internals
1128 documentation when it pops into existence).
1131 4.41 KVM_SET_BOOT_CPU_ID
1133 Capability: KVM_CAP_SET_BOOT_CPU_ID
1136 Parameters: unsigned long vcpu_id
1137 Returns: 0 on success, -1 on error
1139 Define which vcpu is the Bootstrap Processor (BSP). Values are the same
1140 as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default
1146 Capability: KVM_CAP_XSAVE
1149 Parameters: struct kvm_xsave (out)
1150 Returns: 0 on success, -1 on error
1156 This ioctl would copy current vcpu's xsave struct to the userspace.
1161 Capability: KVM_CAP_XSAVE
1164 Parameters: struct kvm_xsave (in)
1165 Returns: 0 on success, -1 on error
1171 This ioctl would copy userspace's xsave struct to the kernel.
1176 Capability: KVM_CAP_XCRS
1179 Parameters: struct kvm_xcrs (out)
1180 Returns: 0 on success, -1 on error
1191 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1195 This ioctl would copy current vcpu's xcrs to the userspace.
1200 Capability: KVM_CAP_XCRS
1203 Parameters: struct kvm_xcrs (in)
1204 Returns: 0 on success, -1 on error
1215 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1219 This ioctl would set vcpu's xcr to the value userspace specified.
1222 4.46 KVM_GET_SUPPORTED_CPUID
1224 Capability: KVM_CAP_EXT_CPUID
1227 Parameters: struct kvm_cpuid2 (in/out)
1228 Returns: 0 on success, -1 on error
1233 struct kvm_cpuid_entry2 entries[0];
1236 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
1237 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
1238 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
1240 struct kvm_cpuid_entry2 {
1251 This ioctl returns x86 cpuid features which are supported by both the hardware
1252 and kvm. Userspace can use the information returned by this ioctl to
1253 construct cpuid information (for KVM_SET_CPUID2) that is consistent with
1254 hardware, kernel, and userspace capabilities, and with user requirements (for
1255 example, the user may wish to constrain cpuid to emulate older hardware,
1256 or for feature consistency across a cluster).
1258 Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure
1259 with the 'nent' field indicating the number of entries in the variable-size
1260 array 'entries'. If the number of entries is too low to describe the cpu
1261 capabilities, an error (E2BIG) is returned. If the number is too high,
1262 the 'nent' field is adjusted and an error (ENOMEM) is returned. If the
1263 number is just right, the 'nent' field is adjusted to the number of valid
1264 entries in the 'entries' array, which is then filled.
1266 The entries returned are the host cpuid as returned by the cpuid instruction,
1267 with unknown or unsupported features masked out. Some features (for example,
1268 x2apic), may not be present in the host cpu, but are exposed by kvm if it can
1269 emulate them efficiently. The fields in each entry are defined as follows:
1271 function: the eax value used to obtain the entry
1272 index: the ecx value used to obtain the entry (for entries that are
1274 flags: an OR of zero or more of the following:
1275 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
1276 if the index field is valid
1277 KVM_CPUID_FLAG_STATEFUL_FUNC:
1278 if cpuid for this function returns different values for successive
1279 invocations; there will be several entries with the same function,
1280 all with this flag set
1281 KVM_CPUID_FLAG_STATE_READ_NEXT:
1282 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
1283 the first entry to be read by a cpu
1284 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
1285 this function/index combination
1287 The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned
1288 as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC
1289 support. Instead it is reported via
1291 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER)
1293 if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the
1294 feature in userspace, then you can enable the feature for KVM_SET_CPUID2.
1297 4.47 KVM_PPC_GET_PVINFO
1299 Capability: KVM_CAP_PPC_GET_PVINFO
1302 Parameters: struct kvm_ppc_pvinfo (out)
1303 Returns: 0 on success, !0 on error
1305 struct kvm_ppc_pvinfo {
1311 This ioctl fetches PV specific information that need to be passed to the guest
1312 using the device tree or other means from vm context.
1314 The hcall array defines 4 instructions that make up a hypercall.
1316 If any additional field gets added to this structure later on, a bit for that
1317 additional piece of information will be set in the flags bitmap.
1319 The flags bitmap is defined as:
1321 /* the host supports the ePAPR idle hcall
1322 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0)
1324 4.48 KVM_ASSIGN_PCI_DEVICE (deprecated)
1329 Parameters: struct kvm_assigned_pci_dev (in)
1330 Returns: 0 on success, -1 on error
1332 Assigns a host PCI device to the VM.
1334 struct kvm_assigned_pci_dev {
1335 __u32 assigned_dev_id;
1345 The PCI device is specified by the triple segnr, busnr, and devfn.
1346 Identification in succeeding service requests is done via assigned_dev_id. The
1347 following flags are specified:
1349 /* Depends on KVM_CAP_IOMMU */
1350 #define KVM_DEV_ASSIGN_ENABLE_IOMMU (1 << 0)
1351 /* The following two depend on KVM_CAP_PCI_2_3 */
1352 #define KVM_DEV_ASSIGN_PCI_2_3 (1 << 1)
1353 #define KVM_DEV_ASSIGN_MASK_INTX (1 << 2)
1355 If KVM_DEV_ASSIGN_PCI_2_3 is set, the kernel will manage legacy INTx interrupts
1356 via the PCI-2.3-compliant device-level mask, thus enable IRQ sharing with other
1357 assigned devices or host devices. KVM_DEV_ASSIGN_MASK_INTX specifies the
1358 guest's view on the INTx mask, see KVM_ASSIGN_SET_INTX_MASK for details.
1360 The KVM_DEV_ASSIGN_ENABLE_IOMMU flag is a mandatory option to ensure
1361 isolation of the device. Usages not specifying this flag are deprecated.
1363 Only PCI header type 0 devices with PCI BAR resources are supported by
1364 device assignment. The user requesting this ioctl must have read/write
1365 access to the PCI sysfs resource files associated with the device.
1368 ENOTTY: kernel does not support this ioctl
1370 Other error conditions may be defined by individual device types or
1371 have their standard meanings.
1374 4.49 KVM_DEASSIGN_PCI_DEVICE (deprecated)
1379 Parameters: struct kvm_assigned_pci_dev (in)
1380 Returns: 0 on success, -1 on error
1382 Ends PCI device assignment, releasing all associated resources.
1384 See KVM_ASSIGN_PCI_DEVICE for the data structure. Only assigned_dev_id is
1385 used in kvm_assigned_pci_dev to identify the device.
1388 ENOTTY: kernel does not support this ioctl
1390 Other error conditions may be defined by individual device types or
1391 have their standard meanings.
1393 4.50 KVM_ASSIGN_DEV_IRQ (deprecated)
1395 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1398 Parameters: struct kvm_assigned_irq (in)
1399 Returns: 0 on success, -1 on error
1401 Assigns an IRQ to a passed-through device.
1403 struct kvm_assigned_irq {
1404 __u32 assigned_dev_id;
1405 __u32 host_irq; /* ignored (legacy field) */
1413 The following flags are defined:
1415 #define KVM_DEV_IRQ_HOST_INTX (1 << 0)
1416 #define KVM_DEV_IRQ_HOST_MSI (1 << 1)
1417 #define KVM_DEV_IRQ_HOST_MSIX (1 << 2)
1419 #define KVM_DEV_IRQ_GUEST_INTX (1 << 8)
1420 #define KVM_DEV_IRQ_GUEST_MSI (1 << 9)
1421 #define KVM_DEV_IRQ_GUEST_MSIX (1 << 10)
1423 It is not valid to specify multiple types per host or guest IRQ. However, the
1424 IRQ type of host and guest can differ or can even be null.
1427 ENOTTY: kernel does not support this ioctl
1429 Other error conditions may be defined by individual device types or
1430 have their standard meanings.
1433 4.51 KVM_DEASSIGN_DEV_IRQ (deprecated)
1435 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1438 Parameters: struct kvm_assigned_irq (in)
1439 Returns: 0 on success, -1 on error
1441 Ends an IRQ assignment to a passed-through device.
1443 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1444 by assigned_dev_id, flags must correspond to the IRQ type specified on
1445 KVM_ASSIGN_DEV_IRQ. Partial deassignment of host or guest IRQ is allowed.
1448 4.52 KVM_SET_GSI_ROUTING
1450 Capability: KVM_CAP_IRQ_ROUTING
1451 Architectures: x86 s390 arm arm64
1453 Parameters: struct kvm_irq_routing (in)
1454 Returns: 0 on success, -1 on error
1456 Sets the GSI routing table entries, overwriting any previously set entries.
1458 On arm/arm64, GSI routing has the following limitation:
1459 - GSI routing does not apply to KVM_IRQ_LINE but only to KVM_IRQFD.
1461 struct kvm_irq_routing {
1464 struct kvm_irq_routing_entry entries[0];
1467 No flags are specified so far, the corresponding field must be set to zero.
1469 struct kvm_irq_routing_entry {
1475 struct kvm_irq_routing_irqchip irqchip;
1476 struct kvm_irq_routing_msi msi;
1477 struct kvm_irq_routing_s390_adapter adapter;
1478 struct kvm_irq_routing_hv_sint hv_sint;
1483 /* gsi routing entry types */
1484 #define KVM_IRQ_ROUTING_IRQCHIP 1
1485 #define KVM_IRQ_ROUTING_MSI 2
1486 #define KVM_IRQ_ROUTING_S390_ADAPTER 3
1487 #define KVM_IRQ_ROUTING_HV_SINT 4
1490 - KVM_MSI_VALID_DEVID: used along with KVM_IRQ_ROUTING_MSI routing entry
1491 type, specifies that the devid field contains a valid value. The per-VM
1492 KVM_CAP_MSI_DEVID capability advertises the requirement to provide
1493 the device ID. If this capability is not available, userspace should
1494 never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail.
1497 struct kvm_irq_routing_irqchip {
1502 struct kvm_irq_routing_msi {
1512 If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier
1513 for the device that wrote the MSI message. For PCI, this is usually a
1514 BFD identifier in the lower 16 bits.
1516 On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS
1517 feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled,
1518 address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of
1519 address_hi must be zero.
1521 struct kvm_irq_routing_s390_adapter {
1525 __u32 summary_offset;
1529 struct kvm_irq_routing_hv_sint {
1534 4.53 KVM_ASSIGN_SET_MSIX_NR (deprecated)
1539 Parameters: struct kvm_assigned_msix_nr (in)
1540 Returns: 0 on success, -1 on error
1542 Set the number of MSI-X interrupts for an assigned device. The number is
1543 reset again by terminating the MSI-X assignment of the device via
1544 KVM_DEASSIGN_DEV_IRQ. Calling this service more than once at any earlier
1547 struct kvm_assigned_msix_nr {
1548 __u32 assigned_dev_id;
1553 #define KVM_MAX_MSIX_PER_DEV 256
1556 4.54 KVM_ASSIGN_SET_MSIX_ENTRY (deprecated)
1561 Parameters: struct kvm_assigned_msix_entry (in)
1562 Returns: 0 on success, -1 on error
1564 Specifies the routing of an MSI-X assigned device interrupt to a GSI. Setting
1565 the GSI vector to zero means disabling the interrupt.
1567 struct kvm_assigned_msix_entry {
1568 __u32 assigned_dev_id;
1570 __u16 entry; /* The index of entry in the MSI-X table */
1575 ENOTTY: kernel does not support this ioctl
1577 Other error conditions may be defined by individual device types or
1578 have their standard meanings.
1581 4.55 KVM_SET_TSC_KHZ
1583 Capability: KVM_CAP_TSC_CONTROL
1586 Parameters: virtual tsc_khz
1587 Returns: 0 on success, -1 on error
1589 Specifies the tsc frequency for the virtual machine. The unit of the
1593 4.56 KVM_GET_TSC_KHZ
1595 Capability: KVM_CAP_GET_TSC_KHZ
1599 Returns: virtual tsc-khz on success, negative value on error
1601 Returns the tsc frequency of the guest. The unit of the return value is
1602 KHz. If the host has unstable tsc this ioctl returns -EIO instead as an
1608 Capability: KVM_CAP_IRQCHIP
1611 Parameters: struct kvm_lapic_state (out)
1612 Returns: 0 on success, -1 on error
1614 #define KVM_APIC_REG_SIZE 0x400
1615 struct kvm_lapic_state {
1616 char regs[KVM_APIC_REG_SIZE];
1619 Reads the Local APIC registers and copies them into the input argument. The
1620 data format and layout are the same as documented in the architecture manual.
1622 If KVM_X2APIC_API_USE_32BIT_IDS feature of KVM_CAP_X2APIC_API is
1623 enabled, then the format of APIC_ID register depends on the APIC mode
1624 (reported by MSR_IA32_APICBASE) of its VCPU. x2APIC stores APIC ID in
1625 the APIC_ID register (bytes 32-35). xAPIC only allows an 8-bit APIC ID
1626 which is stored in bits 31-24 of the APIC register, or equivalently in
1627 byte 35 of struct kvm_lapic_state's regs field. KVM_GET_LAPIC must then
1628 be called after MSR_IA32_APICBASE has been set with KVM_SET_MSR.
1630 If KVM_X2APIC_API_USE_32BIT_IDS feature is disabled, struct kvm_lapic_state
1631 always uses xAPIC format.
1636 Capability: KVM_CAP_IRQCHIP
1639 Parameters: struct kvm_lapic_state (in)
1640 Returns: 0 on success, -1 on error
1642 #define KVM_APIC_REG_SIZE 0x400
1643 struct kvm_lapic_state {
1644 char regs[KVM_APIC_REG_SIZE];
1647 Copies the input argument into the Local APIC registers. The data format
1648 and layout are the same as documented in the architecture manual.
1650 The format of the APIC ID register (bytes 32-35 of struct kvm_lapic_state's
1651 regs field) depends on the state of the KVM_CAP_X2APIC_API capability.
1652 See the note in KVM_GET_LAPIC.
1657 Capability: KVM_CAP_IOEVENTFD
1660 Parameters: struct kvm_ioeventfd (in)
1661 Returns: 0 on success, !0 on error
1663 This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address
1664 within the guest. A guest write in the registered address will signal the
1665 provided event instead of triggering an exit.
1667 struct kvm_ioeventfd {
1669 __u64 addr; /* legal pio/mmio address */
1670 __u32 len; /* 0, 1, 2, 4, or 8 bytes */
1676 For the special case of virtio-ccw devices on s390, the ioevent is matched
1677 to a subchannel/virtqueue tuple instead.
1679 The following flags are defined:
1681 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch)
1682 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio)
1683 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign)
1684 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \
1685 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify)
1687 If datamatch flag is set, the event will be signaled only if the written value
1688 to the registered address is equal to datamatch in struct kvm_ioeventfd.
1690 For virtio-ccw devices, addr contains the subchannel id and datamatch the
1693 With KVM_CAP_IOEVENTFD_ANY_LENGTH, a zero length ioeventfd is allowed, and
1694 the kernel will ignore the length of guest write and may get a faster vmexit.
1695 The speedup may only apply to specific architectures, but the ioeventfd will
1700 Capability: KVM_CAP_SW_TLB
1703 Parameters: struct kvm_dirty_tlb (in)
1704 Returns: 0 on success, -1 on error
1706 struct kvm_dirty_tlb {
1711 This must be called whenever userspace has changed an entry in the shared
1712 TLB, prior to calling KVM_RUN on the associated vcpu.
1714 The "bitmap" field is the userspace address of an array. This array
1715 consists of a number of bits, equal to the total number of TLB entries as
1716 determined by the last successful call to KVM_CONFIG_TLB, rounded up to the
1717 nearest multiple of 64.
1719 Each bit corresponds to one TLB entry, ordered the same as in the shared TLB
1722 The array is little-endian: the bit 0 is the least significant bit of the
1723 first byte, bit 8 is the least significant bit of the second byte, etc.
1724 This avoids any complications with differing word sizes.
1726 The "num_dirty" field is a performance hint for KVM to determine whether it
1727 should skip processing the bitmap and just invalidate everything. It must
1728 be set to the number of set bits in the bitmap.
1731 4.61 KVM_ASSIGN_SET_INTX_MASK (deprecated)
1733 Capability: KVM_CAP_PCI_2_3
1736 Parameters: struct kvm_assigned_pci_dev (in)
1737 Returns: 0 on success, -1 on error
1739 Allows userspace to mask PCI INTx interrupts from the assigned device. The
1740 kernel will not deliver INTx interrupts to the guest between setting and
1741 clearing of KVM_ASSIGN_SET_INTX_MASK via this interface. This enables use of
1742 and emulation of PCI 2.3 INTx disable command register behavior.
1744 This may be used for both PCI 2.3 devices supporting INTx disable natively and
1745 older devices lacking this support. Userspace is responsible for emulating the
1746 read value of the INTx disable bit in the guest visible PCI command register.
1747 When modifying the INTx disable state, userspace should precede updating the
1748 physical device command register by calling this ioctl to inform the kernel of
1749 the new intended INTx mask state.
1751 Note that the kernel uses the device INTx disable bit to internally manage the
1752 device interrupt state for PCI 2.3 devices. Reads of this register may
1753 therefore not match the expected value. Writes should always use the guest
1754 intended INTx disable value rather than attempting to read-copy-update the
1755 current physical device state. Races between user and kernel updates to the
1756 INTx disable bit are handled lazily in the kernel. It's possible the device
1757 may generate unintended interrupts, but they will not be injected into the
1760 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1761 by assigned_dev_id. In the flags field, only KVM_DEV_ASSIGN_MASK_INTX is
1765 4.62 KVM_CREATE_SPAPR_TCE
1767 Capability: KVM_CAP_SPAPR_TCE
1768 Architectures: powerpc
1770 Parameters: struct kvm_create_spapr_tce (in)
1771 Returns: file descriptor for manipulating the created TCE table
1773 This creates a virtual TCE (translation control entry) table, which
1774 is an IOMMU for PAPR-style virtual I/O. It is used to translate
1775 logical addresses used in virtual I/O into guest physical addresses,
1776 and provides a scatter/gather capability for PAPR virtual I/O.
1778 /* for KVM_CAP_SPAPR_TCE */
1779 struct kvm_create_spapr_tce {
1784 The liobn field gives the logical IO bus number for which to create a
1785 TCE table. The window_size field specifies the size of the DMA window
1786 which this TCE table will translate - the table will contain one 64
1787 bit TCE entry for every 4kiB of the DMA window.
1789 When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE
1790 table has been created using this ioctl(), the kernel will handle it
1791 in real mode, updating the TCE table. H_PUT_TCE calls for other
1792 liobns will cause a vm exit and must be handled by userspace.
1794 The return value is a file descriptor which can be passed to mmap(2)
1795 to map the created TCE table into userspace. This lets userspace read
1796 the entries written by kernel-handled H_PUT_TCE calls, and also lets
1797 userspace update the TCE table directly which is useful in some
1801 4.63 KVM_ALLOCATE_RMA
1803 Capability: KVM_CAP_PPC_RMA
1804 Architectures: powerpc
1806 Parameters: struct kvm_allocate_rma (out)
1807 Returns: file descriptor for mapping the allocated RMA
1809 This allocates a Real Mode Area (RMA) from the pool allocated at boot
1810 time by the kernel. An RMA is a physically-contiguous, aligned region
1811 of memory used on older POWER processors to provide the memory which
1812 will be accessed by real-mode (MMU off) accesses in a KVM guest.
1813 POWER processors support a set of sizes for the RMA that usually
1814 includes 64MB, 128MB, 256MB and some larger powers of two.
1816 /* for KVM_ALLOCATE_RMA */
1817 struct kvm_allocate_rma {
1821 The return value is a file descriptor which can be passed to mmap(2)
1822 to map the allocated RMA into userspace. The mapped area can then be
1823 passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the
1824 RMA for a virtual machine. The size of the RMA in bytes (which is
1825 fixed at host kernel boot time) is returned in the rma_size field of
1826 the argument structure.
1828 The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl
1829 is supported; 2 if the processor requires all virtual machines to have
1830 an RMA, or 1 if the processor can use an RMA but doesn't require it,
1831 because it supports the Virtual RMA (VRMA) facility.
1836 Capability: KVM_CAP_USER_NMI
1840 Returns: 0 on success, -1 on error
1842 Queues an NMI on the thread's vcpu. Note this is well defined only
1843 when KVM_CREATE_IRQCHIP has not been called, since this is an interface
1844 between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP
1845 has been called, this interface is completely emulated within the kernel.
1847 To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the
1848 following algorithm:
1851 - read the local APIC's state (KVM_GET_LAPIC)
1852 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1)
1853 - if so, issue KVM_NMI
1856 Some guests configure the LINT1 NMI input to cause a panic, aiding in
1860 4.65 KVM_S390_UCAS_MAP
1862 Capability: KVM_CAP_S390_UCONTROL
1865 Parameters: struct kvm_s390_ucas_mapping (in)
1866 Returns: 0 in case of success
1868 The parameter is defined like this:
1869 struct kvm_s390_ucas_mapping {
1875 This ioctl maps the memory at "user_addr" with the length "length" to
1876 the vcpu's address space starting at "vcpu_addr". All parameters need to
1877 be aligned by 1 megabyte.
1880 4.66 KVM_S390_UCAS_UNMAP
1882 Capability: KVM_CAP_S390_UCONTROL
1885 Parameters: struct kvm_s390_ucas_mapping (in)
1886 Returns: 0 in case of success
1888 The parameter is defined like this:
1889 struct kvm_s390_ucas_mapping {
1895 This ioctl unmaps the memory in the vcpu's address space starting at
1896 "vcpu_addr" with the length "length". The field "user_addr" is ignored.
1897 All parameters need to be aligned by 1 megabyte.
1900 4.67 KVM_S390_VCPU_FAULT
1902 Capability: KVM_CAP_S390_UCONTROL
1905 Parameters: vcpu absolute address (in)
1906 Returns: 0 in case of success
1908 This call creates a page table entry on the virtual cpu's address space
1909 (for user controlled virtual machines) or the virtual machine's address
1910 space (for regular virtual machines). This only works for minor faults,
1911 thus it's recommended to access subject memory page via the user page
1912 table upfront. This is useful to handle validity intercepts for user
1913 controlled virtual machines to fault in the virtual cpu's lowcore pages
1914 prior to calling the KVM_RUN ioctl.
1917 4.68 KVM_SET_ONE_REG
1919 Capability: KVM_CAP_ONE_REG
1922 Parameters: struct kvm_one_reg (in)
1923 Returns: 0 on success, negative value on failure
1925 struct kvm_one_reg {
1930 Using this ioctl, a single vcpu register can be set to a specific value
1931 defined by user space with the passed in struct kvm_one_reg, where id
1932 refers to the register identifier as described below and addr is a pointer
1933 to a variable with the respective size. There can be architecture agnostic
1934 and architecture specific registers. Each have their own range of operation
1935 and their own constants and width. To keep track of the implemented
1936 registers, find a list below:
1938 Arch | Register | Width (bits)
1940 PPC | KVM_REG_PPC_HIOR | 64
1941 PPC | KVM_REG_PPC_IAC1 | 64
1942 PPC | KVM_REG_PPC_IAC2 | 64
1943 PPC | KVM_REG_PPC_IAC3 | 64
1944 PPC | KVM_REG_PPC_IAC4 | 64
1945 PPC | KVM_REG_PPC_DAC1 | 64
1946 PPC | KVM_REG_PPC_DAC2 | 64
1947 PPC | KVM_REG_PPC_DABR | 64
1948 PPC | KVM_REG_PPC_DSCR | 64
1949 PPC | KVM_REG_PPC_PURR | 64
1950 PPC | KVM_REG_PPC_SPURR | 64
1951 PPC | KVM_REG_PPC_DAR | 64
1952 PPC | KVM_REG_PPC_DSISR | 32
1953 PPC | KVM_REG_PPC_AMR | 64
1954 PPC | KVM_REG_PPC_UAMOR | 64
1955 PPC | KVM_REG_PPC_MMCR0 | 64
1956 PPC | KVM_REG_PPC_MMCR1 | 64
1957 PPC | KVM_REG_PPC_MMCRA | 64
1958 PPC | KVM_REG_PPC_MMCR2 | 64
1959 PPC | KVM_REG_PPC_MMCRS | 64
1960 PPC | KVM_REG_PPC_SIAR | 64
1961 PPC | KVM_REG_PPC_SDAR | 64
1962 PPC | KVM_REG_PPC_SIER | 64
1963 PPC | KVM_REG_PPC_PMC1 | 32
1964 PPC | KVM_REG_PPC_PMC2 | 32
1965 PPC | KVM_REG_PPC_PMC3 | 32
1966 PPC | KVM_REG_PPC_PMC4 | 32
1967 PPC | KVM_REG_PPC_PMC5 | 32
1968 PPC | KVM_REG_PPC_PMC6 | 32
1969 PPC | KVM_REG_PPC_PMC7 | 32
1970 PPC | KVM_REG_PPC_PMC8 | 32
1971 PPC | KVM_REG_PPC_FPR0 | 64
1973 PPC | KVM_REG_PPC_FPR31 | 64
1974 PPC | KVM_REG_PPC_VR0 | 128
1976 PPC | KVM_REG_PPC_VR31 | 128
1977 PPC | KVM_REG_PPC_VSR0 | 128
1979 PPC | KVM_REG_PPC_VSR31 | 128
1980 PPC | KVM_REG_PPC_FPSCR | 64
1981 PPC | KVM_REG_PPC_VSCR | 32
1982 PPC | KVM_REG_PPC_VPA_ADDR | 64
1983 PPC | KVM_REG_PPC_VPA_SLB | 128
1984 PPC | KVM_REG_PPC_VPA_DTL | 128
1985 PPC | KVM_REG_PPC_EPCR | 32
1986 PPC | KVM_REG_PPC_EPR | 32
1987 PPC | KVM_REG_PPC_TCR | 32
1988 PPC | KVM_REG_PPC_TSR | 32
1989 PPC | KVM_REG_PPC_OR_TSR | 32
1990 PPC | KVM_REG_PPC_CLEAR_TSR | 32
1991 PPC | KVM_REG_PPC_MAS0 | 32
1992 PPC | KVM_REG_PPC_MAS1 | 32
1993 PPC | KVM_REG_PPC_MAS2 | 64
1994 PPC | KVM_REG_PPC_MAS7_3 | 64
1995 PPC | KVM_REG_PPC_MAS4 | 32
1996 PPC | KVM_REG_PPC_MAS6 | 32
1997 PPC | KVM_REG_PPC_MMUCFG | 32
1998 PPC | KVM_REG_PPC_TLB0CFG | 32
1999 PPC | KVM_REG_PPC_TLB1CFG | 32
2000 PPC | KVM_REG_PPC_TLB2CFG | 32
2001 PPC | KVM_REG_PPC_TLB3CFG | 32
2002 PPC | KVM_REG_PPC_TLB0PS | 32
2003 PPC | KVM_REG_PPC_TLB1PS | 32
2004 PPC | KVM_REG_PPC_TLB2PS | 32
2005 PPC | KVM_REG_PPC_TLB3PS | 32
2006 PPC | KVM_REG_PPC_EPTCFG | 32
2007 PPC | KVM_REG_PPC_ICP_STATE | 64
2008 PPC | KVM_REG_PPC_TB_OFFSET | 64
2009 PPC | KVM_REG_PPC_SPMC1 | 32
2010 PPC | KVM_REG_PPC_SPMC2 | 32
2011 PPC | KVM_REG_PPC_IAMR | 64
2012 PPC | KVM_REG_PPC_TFHAR | 64
2013 PPC | KVM_REG_PPC_TFIAR | 64
2014 PPC | KVM_REG_PPC_TEXASR | 64
2015 PPC | KVM_REG_PPC_FSCR | 64
2016 PPC | KVM_REG_PPC_PSPB | 32
2017 PPC | KVM_REG_PPC_EBBHR | 64
2018 PPC | KVM_REG_PPC_EBBRR | 64
2019 PPC | KVM_REG_PPC_BESCR | 64
2020 PPC | KVM_REG_PPC_TAR | 64
2021 PPC | KVM_REG_PPC_DPDES | 64
2022 PPC | KVM_REG_PPC_DAWR | 64
2023 PPC | KVM_REG_PPC_DAWRX | 64
2024 PPC | KVM_REG_PPC_CIABR | 64
2025 PPC | KVM_REG_PPC_IC | 64
2026 PPC | KVM_REG_PPC_VTB | 64
2027 PPC | KVM_REG_PPC_CSIGR | 64
2028 PPC | KVM_REG_PPC_TACR | 64
2029 PPC | KVM_REG_PPC_TCSCR | 64
2030 PPC | KVM_REG_PPC_PID | 64
2031 PPC | KVM_REG_PPC_ACOP | 64
2032 PPC | KVM_REG_PPC_VRSAVE | 32
2033 PPC | KVM_REG_PPC_LPCR | 32
2034 PPC | KVM_REG_PPC_LPCR_64 | 64
2035 PPC | KVM_REG_PPC_PPR | 64
2036 PPC | KVM_REG_PPC_ARCH_COMPAT | 32
2037 PPC | KVM_REG_PPC_DABRX | 32
2038 PPC | KVM_REG_PPC_WORT | 64
2039 PPC | KVM_REG_PPC_SPRG9 | 64
2040 PPC | KVM_REG_PPC_DBSR | 32
2041 PPC | KVM_REG_PPC_TIDR | 64
2042 PPC | KVM_REG_PPC_PSSCR | 64
2043 PPC | KVM_REG_PPC_TM_GPR0 | 64
2045 PPC | KVM_REG_PPC_TM_GPR31 | 64
2046 PPC | KVM_REG_PPC_TM_VSR0 | 128
2048 PPC | KVM_REG_PPC_TM_VSR63 | 128
2049 PPC | KVM_REG_PPC_TM_CR | 64
2050 PPC | KVM_REG_PPC_TM_LR | 64
2051 PPC | KVM_REG_PPC_TM_CTR | 64
2052 PPC | KVM_REG_PPC_TM_FPSCR | 64
2053 PPC | KVM_REG_PPC_TM_AMR | 64
2054 PPC | KVM_REG_PPC_TM_PPR | 64
2055 PPC | KVM_REG_PPC_TM_VRSAVE | 64
2056 PPC | KVM_REG_PPC_TM_VSCR | 32
2057 PPC | KVM_REG_PPC_TM_DSCR | 64
2058 PPC | KVM_REG_PPC_TM_TAR | 64
2059 PPC | KVM_REG_PPC_TM_XER | 64
2061 MIPS | KVM_REG_MIPS_R0 | 64
2063 MIPS | KVM_REG_MIPS_R31 | 64
2064 MIPS | KVM_REG_MIPS_HI | 64
2065 MIPS | KVM_REG_MIPS_LO | 64
2066 MIPS | KVM_REG_MIPS_PC | 64
2067 MIPS | KVM_REG_MIPS_CP0_INDEX | 32
2068 MIPS | KVM_REG_MIPS_CP0_ENTRYLO0 | 64
2069 MIPS | KVM_REG_MIPS_CP0_ENTRYLO1 | 64
2070 MIPS | KVM_REG_MIPS_CP0_CONTEXT | 64
2071 MIPS | KVM_REG_MIPS_CP0_USERLOCAL | 64
2072 MIPS | KVM_REG_MIPS_CP0_PAGEMASK | 32
2073 MIPS | KVM_REG_MIPS_CP0_WIRED | 32
2074 MIPS | KVM_REG_MIPS_CP0_HWRENA | 32
2075 MIPS | KVM_REG_MIPS_CP0_BADVADDR | 64
2076 MIPS | KVM_REG_MIPS_CP0_COUNT | 32
2077 MIPS | KVM_REG_MIPS_CP0_ENTRYHI | 64
2078 MIPS | KVM_REG_MIPS_CP0_COMPARE | 32
2079 MIPS | KVM_REG_MIPS_CP0_STATUS | 32
2080 MIPS | KVM_REG_MIPS_CP0_INTCTL | 32
2081 MIPS | KVM_REG_MIPS_CP0_CAUSE | 32
2082 MIPS | KVM_REG_MIPS_CP0_EPC | 64
2083 MIPS | KVM_REG_MIPS_CP0_PRID | 32
2084 MIPS | KVM_REG_MIPS_CP0_EBASE | 64
2085 MIPS | KVM_REG_MIPS_CP0_CONFIG | 32
2086 MIPS | KVM_REG_MIPS_CP0_CONFIG1 | 32
2087 MIPS | KVM_REG_MIPS_CP0_CONFIG2 | 32
2088 MIPS | KVM_REG_MIPS_CP0_CONFIG3 | 32
2089 MIPS | KVM_REG_MIPS_CP0_CONFIG4 | 32
2090 MIPS | KVM_REG_MIPS_CP0_CONFIG5 | 32
2091 MIPS | KVM_REG_MIPS_CP0_CONFIG7 | 32
2092 MIPS | KVM_REG_MIPS_CP0_ERROREPC | 64
2093 MIPS | KVM_REG_MIPS_CP0_KSCRATCH1 | 64
2094 MIPS | KVM_REG_MIPS_CP0_KSCRATCH2 | 64
2095 MIPS | KVM_REG_MIPS_CP0_KSCRATCH3 | 64
2096 MIPS | KVM_REG_MIPS_CP0_KSCRATCH4 | 64
2097 MIPS | KVM_REG_MIPS_CP0_KSCRATCH5 | 64
2098 MIPS | KVM_REG_MIPS_CP0_KSCRATCH6 | 64
2099 MIPS | KVM_REG_MIPS_COUNT_CTL | 64
2100 MIPS | KVM_REG_MIPS_COUNT_RESUME | 64
2101 MIPS | KVM_REG_MIPS_COUNT_HZ | 64
2102 MIPS | KVM_REG_MIPS_FPR_32(0..31) | 32
2103 MIPS | KVM_REG_MIPS_FPR_64(0..31) | 64
2104 MIPS | KVM_REG_MIPS_VEC_128(0..31) | 128
2105 MIPS | KVM_REG_MIPS_FCR_IR | 32
2106 MIPS | KVM_REG_MIPS_FCR_CSR | 32
2107 MIPS | KVM_REG_MIPS_MSA_IR | 32
2108 MIPS | KVM_REG_MIPS_MSA_CSR | 32
2110 ARM registers are mapped using the lower 32 bits. The upper 16 of that
2111 is the register group type, or coprocessor number:
2113 ARM core registers have the following id bit patterns:
2114 0x4020 0000 0010 <index into the kvm_regs struct:16>
2116 ARM 32-bit CP15 registers have the following id bit patterns:
2117 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3>
2119 ARM 64-bit CP15 registers have the following id bit patterns:
2120 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3>
2122 ARM CCSIDR registers are demultiplexed by CSSELR value:
2123 0x4020 0000 0011 00 <csselr:8>
2125 ARM 32-bit VFP control registers have the following id bit patterns:
2126 0x4020 0000 0012 1 <regno:12>
2128 ARM 64-bit FP registers have the following id bit patterns:
2129 0x4030 0000 0012 0 <regno:12>
2132 arm64 registers are mapped using the lower 32 bits. The upper 16 of
2133 that is the register group type, or coprocessor number:
2135 arm64 core/FP-SIMD registers have the following id bit patterns. Note
2136 that the size of the access is variable, as the kvm_regs structure
2137 contains elements ranging from 32 to 128 bits. The index is a 32bit
2138 value in the kvm_regs structure seen as a 32bit array.
2139 0x60x0 0000 0010 <index into the kvm_regs struct:16>
2141 arm64 CCSIDR registers are demultiplexed by CSSELR value:
2142 0x6020 0000 0011 00 <csselr:8>
2144 arm64 system registers have the following id bit patterns:
2145 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3>
2148 MIPS registers are mapped using the lower 32 bits. The upper 16 of that is
2149 the register group type:
2151 MIPS core registers (see above) have the following id bit patterns:
2152 0x7030 0000 0000 <reg:16>
2154 MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit
2155 patterns depending on whether they're 32-bit or 64-bit registers:
2156 0x7020 0000 0001 00 <reg:5> <sel:3> (32-bit)
2157 0x7030 0000 0001 00 <reg:5> <sel:3> (64-bit)
2159 Note: KVM_REG_MIPS_CP0_ENTRYLO0 and KVM_REG_MIPS_CP0_ENTRYLO1 are the MIPS64
2160 versions of the EntryLo registers regardless of the word size of the host
2161 hardware, host kernel, guest, and whether XPA is present in the guest, i.e.
2162 with the RI and XI bits (if they exist) in bits 63 and 62 respectively, and
2163 the PFNX field starting at bit 30.
2165 MIPS KVM control registers (see above) have the following id bit patterns:
2166 0x7030 0000 0002 <reg:16>
2168 MIPS FPU registers (see KVM_REG_MIPS_FPR_{32,64}() above) have the following
2169 id bit patterns depending on the size of the register being accessed. They are
2170 always accessed according to the current guest FPU mode (Status.FR and
2171 Config5.FRE), i.e. as the guest would see them, and they become unpredictable
2172 if the guest FPU mode is changed. MIPS SIMD Architecture (MSA) vector
2173 registers (see KVM_REG_MIPS_VEC_128() above) have similar patterns as they
2174 overlap the FPU registers:
2175 0x7020 0000 0003 00 <0:3> <reg:5> (32-bit FPU registers)
2176 0x7030 0000 0003 00 <0:3> <reg:5> (64-bit FPU registers)
2177 0x7040 0000 0003 00 <0:3> <reg:5> (128-bit MSA vector registers)
2179 MIPS FPU control registers (see KVM_REG_MIPS_FCR_{IR,CSR} above) have the
2180 following id bit patterns:
2181 0x7020 0000 0003 01 <0:3> <reg:5>
2183 MIPS MSA control registers (see KVM_REG_MIPS_MSA_{IR,CSR} above) have the
2184 following id bit patterns:
2185 0x7020 0000 0003 02 <0:3> <reg:5>
2188 4.69 KVM_GET_ONE_REG
2190 Capability: KVM_CAP_ONE_REG
2193 Parameters: struct kvm_one_reg (in and out)
2194 Returns: 0 on success, negative value on failure
2196 This ioctl allows to receive the value of a single register implemented
2197 in a vcpu. The register to read is indicated by the "id" field of the
2198 kvm_one_reg struct passed in. On success, the register value can be found
2199 at the memory location pointed to by "addr".
2201 The list of registers accessible using this interface is identical to the
2205 4.70 KVM_KVMCLOCK_CTRL
2207 Capability: KVM_CAP_KVMCLOCK_CTRL
2208 Architectures: Any that implement pvclocks (currently x86 only)
2211 Returns: 0 on success, -1 on error
2213 This signals to the host kernel that the specified guest is being paused by
2214 userspace. The host will set a flag in the pvclock structure that is checked
2215 from the soft lockup watchdog. The flag is part of the pvclock structure that
2216 is shared between guest and host, specifically the second bit of the flags
2217 field of the pvclock_vcpu_time_info structure. It will be set exclusively by
2218 the host and read/cleared exclusively by the guest. The guest operation of
2219 checking and clearing the flag must an atomic operation so
2220 load-link/store-conditional, or equivalent must be used. There are two cases
2221 where the guest will clear the flag: when the soft lockup watchdog timer resets
2222 itself or when a soft lockup is detected. This ioctl can be called any time
2223 after pausing the vcpu, but before it is resumed.
2228 Capability: KVM_CAP_SIGNAL_MSI
2229 Architectures: x86 arm arm64
2231 Parameters: struct kvm_msi (in)
2232 Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error
2234 Directly inject a MSI message. Only valid with in-kernel irqchip that handles
2246 flags: KVM_MSI_VALID_DEVID: devid contains a valid value. The per-VM
2247 KVM_CAP_MSI_DEVID capability advertises the requirement to provide
2248 the device ID. If this capability is not available, userspace
2249 should never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail.
2251 If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier
2252 for the device that wrote the MSI message. For PCI, this is usually a
2253 BFD identifier in the lower 16 bits.
2255 On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS
2256 feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled,
2257 address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of
2258 address_hi must be zero.
2261 4.71 KVM_CREATE_PIT2
2263 Capability: KVM_CAP_PIT2
2266 Parameters: struct kvm_pit_config (in)
2267 Returns: 0 on success, -1 on error
2269 Creates an in-kernel device model for the i8254 PIT. This call is only valid
2270 after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following
2271 parameters have to be passed:
2273 struct kvm_pit_config {
2280 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */
2282 PIT timer interrupts may use a per-VM kernel thread for injection. If it
2283 exists, this thread will have a name of the following pattern:
2285 kvm-pit/<owner-process-pid>
2287 When running a guest with elevated priorities, the scheduling parameters of
2288 this thread may have to be adjusted accordingly.
2290 This IOCTL replaces the obsolete KVM_CREATE_PIT.
2295 Capability: KVM_CAP_PIT_STATE2
2298 Parameters: struct kvm_pit_state2 (out)
2299 Returns: 0 on success, -1 on error
2301 Retrieves the state of the in-kernel PIT model. Only valid after
2302 KVM_CREATE_PIT2. The state is returned in the following structure:
2304 struct kvm_pit_state2 {
2305 struct kvm_pit_channel_state channels[3];
2312 /* disable PIT in HPET legacy mode */
2313 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001
2315 This IOCTL replaces the obsolete KVM_GET_PIT.
2320 Capability: KVM_CAP_PIT_STATE2
2323 Parameters: struct kvm_pit_state2 (in)
2324 Returns: 0 on success, -1 on error
2326 Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2.
2327 See KVM_GET_PIT2 for details on struct kvm_pit_state2.
2329 This IOCTL replaces the obsolete KVM_SET_PIT.
2332 4.74 KVM_PPC_GET_SMMU_INFO
2334 Capability: KVM_CAP_PPC_GET_SMMU_INFO
2335 Architectures: powerpc
2338 Returns: 0 on success, -1 on error
2340 This populates and returns a structure describing the features of
2341 the "Server" class MMU emulation supported by KVM.
2342 This can in turn be used by userspace to generate the appropriate
2343 device-tree properties for the guest operating system.
2345 The structure contains some global information, followed by an
2346 array of supported segment page sizes:
2348 struct kvm_ppc_smmu_info {
2352 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ];
2355 The supported flags are:
2357 - KVM_PPC_PAGE_SIZES_REAL:
2358 When that flag is set, guest page sizes must "fit" the backing
2359 store page sizes. When not set, any page size in the list can
2360 be used regardless of how they are backed by userspace.
2362 - KVM_PPC_1T_SEGMENTS
2363 The emulated MMU supports 1T segments in addition to the
2366 The "slb_size" field indicates how many SLB entries are supported
2368 The "sps" array contains 8 entries indicating the supported base
2369 page sizes for a segment in increasing order. Each entry is defined
2372 struct kvm_ppc_one_seg_page_size {
2373 __u32 page_shift; /* Base page shift of segment (or 0) */
2374 __u32 slb_enc; /* SLB encoding for BookS */
2375 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ];
2378 An entry with a "page_shift" of 0 is unused. Because the array is
2379 organized in increasing order, a lookup can stop when encoutering
2382 The "slb_enc" field provides the encoding to use in the SLB for the
2383 page size. The bits are in positions such as the value can directly
2384 be OR'ed into the "vsid" argument of the slbmte instruction.
2386 The "enc" array is a list which for each of those segment base page
2387 size provides the list of supported actual page sizes (which can be
2388 only larger or equal to the base page size), along with the
2389 corresponding encoding in the hash PTE. Similarly, the array is
2390 8 entries sorted by increasing sizes and an entry with a "0" shift
2391 is an empty entry and a terminator:
2393 struct kvm_ppc_one_page_size {
2394 __u32 page_shift; /* Page shift (or 0) */
2395 __u32 pte_enc; /* Encoding in the HPTE (>>12) */
2398 The "pte_enc" field provides a value that can OR'ed into the hash
2399 PTE's RPN field (ie, it needs to be shifted left by 12 to OR it
2400 into the hash PTE second double word).
2404 Capability: KVM_CAP_IRQFD
2405 Architectures: x86 s390 arm arm64
2407 Parameters: struct kvm_irqfd (in)
2408 Returns: 0 on success, -1 on error
2410 Allows setting an eventfd to directly trigger a guest interrupt.
2411 kvm_irqfd.fd specifies the file descriptor to use as the eventfd and
2412 kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When
2413 an event is triggered on the eventfd, an interrupt is injected into
2414 the guest using the specified gsi pin. The irqfd is removed using
2415 the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd
2418 With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify
2419 mechanism allowing emulation of level-triggered, irqfd-based
2420 interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an
2421 additional eventfd in the kvm_irqfd.resamplefd field. When operating
2422 in resample mode, posting of an interrupt through kvm_irq.fd asserts
2423 the specified gsi in the irqchip. When the irqchip is resampled, such
2424 as from an EOI, the gsi is de-asserted and the user is notified via
2425 kvm_irqfd.resamplefd. It is the user's responsibility to re-queue
2426 the interrupt if the device making use of it still requires service.
2427 Note that closing the resamplefd is not sufficient to disable the
2428 irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment
2429 and need not be specified with KVM_IRQFD_FLAG_DEASSIGN.
2431 On arm/arm64, gsi routing being supported, the following can happen:
2432 - in case no routing entry is associated to this gsi, injection fails
2433 - in case the gsi is associated to an irqchip routing entry,
2434 irqchip.pin + 32 corresponds to the injected SPI ID.
2435 - in case the gsi is associated to an MSI routing entry, the MSI
2436 message and device ID are translated into an LPI (support restricted
2437 to GICv3 ITS in-kernel emulation).
2439 4.76 KVM_PPC_ALLOCATE_HTAB
2441 Capability: KVM_CAP_PPC_ALLOC_HTAB
2442 Architectures: powerpc
2444 Parameters: Pointer to u32 containing hash table order (in/out)
2445 Returns: 0 on success, -1 on error
2447 This requests the host kernel to allocate an MMU hash table for a
2448 guest using the PAPR paravirtualization interface. This only does
2449 anything if the kernel is configured to use the Book 3S HV style of
2450 virtualization. Otherwise the capability doesn't exist and the ioctl
2451 returns an ENOTTY error. The rest of this description assumes Book 3S
2454 There must be no vcpus running when this ioctl is called; if there
2455 are, it will do nothing and return an EBUSY error.
2457 The parameter is a pointer to a 32-bit unsigned integer variable
2458 containing the order (log base 2) of the desired size of the hash
2459 table, which must be between 18 and 46. On successful return from the
2460 ioctl, the value will not be changed by the kernel.
2462 If no hash table has been allocated when any vcpu is asked to run
2463 (with the KVM_RUN ioctl), the host kernel will allocate a
2464 default-sized hash table (16 MB).
2466 If this ioctl is called when a hash table has already been allocated,
2467 with a different order from the existing hash table, the existing hash
2468 table will be freed and a new one allocated. If this is ioctl is
2469 called when a hash table has already been allocated of the same order
2470 as specified, the kernel will clear out the existing hash table (zero
2471 all HPTEs). In either case, if the guest is using the virtualized
2472 real-mode area (VRMA) facility, the kernel will re-create the VMRA
2473 HPTEs on the next KVM_RUN of any vcpu.
2475 4.77 KVM_S390_INTERRUPT
2479 Type: vm ioctl, vcpu ioctl
2480 Parameters: struct kvm_s390_interrupt (in)
2481 Returns: 0 on success, -1 on error
2483 Allows to inject an interrupt to the guest. Interrupts can be floating
2484 (vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type.
2486 Interrupt parameters are passed via kvm_s390_interrupt:
2488 struct kvm_s390_interrupt {
2494 type can be one of the following:
2496 KVM_S390_SIGP_STOP (vcpu) - sigp stop; optional flags in parm
2497 KVM_S390_PROGRAM_INT (vcpu) - program check; code in parm
2498 KVM_S390_SIGP_SET_PREFIX (vcpu) - sigp set prefix; prefix address in parm
2499 KVM_S390_RESTART (vcpu) - restart
2500 KVM_S390_INT_CLOCK_COMP (vcpu) - clock comparator interrupt
2501 KVM_S390_INT_CPU_TIMER (vcpu) - CPU timer interrupt
2502 KVM_S390_INT_VIRTIO (vm) - virtio external interrupt; external interrupt
2503 parameters in parm and parm64
2504 KVM_S390_INT_SERVICE (vm) - sclp external interrupt; sclp parameter in parm
2505 KVM_S390_INT_EMERGENCY (vcpu) - sigp emergency; source cpu in parm
2506 KVM_S390_INT_EXTERNAL_CALL (vcpu) - sigp external call; source cpu in parm
2507 KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) - compound value to indicate an
2508 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel);
2509 I/O interruption parameters in parm (subchannel) and parm64 (intparm,
2510 interruption subclass)
2511 KVM_S390_MCHK (vm, vcpu) - machine check interrupt; cr 14 bits in parm,
2512 machine check interrupt code in parm64 (note that
2513 machine checks needing further payload are not
2514 supported by this ioctl)
2516 Note that the vcpu ioctl is asynchronous to vcpu execution.
2518 4.78 KVM_PPC_GET_HTAB_FD
2520 Capability: KVM_CAP_PPC_HTAB_FD
2521 Architectures: powerpc
2523 Parameters: Pointer to struct kvm_get_htab_fd (in)
2524 Returns: file descriptor number (>= 0) on success, -1 on error
2526 This returns a file descriptor that can be used either to read out the
2527 entries in the guest's hashed page table (HPT), or to write entries to
2528 initialize the HPT. The returned fd can only be written to if the
2529 KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and
2530 can only be read if that bit is clear. The argument struct looks like
2533 /* For KVM_PPC_GET_HTAB_FD */
2534 struct kvm_get_htab_fd {
2540 /* Values for kvm_get_htab_fd.flags */
2541 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1)
2542 #define KVM_GET_HTAB_WRITE ((__u64)0x2)
2544 The `start_index' field gives the index in the HPT of the entry at
2545 which to start reading. It is ignored when writing.
2547 Reads on the fd will initially supply information about all
2548 "interesting" HPT entries. Interesting entries are those with the
2549 bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise
2550 all entries. When the end of the HPT is reached, the read() will
2551 return. If read() is called again on the fd, it will start again from
2552 the beginning of the HPT, but will only return HPT entries that have
2553 changed since they were last read.
2555 Data read or written is structured as a header (8 bytes) followed by a
2556 series of valid HPT entries (16 bytes) each. The header indicates how
2557 many valid HPT entries there are and how many invalid entries follow
2558 the valid entries. The invalid entries are not represented explicitly
2559 in the stream. The header format is:
2561 struct kvm_get_htab_header {
2567 Writes to the fd create HPT entries starting at the index given in the
2568 header; first `n_valid' valid entries with contents from the data
2569 written, then `n_invalid' invalid entries, invalidating any previously
2570 valid entries found.
2572 4.79 KVM_CREATE_DEVICE
2574 Capability: KVM_CAP_DEVICE_CTRL
2576 Parameters: struct kvm_create_device (in/out)
2577 Returns: 0 on success, -1 on error
2579 ENODEV: The device type is unknown or unsupported
2580 EEXIST: Device already created, and this type of device may not
2581 be instantiated multiple times
2583 Other error conditions may be defined by individual device types or
2584 have their standard meanings.
2586 Creates an emulated device in the kernel. The file descriptor returned
2587 in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR.
2589 If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the
2590 device type is supported (not necessarily whether it can be created
2593 Individual devices should not define flags. Attributes should be used
2594 for specifying any behavior that is not implied by the device type
2597 struct kvm_create_device {
2598 __u32 type; /* in: KVM_DEV_TYPE_xxx */
2599 __u32 fd; /* out: device handle */
2600 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */
2603 4.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR
2605 Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device,
2606 KVM_CAP_VCPU_ATTRIBUTES for vcpu device
2607 Type: device ioctl, vm ioctl, vcpu ioctl
2608 Parameters: struct kvm_device_attr
2609 Returns: 0 on success, -1 on error
2611 ENXIO: The group or attribute is unknown/unsupported for this device
2612 or hardware support is missing.
2613 EPERM: The attribute cannot (currently) be accessed this way
2614 (e.g. read-only attribute, or attribute that only makes
2615 sense when the device is in a different state)
2617 Other error conditions may be defined by individual device types.
2619 Gets/sets a specified piece of device configuration and/or state. The
2620 semantics are device-specific. See individual device documentation in
2621 the "devices" directory. As with ONE_REG, the size of the data
2622 transferred is defined by the particular attribute.
2624 struct kvm_device_attr {
2625 __u32 flags; /* no flags currently defined */
2626 __u32 group; /* device-defined */
2627 __u64 attr; /* group-defined */
2628 __u64 addr; /* userspace address of attr data */
2631 4.81 KVM_HAS_DEVICE_ATTR
2633 Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device,
2634 KVM_CAP_VCPU_ATTRIBUTES for vcpu device
2635 Type: device ioctl, vm ioctl, vcpu ioctl
2636 Parameters: struct kvm_device_attr
2637 Returns: 0 on success, -1 on error
2639 ENXIO: The group or attribute is unknown/unsupported for this device
2640 or hardware support is missing.
2642 Tests whether a device supports a particular attribute. A successful
2643 return indicates the attribute is implemented. It does not necessarily
2644 indicate that the attribute can be read or written in the device's
2645 current state. "addr" is ignored.
2647 4.82 KVM_ARM_VCPU_INIT
2650 Architectures: arm, arm64
2652 Parameters: struct kvm_vcpu_init (in)
2653 Returns: 0 on success; -1 on error
2655 Â EINVAL: Â Â Â the target is unknown, or the combination of features is invalid.
2656 Â ENOENT: Â Â Â a features bit specified is unknown.
2658 This tells KVM what type of CPU to present to the guest, and what
2659 optional features it should have. Â This will cause a reset of the cpu
2660 registers to their initial values. Â If this is not called, KVM_RUN will
2661 return ENOEXEC for that vcpu.
2663 Note that because some registers reflect machine topology, all vcpus
2664 should be created before this ioctl is invoked.
2666 Userspace can call this function multiple times for a given vcpu, including
2667 after the vcpu has been run. This will reset the vcpu to its initial
2668 state. All calls to this function after the initial call must use the same
2669 target and same set of feature flags, otherwise EINVAL will be returned.
2672 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state.
2673 Depends on KVM_CAP_ARM_PSCI. If not set, the CPU will be powered on
2674 and execute guest code when KVM_RUN is called.
2675 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode.
2676 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only).
2677 - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 for the CPU.
2678 Depends on KVM_CAP_ARM_PSCI_0_2.
2679 - KVM_ARM_VCPU_PMU_V3: Emulate PMUv3 for the CPU.
2680 Depends on KVM_CAP_ARM_PMU_V3.
2683 4.83 KVM_ARM_PREFERRED_TARGET
2686 Architectures: arm, arm64
2688 Parameters: struct struct kvm_vcpu_init (out)
2689 Returns: 0 on success; -1 on error
2691 ENODEV: no preferred target available for the host
2693 This queries KVM for preferred CPU target type which can be emulated
2694 by KVM on underlying host.
2696 The ioctl returns struct kvm_vcpu_init instance containing information
2697 about preferred CPU target type and recommended features for it. The
2698 kvm_vcpu_init->features bitmap returned will have feature bits set if
2699 the preferred target recommends setting these features, but this is
2702 The information returned by this ioctl can be used to prepare an instance
2703 of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in
2704 in VCPU matching underlying host.
2707 4.84 KVM_GET_REG_LIST
2710 Architectures: arm, arm64, mips
2712 Parameters: struct kvm_reg_list (in/out)
2713 Returns: 0 on success; -1 on error
2715 Â E2BIG: Â Â Â Â the reg index list is too big to fit in the array specified by
2716 Â Â Â Â Â Â Â Â Â Â Â Â the user (the number required will be written into n).
2718 struct kvm_reg_list {
2719 __u64 n; /* number of registers in reg[] */
2723 This ioctl returns the guest registers that are supported for the
2724 KVM_GET_ONE_REG/KVM_SET_ONE_REG calls.
2727 4.85 KVM_ARM_SET_DEVICE_ADDR (deprecated)
2729 Capability: KVM_CAP_ARM_SET_DEVICE_ADDR
2730 Architectures: arm, arm64
2732 Parameters: struct kvm_arm_device_address (in)
2733 Returns: 0 on success, -1 on error
2735 ENODEV: The device id is unknown
2736 ENXIO: Device not supported on current system
2737 EEXIST: Address already set
2738 E2BIG: Address outside guest physical address space
2739 EBUSY: Address overlaps with other device range
2741 struct kvm_arm_device_addr {
2746 Specify a device address in the guest's physical address space where guests
2747 can access emulated or directly exposed devices, which the host kernel needs
2748 to know about. The id field is an architecture specific identifier for a
2751 ARM/arm64 divides the id field into two parts, a device id and an
2752 address type id specific to the individual device.
2754 Â bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 |
2755 field: | 0x00000000 | device id | addr type id |
2757 ARM/arm64 currently only require this when using the in-kernel GIC
2758 support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2
2759 as the device id. When setting the base address for the guest's
2760 mapping of the VGIC virtual CPU and distributor interface, the ioctl
2761 must be called after calling KVM_CREATE_IRQCHIP, but before calling
2762 KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the
2763 base addresses will return -EEXIST.
2765 Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API
2766 should be used instead.
2769 4.86 KVM_PPC_RTAS_DEFINE_TOKEN
2771 Capability: KVM_CAP_PPC_RTAS
2774 Parameters: struct kvm_rtas_token_args
2775 Returns: 0 on success, -1 on error
2777 Defines a token value for a RTAS (Run Time Abstraction Services)
2778 service in order to allow it to be handled in the kernel. The
2779 argument struct gives the name of the service, which must be the name
2780 of a service that has a kernel-side implementation. If the token
2781 value is non-zero, it will be associated with that service, and
2782 subsequent RTAS calls by the guest specifying that token will be
2783 handled by the kernel. If the token value is 0, then any token
2784 associated with the service will be forgotten, and subsequent RTAS
2785 calls by the guest for that service will be passed to userspace to be
2788 4.87 KVM_SET_GUEST_DEBUG
2790 Capability: KVM_CAP_SET_GUEST_DEBUG
2791 Architectures: x86, s390, ppc, arm64
2793 Parameters: struct kvm_guest_debug (in)
2794 Returns: 0 on success; -1 on error
2796 struct kvm_guest_debug {
2799 struct kvm_guest_debug_arch arch;
2802 Set up the processor specific debug registers and configure vcpu for
2803 handling guest debug events. There are two parts to the structure, the
2804 first a control bitfield indicates the type of debug events to handle
2805 when running. Common control bits are:
2807 - KVM_GUESTDBG_ENABLE: guest debugging is enabled
2808 - KVM_GUESTDBG_SINGLESTEP: the next run should single-step
2810 The top 16 bits of the control field are architecture specific control
2811 flags which can include the following:
2813 - KVM_GUESTDBG_USE_SW_BP: using software breakpoints [x86, arm64]
2814 - KVM_GUESTDBG_USE_HW_BP: using hardware breakpoints [x86, s390, arm64]
2815 - KVM_GUESTDBG_INJECT_DB: inject DB type exception [x86]
2816 - KVM_GUESTDBG_INJECT_BP: inject BP type exception [x86]
2817 - KVM_GUESTDBG_EXIT_PENDING: trigger an immediate guest exit [s390]
2819 For example KVM_GUESTDBG_USE_SW_BP indicates that software breakpoints
2820 are enabled in memory so we need to ensure breakpoint exceptions are
2821 correctly trapped and the KVM run loop exits at the breakpoint and not
2822 running off into the normal guest vector. For KVM_GUESTDBG_USE_HW_BP
2823 we need to ensure the guest vCPUs architecture specific registers are
2824 updated to the correct (supplied) values.
2826 The second part of the structure is architecture specific and
2827 typically contains a set of debug registers.
2829 For arm64 the number of debug registers is implementation defined and
2830 can be determined by querying the KVM_CAP_GUEST_DEBUG_HW_BPS and
2831 KVM_CAP_GUEST_DEBUG_HW_WPS capabilities which return a positive number
2832 indicating the number of supported registers.
2834 When debug events exit the main run loop with the reason
2835 KVM_EXIT_DEBUG with the kvm_debug_exit_arch part of the kvm_run
2836 structure containing architecture specific debug information.
2838 4.88 KVM_GET_EMULATED_CPUID
2840 Capability: KVM_CAP_EXT_EMUL_CPUID
2843 Parameters: struct kvm_cpuid2 (in/out)
2844 Returns: 0 on success, -1 on error
2849 struct kvm_cpuid_entry2 entries[0];
2852 The member 'flags' is used for passing flags from userspace.
2854 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
2855 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
2856 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
2858 struct kvm_cpuid_entry2 {
2869 This ioctl returns x86 cpuid features which are emulated by
2870 kvm.Userspace can use the information returned by this ioctl to query
2871 which features are emulated by kvm instead of being present natively.
2873 Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2
2874 structure with the 'nent' field indicating the number of entries in
2875 the variable-size array 'entries'. If the number of entries is too low
2876 to describe the cpu capabilities, an error (E2BIG) is returned. If the
2877 number is too high, the 'nent' field is adjusted and an error (ENOMEM)
2878 is returned. If the number is just right, the 'nent' field is adjusted
2879 to the number of valid entries in the 'entries' array, which is then
2882 The entries returned are the set CPUID bits of the respective features
2883 which kvm emulates, as returned by the CPUID instruction, with unknown
2884 or unsupported feature bits cleared.
2886 Features like x2apic, for example, may not be present in the host cpu
2887 but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be
2888 emulated efficiently and thus not included here.
2890 The fields in each entry are defined as follows:
2892 function: the eax value used to obtain the entry
2893 index: the ecx value used to obtain the entry (for entries that are
2895 flags: an OR of zero or more of the following:
2896 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
2897 if the index field is valid
2898 KVM_CPUID_FLAG_STATEFUL_FUNC:
2899 if cpuid for this function returns different values for successive
2900 invocations; there will be several entries with the same function,
2901 all with this flag set
2902 KVM_CPUID_FLAG_STATE_READ_NEXT:
2903 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
2904 the first entry to be read by a cpu
2905 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
2906 this function/index combination
2908 4.89 KVM_S390_MEM_OP
2910 Capability: KVM_CAP_S390_MEM_OP
2913 Parameters: struct kvm_s390_mem_op (in)
2914 Returns: = 0 on success,
2915 < 0 on generic error (e.g. -EFAULT or -ENOMEM),
2916 > 0 if an exception occurred while walking the page tables
2918 Read or write data from/to the logical (virtual) memory of a VCPU.
2920 Parameters are specified via the following structure:
2922 struct kvm_s390_mem_op {
2923 __u64 gaddr; /* the guest address */
2924 __u64 flags; /* flags */
2925 __u32 size; /* amount of bytes */
2926 __u32 op; /* type of operation */
2927 __u64 buf; /* buffer in userspace */
2928 __u8 ar; /* the access register number */
2929 __u8 reserved[31]; /* should be set to 0 */
2932 The type of operation is specified in the "op" field. It is either
2933 KVM_S390_MEMOP_LOGICAL_READ for reading from logical memory space or
2934 KVM_S390_MEMOP_LOGICAL_WRITE for writing to logical memory space. The
2935 KVM_S390_MEMOP_F_CHECK_ONLY flag can be set in the "flags" field to check
2936 whether the corresponding memory access would create an access exception
2937 (without touching the data in the memory at the destination). In case an
2938 access exception occurred while walking the MMU tables of the guest, the
2939 ioctl returns a positive error number to indicate the type of exception.
2940 This exception is also raised directly at the corresponding VCPU if the
2941 flag KVM_S390_MEMOP_F_INJECT_EXCEPTION is set in the "flags" field.
2943 The start address of the memory region has to be specified in the "gaddr"
2944 field, and the length of the region in the "size" field. "buf" is the buffer
2945 supplied by the userspace application where the read data should be written
2946 to for KVM_S390_MEMOP_LOGICAL_READ, or where the data that should be written
2947 is stored for a KVM_S390_MEMOP_LOGICAL_WRITE. "buf" is unused and can be NULL
2948 when KVM_S390_MEMOP_F_CHECK_ONLY is specified. "ar" designates the access
2949 register number to be used.
2951 The "reserved" field is meant for future extensions. It is not used by
2952 KVM with the currently defined set of flags.
2954 4.90 KVM_S390_GET_SKEYS
2956 Capability: KVM_CAP_S390_SKEYS
2959 Parameters: struct kvm_s390_skeys
2960 Returns: 0 on success, KVM_S390_GET_KEYS_NONE if guest is not using storage
2961 keys, negative value on error
2963 This ioctl is used to get guest storage key values on the s390
2964 architecture. The ioctl takes parameters via the kvm_s390_skeys struct.
2966 struct kvm_s390_skeys {
2969 __u64 skeydata_addr;
2974 The start_gfn field is the number of the first guest frame whose storage keys
2977 The count field is the number of consecutive frames (starting from start_gfn)
2978 whose storage keys to get. The count field must be at least 1 and the maximum
2979 allowed value is defined as KVM_S390_SKEYS_ALLOC_MAX. Values outside this range
2980 will cause the ioctl to return -EINVAL.
2982 The skeydata_addr field is the address to a buffer large enough to hold count
2983 bytes. This buffer will be filled with storage key data by the ioctl.
2985 4.91 KVM_S390_SET_SKEYS
2987 Capability: KVM_CAP_S390_SKEYS
2990 Parameters: struct kvm_s390_skeys
2991 Returns: 0 on success, negative value on error
2993 This ioctl is used to set guest storage key values on the s390
2994 architecture. The ioctl takes parameters via the kvm_s390_skeys struct.
2995 See section on KVM_S390_GET_SKEYS for struct definition.
2997 The start_gfn field is the number of the first guest frame whose storage keys
3000 The count field is the number of consecutive frames (starting from start_gfn)
3001 whose storage keys to get. The count field must be at least 1 and the maximum
3002 allowed value is defined as KVM_S390_SKEYS_ALLOC_MAX. Values outside this range
3003 will cause the ioctl to return -EINVAL.
3005 The skeydata_addr field is the address to a buffer containing count bytes of
3006 storage keys. Each byte in the buffer will be set as the storage key for a
3007 single frame starting at start_gfn for count frames.
3009 Note: If any architecturally invalid key value is found in the given data then
3010 the ioctl will return -EINVAL.
3014 Capability: KVM_CAP_S390_INJECT_IRQ
3017 Parameters: struct kvm_s390_irq (in)
3018 Returns: 0 on success, -1 on error
3020 EINVAL: interrupt type is invalid
3021 type is KVM_S390_SIGP_STOP and flag parameter is invalid value
3022 type is KVM_S390_INT_EXTERNAL_CALL and code is bigger
3023 than the maximum of VCPUs
3024 EBUSY: type is KVM_S390_SIGP_SET_PREFIX and vcpu is not stopped
3025 type is KVM_S390_SIGP_STOP and a stop irq is already pending
3026 type is KVM_S390_INT_EXTERNAL_CALL and an external call interrupt
3029 Allows to inject an interrupt to the guest.
3031 Using struct kvm_s390_irq as a parameter allows
3032 to inject additional payload which is not
3033 possible via KVM_S390_INTERRUPT.
3035 Interrupt parameters are passed via kvm_s390_irq:
3037 struct kvm_s390_irq {
3040 struct kvm_s390_io_info io;
3041 struct kvm_s390_ext_info ext;
3042 struct kvm_s390_pgm_info pgm;
3043 struct kvm_s390_emerg_info emerg;
3044 struct kvm_s390_extcall_info extcall;
3045 struct kvm_s390_prefix_info prefix;
3046 struct kvm_s390_stop_info stop;
3047 struct kvm_s390_mchk_info mchk;
3052 type can be one of the following:
3054 KVM_S390_SIGP_STOP - sigp stop; parameter in .stop
3055 KVM_S390_PROGRAM_INT - program check; parameters in .pgm
3056 KVM_S390_SIGP_SET_PREFIX - sigp set prefix; parameters in .prefix
3057 KVM_S390_RESTART - restart; no parameters
3058 KVM_S390_INT_CLOCK_COMP - clock comparator interrupt; no parameters
3059 KVM_S390_INT_CPU_TIMER - CPU timer interrupt; no parameters
3060 KVM_S390_INT_EMERGENCY - sigp emergency; parameters in .emerg
3061 KVM_S390_INT_EXTERNAL_CALL - sigp external call; parameters in .extcall
3062 KVM_S390_MCHK - machine check interrupt; parameters in .mchk
3065 Note that the vcpu ioctl is asynchronous to vcpu execution.
3067 4.94 KVM_S390_GET_IRQ_STATE
3069 Capability: KVM_CAP_S390_IRQ_STATE
3072 Parameters: struct kvm_s390_irq_state (out)
3073 Returns: >= number of bytes copied into buffer,
3074 -EINVAL if buffer size is 0,
3075 -ENOBUFS if buffer size is too small to fit all pending interrupts,
3076 -EFAULT if the buffer address was invalid
3078 This ioctl allows userspace to retrieve the complete state of all currently
3079 pending interrupts in a single buffer. Use cases include migration
3080 and introspection. The parameter structure contains the address of a
3081 userspace buffer and its length:
3083 struct kvm_s390_irq_state {
3090 Userspace passes in the above struct and for each pending interrupt a
3091 struct kvm_s390_irq is copied to the provided buffer.
3093 If -ENOBUFS is returned the buffer provided was too small and userspace
3094 may retry with a bigger buffer.
3096 4.95 KVM_S390_SET_IRQ_STATE
3098 Capability: KVM_CAP_S390_IRQ_STATE
3101 Parameters: struct kvm_s390_irq_state (in)
3102 Returns: 0 on success,
3103 -EFAULT if the buffer address was invalid,
3104 -EINVAL for an invalid buffer length (see below),
3105 -EBUSY if there were already interrupts pending,
3106 errors occurring when actually injecting the
3107 interrupt. See KVM_S390_IRQ.
3109 This ioctl allows userspace to set the complete state of all cpu-local
3110 interrupts currently pending for the vcpu. It is intended for restoring
3111 interrupt state after a migration. The input parameter is a userspace buffer
3112 containing a struct kvm_s390_irq_state:
3114 struct kvm_s390_irq_state {
3120 The userspace memory referenced by buf contains a struct kvm_s390_irq
3121 for each interrupt to be injected into the guest.
3122 If one of the interrupts could not be injected for some reason the
3125 len must be a multiple of sizeof(struct kvm_s390_irq). It must be > 0
3126 and it must not exceed (max_vcpus + 32) * sizeof(struct kvm_s390_irq),
3127 which is the maximum number of possibly pending cpu-local interrupts.
3131 Capability: KVM_CAP_X86_SMM
3135 Returns: 0 on success, -1 on error
3137 Queues an SMI on the thread's vcpu.
3139 4.97 KVM_CAP_PPC_MULTITCE
3141 Capability: KVM_CAP_PPC_MULTITCE
3145 This capability means the kernel is capable of handling hypercalls
3146 H_PUT_TCE_INDIRECT and H_STUFF_TCE without passing those into the user
3147 space. This significantly accelerates DMA operations for PPC KVM guests.
3148 User space should expect that its handlers for these hypercalls
3149 are not going to be called if user space previously registered LIOBN
3150 in KVM (via KVM_CREATE_SPAPR_TCE or similar calls).
3152 In order to enable H_PUT_TCE_INDIRECT and H_STUFF_TCE use in the guest,
3153 user space might have to advertise it for the guest. For example,
3154 IBM pSeries (sPAPR) guest starts using them if "hcall-multi-tce" is
3155 present in the "ibm,hypertas-functions" device-tree property.
3157 The hypercalls mentioned above may or may not be processed successfully
3158 in the kernel based fast path. If they can not be handled by the kernel,
3159 they will get passed on to user space. So user space still has to have
3160 an implementation for these despite the in kernel acceleration.
3162 This capability is always enabled.
3164 4.98 KVM_CREATE_SPAPR_TCE_64
3166 Capability: KVM_CAP_SPAPR_TCE_64
3167 Architectures: powerpc
3169 Parameters: struct kvm_create_spapr_tce_64 (in)
3170 Returns: file descriptor for manipulating the created TCE table
3172 This is an extension for KVM_CAP_SPAPR_TCE which only supports 32bit
3173 windows, described in 4.62 KVM_CREATE_SPAPR_TCE
3175 This capability uses extended struct in ioctl interface:
3177 /* for KVM_CAP_SPAPR_TCE_64 */
3178 struct kvm_create_spapr_tce_64 {
3182 __u64 offset; /* in pages */
3183 __u64 size; /* in pages */
3186 The aim of extension is to support an additional bigger DMA window with
3187 a variable page size.
3188 KVM_CREATE_SPAPR_TCE_64 receives a 64bit window size, an IOMMU page shift and
3189 a bus offset of the corresponding DMA window, @size and @offset are numbers
3192 @flags are not used at the moment.
3194 The rest of functionality is identical to KVM_CREATE_SPAPR_TCE.
3196 4.99 KVM_REINJECT_CONTROL
3198 Capability: KVM_CAP_REINJECT_CONTROL
3201 Parameters: struct kvm_reinject_control (in)
3202 Returns: 0 on success,
3203 -EFAULT if struct kvm_reinject_control cannot be read,
3204 -ENXIO if KVM_CREATE_PIT or KVM_CREATE_PIT2 didn't succeed earlier.
3206 i8254 (PIT) has two modes, reinject and !reinject. The default is reinject,
3207 where KVM queues elapsed i8254 ticks and monitors completion of interrupt from
3208 vector(s) that i8254 injects. Reinject mode dequeues a tick and injects its
3209 interrupt whenever there isn't a pending interrupt from i8254.
3210 !reinject mode injects an interrupt as soon as a tick arrives.
3212 struct kvm_reinject_control {
3217 pit_reinject = 0 (!reinject mode) is recommended, unless running an old
3218 operating system that uses the PIT for timing (e.g. Linux 2.4.x).
3220 4.100 KVM_PPC_CONFIGURE_V3_MMU
3222 Capability: KVM_CAP_PPC_RADIX_MMU or KVM_CAP_PPC_HASH_MMU_V3
3225 Parameters: struct kvm_ppc_mmuv3_cfg (in)
3226 Returns: 0 on success,
3227 -EFAULT if struct kvm_ppc_mmuv3_cfg cannot be read,
3228 -EINVAL if the configuration is invalid
3230 This ioctl controls whether the guest will use radix or HPT (hashed
3231 page table) translation, and sets the pointer to the process table for
3234 struct kvm_ppc_mmuv3_cfg {
3236 __u64 process_table;
3239 There are two bits that can be set in flags; KVM_PPC_MMUV3_RADIX and
3240 KVM_PPC_MMUV3_GTSE. KVM_PPC_MMUV3_RADIX, if set, configures the guest
3241 to use radix tree translation, and if clear, to use HPT translation.
3242 KVM_PPC_MMUV3_GTSE, if set and if KVM permits it, configures the guest
3243 to be able to use the global TLB and SLB invalidation instructions;
3244 if clear, the guest may not use these instructions.
3246 The process_table field specifies the address and size of the guest
3247 process table, which is in the guest's space. This field is formatted
3248 as the second doubleword of the partition table entry, as defined in
3249 the Power ISA V3.00, Book III section 5.7.6.1.
3251 4.101 KVM_PPC_GET_RMMU_INFO
3253 Capability: KVM_CAP_PPC_RADIX_MMU
3256 Parameters: struct kvm_ppc_rmmu_info (out)
3257 Returns: 0 on success,
3258 -EFAULT if struct kvm_ppc_rmmu_info cannot be written,
3259 -EINVAL if no useful information can be returned
3261 This ioctl returns a structure containing two things: (a) a list
3262 containing supported radix tree geometries, and (b) a list that maps
3263 page sizes to put in the "AP" (actual page size) field for the tlbie
3264 (TLB invalidate entry) instruction.
3266 struct kvm_ppc_rmmu_info {
3267 struct kvm_ppc_radix_geom {
3272 __u32 ap_encodings[8];
3275 The geometries[] field gives up to 8 supported geometries for the
3276 radix page table, in terms of the log base 2 of the smallest page
3277 size, and the number of bits indexed at each level of the tree, from
3278 the PTE level up to the PGD level in that order. Any unused entries
3279 will have 0 in the page_shift field.
3281 The ap_encodings gives the supported page sizes and their AP field
3282 encodings, encoded with the AP value in the top 3 bits and the log
3283 base 2 of the page size in the bottom 6 bits.
3285 4.102 KVM_PPC_RESIZE_HPT_PREPARE
3287 Capability: KVM_CAP_SPAPR_RESIZE_HPT
3288 Architectures: powerpc
3290 Parameters: struct kvm_ppc_resize_hpt (in)
3291 Returns: 0 on successful completion,
3292 >0 if a new HPT is being prepared, the value is an estimated
3293 number of milliseconds until preparation is complete
3294 -EFAULT if struct kvm_reinject_control cannot be read,
3295 -EINVAL if the supplied shift or flags are invalid
3296 -ENOMEM if unable to allocate the new HPT
3297 -ENOSPC if there was a hash collision when moving existing
3298 HPT entries to the new HPT
3299 -EIO on other error conditions
3301 Used to implement the PAPR extension for runtime resizing of a guest's
3302 Hashed Page Table (HPT). Specifically this starts, stops or monitors
3303 the preparation of a new potential HPT for the guest, essentially
3304 implementing the H_RESIZE_HPT_PREPARE hypercall.
3306 If called with shift > 0 when there is no pending HPT for the guest,
3307 this begins preparation of a new pending HPT of size 2^(shift) bytes.
3308 It then returns a positive integer with the estimated number of
3309 milliseconds until preparation is complete.
3311 If called when there is a pending HPT whose size does not match that
3312 requested in the parameters, discards the existing pending HPT and
3313 creates a new one as above.
3315 If called when there is a pending HPT of the size requested, will:
3316 * If preparation of the pending HPT is already complete, return 0
3317 * If preparation of the pending HPT has failed, return an error
3318 code, then discard the pending HPT.
3319 * If preparation of the pending HPT is still in progress, return an
3320 estimated number of milliseconds until preparation is complete.
3322 If called with shift == 0, discards any currently pending HPT and
3323 returns 0 (i.e. cancels any in-progress preparation).
3325 flags is reserved for future expansion, currently setting any bits in
3326 flags will result in an -EINVAL.
3328 Normally this will be called repeatedly with the same parameters until
3329 it returns <= 0. The first call will initiate preparation, subsequent
3330 ones will monitor preparation until it completes or fails.
3332 struct kvm_ppc_resize_hpt {
3338 4.103 KVM_PPC_RESIZE_HPT_COMMIT
3340 Capability: KVM_CAP_SPAPR_RESIZE_HPT
3341 Architectures: powerpc
3343 Parameters: struct kvm_ppc_resize_hpt (in)
3344 Returns: 0 on successful completion,
3345 -EFAULT if struct kvm_reinject_control cannot be read,
3346 -EINVAL if the supplied shift or flags are invalid
3347 -ENXIO is there is no pending HPT, or the pending HPT doesn't
3348 have the requested size
3349 -EBUSY if the pending HPT is not fully prepared
3350 -ENOSPC if there was a hash collision when moving existing
3351 HPT entries to the new HPT
3352 -EIO on other error conditions
3354 Used to implement the PAPR extension for runtime resizing of a guest's
3355 Hashed Page Table (HPT). Specifically this requests that the guest be
3356 transferred to working with the new HPT, essentially implementing the
3357 H_RESIZE_HPT_COMMIT hypercall.
3359 This should only be called after KVM_PPC_RESIZE_HPT_PREPARE has
3360 returned 0 with the same parameters. In other cases
3361 KVM_PPC_RESIZE_HPT_COMMIT will return an error (usually -ENXIO or
3362 -EBUSY, though others may be possible if the preparation was started,
3365 This will have undefined effects on the guest if it has not already
3366 placed itself in a quiescent state where no vcpu will make MMU enabled
3369 On succsful completion, the pending HPT will become the guest's active
3370 HPT and the previous HPT will be discarded.
3372 On failure, the guest will still be operating on its previous HPT.
3374 struct kvm_ppc_resize_hpt {
3380 5. The kvm_run structure
3381 ------------------------
3383 Application code obtains a pointer to the kvm_run structure by
3384 mmap()ing a vcpu fd. From that point, application code can control
3385 execution by changing fields in kvm_run prior to calling the KVM_RUN
3386 ioctl, and obtain information about the reason KVM_RUN returned by
3387 looking up structure members.
3391 __u8 request_interrupt_window;
3393 Request that KVM_RUN return when it becomes possible to inject external
3394 interrupts into the guest. Useful in conjunction with KVM_INTERRUPT.
3396 __u8 immediate_exit;
3398 This field is polled once when KVM_RUN starts; if non-zero, KVM_RUN
3399 exits immediately, returning -EINTR. In the common scenario where a
3400 signal is used to "kick" a VCPU out of KVM_RUN, this field can be used
3401 to avoid usage of KVM_SET_SIGNAL_MASK, which has worse scalability.
3402 Rather than blocking the signal outside KVM_RUN, userspace can set up
3403 a signal handler that sets run->immediate_exit to a non-zero value.
3405 This field is ignored if KVM_CAP_IMMEDIATE_EXIT is not available.
3412 When KVM_RUN has returned successfully (return value 0), this informs
3413 application code why KVM_RUN has returned. Allowable values for this
3414 field are detailed below.
3416 __u8 ready_for_interrupt_injection;
3418 If request_interrupt_window has been specified, this field indicates
3419 an interrupt can be injected now with KVM_INTERRUPT.
3423 The value of the current interrupt flag. Only valid if in-kernel
3424 local APIC is not used.
3428 More architecture-specific flags detailing state of the VCPU that may
3429 affect the device's behavior. The only currently defined flag is
3430 KVM_RUN_X86_SMM, which is valid on x86 machines and is set if the
3431 VCPU is in system management mode.
3433 /* in (pre_kvm_run), out (post_kvm_run) */
3436 The value of the cr8 register. Only valid if in-kernel local APIC is
3437 not used. Both input and output.
3441 The value of the APIC BASE msr. Only valid if in-kernel local
3442 APIC is not used. Both input and output.
3445 /* KVM_EXIT_UNKNOWN */
3447 __u64 hardware_exit_reason;
3450 If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown
3451 reasons. Further architecture-specific information is available in
3452 hardware_exit_reason.
3454 /* KVM_EXIT_FAIL_ENTRY */
3456 __u64 hardware_entry_failure_reason;
3459 If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due
3460 to unknown reasons. Further architecture-specific information is
3461 available in hardware_entry_failure_reason.
3463 /* KVM_EXIT_EXCEPTION */
3473 #define KVM_EXIT_IO_IN 0
3474 #define KVM_EXIT_IO_OUT 1
3476 __u8 size; /* bytes */
3479 __u64 data_offset; /* relative to kvm_run start */
3482 If exit_reason is KVM_EXIT_IO, then the vcpu has
3483 executed a port I/O instruction which could not be satisfied by kvm.
3484 data_offset describes where the data is located (KVM_EXIT_IO_OUT) or
3485 where kvm expects application code to place the data for the next
3486 KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array.
3488 /* KVM_EXIT_DEBUG */
3490 struct kvm_debug_exit_arch arch;
3493 If the exit_reason is KVM_EXIT_DEBUG, then a vcpu is processing a debug event
3494 for which architecture specific information is returned.
3504 If exit_reason is KVM_EXIT_MMIO, then the vcpu has
3505 executed a memory-mapped I/O instruction which could not be satisfied
3506 by kvm. The 'data' member contains the written data if 'is_write' is
3507 true, and should be filled by application code otherwise.
3509 The 'data' member contains, in its first 'len' bytes, the value as it would
3510 appear if the VCPU performed a load or store of the appropriate width directly
3513 NOTE: For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_PAPR and
3514 KVM_EXIT_EPR the corresponding
3515 operations are complete (and guest state is consistent) only after userspace
3516 has re-entered the kernel with KVM_RUN. The kernel side will first finish
3517 incomplete operations and then check for pending signals. Userspace
3518 can re-enter the guest with an unmasked signal pending to complete
3521 /* KVM_EXIT_HYPERCALL */
3530 Unused. This was once used for 'hypercall to userspace'. To implement
3531 such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390).
3532 Note KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO.
3534 /* KVM_EXIT_TPR_ACCESS */
3541 To be documented (KVM_TPR_ACCESS_REPORTING).
3543 /* KVM_EXIT_S390_SIEIC */
3546 __u64 mask; /* psw upper half */
3547 __u64 addr; /* psw lower half */
3554 /* KVM_EXIT_S390_RESET */
3555 #define KVM_S390_RESET_POR 1
3556 #define KVM_S390_RESET_CLEAR 2
3557 #define KVM_S390_RESET_SUBSYSTEM 4
3558 #define KVM_S390_RESET_CPU_INIT 8
3559 #define KVM_S390_RESET_IPL 16
3560 __u64 s390_reset_flags;
3564 /* KVM_EXIT_S390_UCONTROL */
3566 __u64 trans_exc_code;
3570 s390 specific. A page fault has occurred for a user controlled virtual
3571 machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be
3572 resolved by the kernel.
3573 The program code and the translation exception code that were placed
3574 in the cpu's lowcore are presented here as defined by the z Architecture
3575 Principles of Operation Book in the Chapter for Dynamic Address Translation
3585 Deprecated - was used for 440 KVM.
3592 MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch
3593 hypercalls and exit with this exit struct that contains all the guest gprs.
3595 If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall.
3596 Userspace can now handle the hypercall and when it's done modify the gprs as
3597 necessary. Upon guest entry all guest GPRs will then be replaced by the values
3600 /* KVM_EXIT_PAPR_HCALL */
3607 This is used on 64-bit PowerPC when emulating a pSeries partition,
3608 e.g. with the 'pseries' machine type in qemu. It occurs when the
3609 guest does a hypercall using the 'sc 1' instruction. The 'nr' field
3610 contains the hypercall number (from the guest R3), and 'args' contains
3611 the arguments (from the guest R4 - R12). Userspace should put the
3612 return code in 'ret' and any extra returned values in args[].
3613 The possible hypercalls are defined in the Power Architecture Platform
3614 Requirements (PAPR) document available from www.power.org (free
3615 developer registration required to access it).
3617 /* KVM_EXIT_S390_TSCH */
3619 __u16 subchannel_id;
3620 __u16 subchannel_nr;
3627 s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled
3628 and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O
3629 interrupt for the target subchannel has been dequeued and subchannel_id,
3630 subchannel_nr, io_int_parm and io_int_word contain the parameters for that
3631 interrupt. ipb is needed for instruction parameter decoding.
3638 On FSL BookE PowerPC chips, the interrupt controller has a fast patch
3639 interrupt acknowledge path to the core. When the core successfully
3640 delivers an interrupt, it automatically populates the EPR register with
3641 the interrupt vector number and acknowledges the interrupt inside
3642 the interrupt controller.
3644 In case the interrupt controller lives in user space, we need to do
3645 the interrupt acknowledge cycle through it to fetch the next to be
3646 delivered interrupt vector using this exit.
3648 It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an
3649 external interrupt has just been delivered into the guest. User space
3650 should put the acknowledged interrupt vector into the 'epr' field.
3652 /* KVM_EXIT_SYSTEM_EVENT */
3654 #define KVM_SYSTEM_EVENT_SHUTDOWN 1
3655 #define KVM_SYSTEM_EVENT_RESET 2
3656 #define KVM_SYSTEM_EVENT_CRASH 3
3661 If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered
3662 a system-level event using some architecture specific mechanism (hypercall
3663 or some special instruction). In case of ARM/ARM64, this is triggered using
3664 HVC instruction based PSCI call from the vcpu. The 'type' field describes
3665 the system-level event type. The 'flags' field describes architecture
3666 specific flags for the system-level event.
3668 Valid values for 'type' are:
3669 KVM_SYSTEM_EVENT_SHUTDOWN -- the guest has requested a shutdown of the
3670 VM. Userspace is not obliged to honour this, and if it does honour
3671 this does not need to destroy the VM synchronously (ie it may call
3672 KVM_RUN again before shutdown finally occurs).
3673 KVM_SYSTEM_EVENT_RESET -- the guest has requested a reset of the VM.
3674 As with SHUTDOWN, userspace can choose to ignore the request, or
3675 to schedule the reset to occur in the future and may call KVM_RUN again.
3676 KVM_SYSTEM_EVENT_CRASH -- the guest crash occurred and the guest
3677 has requested a crash condition maintenance. Userspace can choose
3678 to ignore the request, or to gather VM memory core dump and/or
3679 reset/shutdown of the VM.
3681 /* KVM_EXIT_IOAPIC_EOI */
3686 Indicates that the VCPU's in-kernel local APIC received an EOI for a
3687 level-triggered IOAPIC interrupt. This exit only triggers when the
3688 IOAPIC is implemented in userspace (i.e. KVM_CAP_SPLIT_IRQCHIP is enabled);
3689 the userspace IOAPIC should process the EOI and retrigger the interrupt if
3690 it is still asserted. Vector is the LAPIC interrupt vector for which the
3693 struct kvm_hyperv_exit {
3694 #define KVM_EXIT_HYPERV_SYNIC 1
3695 #define KVM_EXIT_HYPERV_HCALL 2
3711 /* KVM_EXIT_HYPERV */
3712 struct kvm_hyperv_exit hyperv;
3713 Indicates that the VCPU exits into userspace to process some tasks
3714 related to Hyper-V emulation.
3715 Valid values for 'type' are:
3716 KVM_EXIT_HYPERV_SYNIC -- synchronously notify user-space about
3717 Hyper-V SynIC state change. Notification is used to remap SynIC
3718 event/message pages and to enable/disable SynIC messages/events processing
3721 /* Fix the size of the union. */
3726 * shared registers between kvm and userspace.
3727 * kvm_valid_regs specifies the register classes set by the host
3728 * kvm_dirty_regs specified the register classes dirtied by userspace
3729 * struct kvm_sync_regs is architecture specific, as well as the
3730 * bits for kvm_valid_regs and kvm_dirty_regs
3732 __u64 kvm_valid_regs;
3733 __u64 kvm_dirty_regs;
3735 struct kvm_sync_regs regs;
3739 If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access
3740 certain guest registers without having to call SET/GET_*REGS. Thus we can
3741 avoid some system call overhead if userspace has to handle the exit.
3742 Userspace can query the validity of the structure by checking
3743 kvm_valid_regs for specific bits. These bits are architecture specific
3744 and usually define the validity of a groups of registers. (e.g. one bit
3745 for general purpose registers)
3747 Please note that the kernel is allowed to use the kvm_run structure as the
3748 primary storage for certain register types. Therefore, the kernel may use the
3749 values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set.
3755 6. Capabilities that can be enabled on vCPUs
3756 --------------------------------------------
3758 There are certain capabilities that change the behavior of the virtual CPU or
3759 the virtual machine when enabled. To enable them, please see section 4.37.
3760 Below you can find a list of capabilities and what their effect on the vCPU or
3761 the virtual machine is when enabling them.
3763 The following information is provided along with the description:
3765 Architectures: which instruction set architectures provide this ioctl.
3766 x86 includes both i386 and x86_64.
3768 Target: whether this is a per-vcpu or per-vm capability.
3770 Parameters: what parameters are accepted by the capability.
3772 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
3773 are not detailed, but errors with specific meanings are.
3781 Returns: 0 on success; -1 on error
3783 This capability enables interception of OSI hypercalls that otherwise would
3784 be treated as normal system calls to be injected into the guest. OSI hypercalls
3785 were invented by Mac-on-Linux to have a standardized communication mechanism
3786 between the guest and the host.
3788 When this capability is enabled, KVM_EXIT_OSI can occur.
3791 6.2 KVM_CAP_PPC_PAPR
3796 Returns: 0 on success; -1 on error
3798 This capability enables interception of PAPR hypercalls. PAPR hypercalls are
3799 done using the hypercall instruction "sc 1".
3801 It also sets the guest privilege level to "supervisor" mode. Usually the guest
3802 runs in "hypervisor" privilege mode with a few missing features.
3804 In addition to the above, it changes the semantics of SDR1. In this mode, the
3805 HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the
3806 HTAB invisible to the guest.
3808 When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur.
3815 Parameters: args[0] is the address of a struct kvm_config_tlb
3816 Returns: 0 on success; -1 on error
3818 struct kvm_config_tlb {
3825 Configures the virtual CPU's TLB array, establishing a shared memory area
3826 between userspace and KVM. The "params" and "array" fields are userspace
3827 addresses of mmu-type-specific data structures. The "array_len" field is an
3828 safety mechanism, and should be set to the size in bytes of the memory that
3829 userspace has reserved for the array. It must be at least the size dictated
3830 by "mmu_type" and "params".
3832 While KVM_RUN is active, the shared region is under control of KVM. Its
3833 contents are undefined, and any modification by userspace results in
3834 boundedly undefined behavior.
3836 On return from KVM_RUN, the shared region will reflect the current state of
3837 the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB
3838 to tell KVM which entries have been changed, prior to calling KVM_RUN again
3841 For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV:
3842 - The "params" field is of type "struct kvm_book3e_206_tlb_params".
3843 - The "array" field points to an array of type "struct
3844 kvm_book3e_206_tlb_entry".
3845 - The array consists of all entries in the first TLB, followed by all
3846 entries in the second TLB.
3847 - Within a TLB, entries are ordered first by increasing set number. Within a
3848 set, entries are ordered by way (increasing ESEL).
3849 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1)
3850 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value.
3851 - The tsize field of mas1 shall be set to 4K on TLB0, even though the
3852 hardware ignores this value for TLB0.
3854 6.4 KVM_CAP_S390_CSS_SUPPORT
3859 Returns: 0 on success; -1 on error
3861 This capability enables support for handling of channel I/O instructions.
3863 TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are
3864 handled in-kernel, while the other I/O instructions are passed to userspace.
3866 When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST
3867 SUBCHANNEL intercepts.
3869 Note that even though this capability is enabled per-vcpu, the complete
3870 virtual machine is affected.
3876 Parameters: args[0] defines whether the proxy facility is active
3877 Returns: 0 on success; -1 on error
3879 This capability enables or disables the delivery of interrupts through the
3880 external proxy facility.
3882 When enabled (args[0] != 0), every time the guest gets an external interrupt
3883 delivered, it automatically exits into user space with a KVM_EXIT_EPR exit
3884 to receive the topmost interrupt vector.
3886 When disabled (args[0] == 0), behavior is as if this facility is unsupported.
3888 When this capability is enabled, KVM_EXIT_EPR can occur.
3890 6.6 KVM_CAP_IRQ_MPIC
3893 Parameters: args[0] is the MPIC device fd
3894 args[1] is the MPIC CPU number for this vcpu
3896 This capability connects the vcpu to an in-kernel MPIC device.
3898 6.7 KVM_CAP_IRQ_XICS
3902 Parameters: args[0] is the XICS device fd
3903 args[1] is the XICS CPU number (server ID) for this vcpu
3905 This capability connects the vcpu to an in-kernel XICS device.
3907 6.8 KVM_CAP_S390_IRQCHIP
3913 This capability enables the in-kernel irqchip for s390. Please refer to
3914 "4.24 KVM_CREATE_IRQCHIP" for details.
3916 6.9 KVM_CAP_MIPS_FPU
3920 Parameters: args[0] is reserved for future use (should be 0).
3922 This capability allows the use of the host Floating Point Unit by the guest. It
3923 allows the Config1.FP bit to be set to enable the FPU in the guest. Once this is
3924 done the KVM_REG_MIPS_FPR_* and KVM_REG_MIPS_FCR_* registers can be accessed
3925 (depending on the current guest FPU register mode), and the Status.FR,
3926 Config5.FRE bits are accessible via the KVM API and also from the guest,
3927 depending on them being supported by the FPU.
3929 6.10 KVM_CAP_MIPS_MSA
3933 Parameters: args[0] is reserved for future use (should be 0).
3935 This capability allows the use of the MIPS SIMD Architecture (MSA) by the guest.
3936 It allows the Config3.MSAP bit to be set to enable the use of MSA by the guest.
3937 Once this is done the KVM_REG_MIPS_VEC_* and KVM_REG_MIPS_MSA_* registers can be
3938 accessed, and the Config5.MSAEn bit is accessible via the KVM API and also from
3941 7. Capabilities that can be enabled on VMs
3942 ------------------------------------------
3944 There are certain capabilities that change the behavior of the virtual
3945 machine when enabled. To enable them, please see section 4.37. Below
3946 you can find a list of capabilities and what their effect on the VM
3947 is when enabling them.
3949 The following information is provided along with the description:
3951 Architectures: which instruction set architectures provide this ioctl.
3952 x86 includes both i386 and x86_64.
3954 Parameters: what parameters are accepted by the capability.
3956 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
3957 are not detailed, but errors with specific meanings are.
3960 7.1 KVM_CAP_PPC_ENABLE_HCALL
3963 Parameters: args[0] is the sPAPR hcall number
3964 args[1] is 0 to disable, 1 to enable in-kernel handling
3966 This capability controls whether individual sPAPR hypercalls (hcalls)
3967 get handled by the kernel or not. Enabling or disabling in-kernel
3968 handling of an hcall is effective across the VM. On creation, an
3969 initial set of hcalls are enabled for in-kernel handling, which
3970 consists of those hcalls for which in-kernel handlers were implemented
3971 before this capability was implemented. If disabled, the kernel will
3972 not to attempt to handle the hcall, but will always exit to userspace
3973 to handle it. Note that it may not make sense to enable some and
3974 disable others of a group of related hcalls, but KVM does not prevent
3975 userspace from doing that.
3977 If the hcall number specified is not one that has an in-kernel
3978 implementation, the KVM_ENABLE_CAP ioctl will fail with an EINVAL
3981 7.2 KVM_CAP_S390_USER_SIGP
3986 This capability controls which SIGP orders will be handled completely in user
3987 space. With this capability enabled, all fast orders will be handled completely
3993 - CONDITIONAL EMERGENCY SIGNAL
3995 All other orders will be handled completely in user space.
3997 Only privileged operation exceptions will be checked for in the kernel (or even
3998 in the hardware prior to interception). If this capability is not enabled, the
3999 old way of handling SIGP orders is used (partially in kernel and user space).
4001 7.3 KVM_CAP_S390_VECTOR_REGISTERS
4005 Returns: 0 on success, negative value on error
4007 Allows use of the vector registers introduced with z13 processor, and
4008 provides for the synchronization between host and user space. Will
4009 return -EINVAL if the machine does not support vectors.
4011 7.4 KVM_CAP_S390_USER_STSI
4016 This capability allows post-handlers for the STSI instruction. After
4017 initial handling in the kernel, KVM exits to user space with
4018 KVM_EXIT_S390_STSI to allow user space to insert further data.
4020 Before exiting to userspace, kvm handlers should fill in s390_stsi field of
4031 @addr - guest address of STSI SYSIB
4035 @ar - access register number
4037 KVM handlers should exit to userspace with rc = -EREMOTE.
4039 7.5 KVM_CAP_SPLIT_IRQCHIP
4042 Parameters: args[0] - number of routes reserved for userspace IOAPICs
4043 Returns: 0 on success, -1 on error
4045 Create a local apic for each processor in the kernel. This can be used
4046 instead of KVM_CREATE_IRQCHIP if the userspace VMM wishes to emulate the
4047 IOAPIC and PIC (and also the PIT, even though this has to be enabled
4050 This capability also enables in kernel routing of interrupt requests;
4051 when KVM_CAP_SPLIT_IRQCHIP only routes of KVM_IRQ_ROUTING_MSI type are
4052 used in the IRQ routing table. The first args[0] MSI routes are reserved
4053 for the IOAPIC pins. Whenever the LAPIC receives an EOI for these routes,
4054 a KVM_EXIT_IOAPIC_EOI vmexit will be reported to userspace.
4056 Fails if VCPU has already been created, or if the irqchip is already in the
4057 kernel (i.e. KVM_CREATE_IRQCHIP has already been called).
4064 Allows use of runtime-instrumentation introduced with zEC12 processor.
4065 Will return -EINVAL if the machine does not support runtime-instrumentation.
4066 Will return -EBUSY if a VCPU has already been created.
4068 7.7 KVM_CAP_X2APIC_API
4071 Parameters: args[0] - features that should be enabled
4072 Returns: 0 on success, -EINVAL when args[0] contains invalid features
4074 Valid feature flags in args[0] are
4076 #define KVM_X2APIC_API_USE_32BIT_IDS (1ULL << 0)
4077 #define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK (1ULL << 1)
4079 Enabling KVM_X2APIC_API_USE_32BIT_IDS changes the behavior of
4080 KVM_SET_GSI_ROUTING, KVM_SIGNAL_MSI, KVM_SET_LAPIC, and KVM_GET_LAPIC,
4081 allowing the use of 32-bit APIC IDs. See KVM_CAP_X2APIC_API in their
4082 respective sections.
4084 KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK must be enabled for x2APIC to work
4085 in logical mode or with more than 255 VCPUs. Otherwise, KVM treats 0xff
4086 as a broadcast even in x2APIC mode in order to support physical x2APIC
4087 without interrupt remapping. This is undesirable in logical mode,
4088 where 0xff represents CPUs 0-7 in cluster 0.
4090 7.8 KVM_CAP_S390_USER_INSTR0
4095 With this capability enabled, all illegal instructions 0x0000 (2 bytes) will
4096 be intercepted and forwarded to user space. User space can use this
4097 mechanism e.g. to realize 2-byte software breakpoints. The kernel will
4098 not inject an operating exception for these instructions, user space has
4099 to take care of that.
4101 This capability can be enabled dynamically even if VCPUs were already
4102 created and are running.
4104 8. Other capabilities.
4105 ----------------------
4107 This section lists capabilities that give information about other
4108 features of the KVM implementation.
4110 8.1 KVM_CAP_PPC_HWRNG
4114 This capability, if KVM_CHECK_EXTENSION indicates that it is
4115 available, means that that the kernel has an implementation of the
4116 H_RANDOM hypercall backed by a hardware random-number generator.
4117 If present, the kernel H_RANDOM handler can be enabled for guest use
4118 with the KVM_CAP_PPC_ENABLE_HCALL capability.
4120 8.2 KVM_CAP_HYPERV_SYNIC
4123 This capability, if KVM_CHECK_EXTENSION indicates that it is
4124 available, means that that the kernel has an implementation of the
4125 Hyper-V Synthetic interrupt controller(SynIC). Hyper-V SynIC is
4126 used to support Windows Hyper-V based guest paravirt drivers(VMBus).
4128 In order to use SynIC, it has to be activated by setting this
4129 capability via KVM_ENABLE_CAP ioctl on the vcpu fd. Note that this
4130 will disable the use of APIC hardware virtualization even if supported
4131 by the CPU, as it's incompatible with SynIC auto-EOI behavior.
4133 8.3 KVM_CAP_PPC_RADIX_MMU
4137 This capability, if KVM_CHECK_EXTENSION indicates that it is
4138 available, means that that the kernel can support guests using the
4139 radix MMU defined in Power ISA V3.00 (as implemented in the POWER9
4142 8.4 KVM_CAP_PPC_HASH_MMU_V3
4146 This capability, if KVM_CHECK_EXTENSION indicates that it is
4147 available, means that that the kernel can support guests using the
4148 hashed page table MMU defined in Power ISA V3.00 (as implemented in
4149 the POWER9 processor), including in-memory segment tables.