1 The Definitive KVM (Kernel-based Virtual Machine) API Documentation
2 ===================================================================
7 The kvm API is a set of ioctls that are issued to control various aspects
8 of a virtual machine. The ioctls belong to three classes
10 - System ioctls: These query and set global attributes which affect the
11 whole kvm subsystem. In addition a system ioctl is used to create
14 - VM ioctls: These query and set attributes that affect an entire virtual
15 machine, for example memory layout. In addition a VM ioctl is used to
16 create virtual cpus (vcpus).
18 Only run VM ioctls from the same process (address space) that was used
21 - vcpu ioctls: These query and set attributes that control the operation
22 of a single virtual cpu.
24 Only run vcpu ioctls from the same thread that was used to create the
31 The kvm API is centered around file descriptors. An initial
32 open("/dev/kvm") obtains a handle to the kvm subsystem; this handle
33 can be used to issue system ioctls. A KVM_CREATE_VM ioctl on this
34 handle will create a VM file descriptor which can be used to issue VM
35 ioctls. A KVM_CREATE_VCPU ioctl on a VM fd will create a virtual cpu
36 and return a file descriptor pointing to it. Finally, ioctls on a vcpu
37 fd can be used to control the vcpu, including the important task of
38 actually running guest code.
40 In general file descriptors can be migrated among processes by means
41 of fork() and the SCM_RIGHTS facility of unix domain socket. These
42 kinds of tricks are explicitly not supported by kvm. While they will
43 not cause harm to the host, their actual behavior is not guaranteed by
44 the API. The only supported use is one virtual machine per process,
45 and one vcpu per thread.
51 As of Linux 2.6.22, the KVM ABI has been stabilized: no backward
52 incompatible change are allowed. However, there is an extension
53 facility that allows backward-compatible extensions to the API to be
56 The extension mechanism is not based on the Linux version number.
57 Instead, kvm defines extension identifiers and a facility to query
58 whether a particular extension identifier is available. If it is, a
59 set of ioctls is available for application use.
65 This section describes ioctls that can be used to control kvm guests.
66 For each ioctl, the following information is provided along with a
69 Capability: which KVM extension provides this ioctl. Can be 'basic',
70 which means that is will be provided by any kernel that supports
71 API version 12 (see section 4.1), a KVM_CAP_xyz constant, which
72 means availability needs to be checked with KVM_CHECK_EXTENSION
73 (see section 4.4), or 'none' which means that while not all kernels
74 support this ioctl, there's no capability bit to check its
75 availability: for kernels that don't support the ioctl,
76 the ioctl returns -ENOTTY.
78 Architectures: which instruction set architectures provide this ioctl.
79 x86 includes both i386 and x86_64.
81 Type: system, vm, or vcpu.
83 Parameters: what parameters are accepted by the ioctl.
85 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
86 are not detailed, but errors with specific meanings are.
89 4.1 KVM_GET_API_VERSION
95 Returns: the constant KVM_API_VERSION (=12)
97 This identifies the API version as the stable kvm API. It is not
98 expected that this number will change. However, Linux 2.6.20 and
99 2.6.21 report earlier versions; these are not documented and not
100 supported. Applications should refuse to run if KVM_GET_API_VERSION
101 returns a value other than 12. If this check passes, all ioctls
102 described as 'basic' will be available.
110 Parameters: machine type identifier (KVM_VM_*)
111 Returns: a VM fd that can be used to control the new virtual machine.
113 The new VM has no virtual cpus and no memory. An mmap() of a VM fd
114 will access the virtual machine's physical address space; offset zero
115 corresponds to guest physical address zero. Use of mmap() on a VM fd
116 is discouraged if userspace memory allocation (KVM_CAP_USER_MEMORY) is
118 You most certainly want to use 0 as machine type.
120 In order to create user controlled virtual machines on S390, check
121 KVM_CAP_S390_UCONTROL and use the flag KVM_VM_S390_UCONTROL as
122 privileged user (CAP_SYS_ADMIN).
125 4.3 KVM_GET_MSR_INDEX_LIST
130 Parameters: struct kvm_msr_list (in/out)
131 Returns: 0 on success; -1 on error
133 E2BIG: the msr index list is to be to fit in the array specified by
136 struct kvm_msr_list {
137 __u32 nmsrs; /* number of msrs in entries */
141 This ioctl returns the guest msrs that are supported. The list varies
142 by kvm version and host processor, but does not change otherwise. The
143 user fills in the size of the indices array in nmsrs, and in return
144 kvm adjusts nmsrs to reflect the actual number of msrs and fills in
145 the indices array with their numbers.
147 Note: if kvm indicates supports MCE (KVM_CAP_MCE), then the MCE bank MSRs are
148 not returned in the MSR list, as different vcpus can have a different number
149 of banks, as set via the KVM_X86_SETUP_MCE ioctl.
152 4.4 KVM_CHECK_EXTENSION
154 Capability: basic, KVM_CAP_CHECK_EXTENSION_VM for vm ioctl
156 Type: system ioctl, vm ioctl
157 Parameters: extension identifier (KVM_CAP_*)
158 Returns: 0 if unsupported; 1 (or some other positive integer) if supported
160 The API allows the application to query about extensions to the core
161 kvm API. Userspace passes an extension identifier (an integer) and
162 receives an integer that describes the extension availability.
163 Generally 0 means no and 1 means yes, but some extensions may report
164 additional information in the integer return value.
166 Based on their initialization different VMs may have different capabilities.
167 It is thus encouraged to use the vm ioctl to query for capabilities (available
168 with KVM_CAP_CHECK_EXTENSION_VM on the vm fd)
170 4.5 KVM_GET_VCPU_MMAP_SIZE
176 Returns: size of vcpu mmap area, in bytes
178 The KVM_RUN ioctl (cf.) communicates with userspace via a shared
179 memory region. This ioctl returns the size of that region. See the
180 KVM_RUN documentation for details.
183 4.6 KVM_SET_MEMORY_REGION
188 Parameters: struct kvm_memory_region (in)
189 Returns: 0 on success, -1 on error
191 This ioctl is obsolete and has been removed.
199 Parameters: vcpu id (apic id on x86)
200 Returns: vcpu fd on success, -1 on error
202 This API adds a vcpu to a virtual machine. The vcpu id is a small integer
203 in the range [0, max_vcpus).
205 The recommended max_vcpus value can be retrieved using the KVM_CAP_NR_VCPUS of
206 the KVM_CHECK_EXTENSION ioctl() at run-time.
207 The maximum possible value for max_vcpus can be retrieved using the
208 KVM_CAP_MAX_VCPUS of the KVM_CHECK_EXTENSION ioctl() at run-time.
210 If the KVM_CAP_NR_VCPUS does not exist, you should assume that max_vcpus is 4
212 If the KVM_CAP_MAX_VCPUS does not exist, you should assume that max_vcpus is
213 same as the value returned from KVM_CAP_NR_VCPUS.
215 On powerpc using book3s_hv mode, the vcpus are mapped onto virtual
216 threads in one or more virtual CPU cores. (This is because the
217 hardware requires all the hardware threads in a CPU core to be in the
218 same partition.) The KVM_CAP_PPC_SMT capability indicates the number
219 of vcpus per virtual core (vcore). The vcore id is obtained by
220 dividing the vcpu id by the number of vcpus per vcore. The vcpus in a
221 given vcore will always be in the same physical core as each other
222 (though that might be a different physical core from time to time).
223 Userspace can control the threading (SMT) mode of the guest by its
224 allocation of vcpu ids. For example, if userspace wants
225 single-threaded guest vcpus, it should make all vcpu ids be a multiple
226 of the number of vcpus per vcore.
228 For virtual cpus that have been created with S390 user controlled virtual
229 machines, the resulting vcpu fd can be memory mapped at page offset
230 KVM_S390_SIE_PAGE_OFFSET in order to obtain a memory map of the virtual
231 cpu's hardware control block.
234 4.8 KVM_GET_DIRTY_LOG (vm ioctl)
239 Parameters: struct kvm_dirty_log (in/out)
240 Returns: 0 on success, -1 on error
242 /* for KVM_GET_DIRTY_LOG */
243 struct kvm_dirty_log {
247 void __user *dirty_bitmap; /* one bit per page */
252 Given a memory slot, return a bitmap containing any pages dirtied
253 since the last call to this ioctl. Bit 0 is the first page in the
254 memory slot. Ensure the entire structure is cleared to avoid padding
258 4.9 KVM_SET_MEMORY_ALIAS
263 Parameters: struct kvm_memory_alias (in)
264 Returns: 0 (success), -1 (error)
266 This ioctl is obsolete and has been removed.
275 Returns: 0 on success, -1 on error
277 EINTR: an unmasked signal is pending
279 This ioctl is used to run a guest virtual cpu. While there are no
280 explicit parameters, there is an implicit parameter block that can be
281 obtained by mmap()ing the vcpu fd at offset 0, with the size given by
282 KVM_GET_VCPU_MMAP_SIZE. The parameter block is formatted as a 'struct
283 kvm_run' (see below).
289 Architectures: all except ARM, arm64
291 Parameters: struct kvm_regs (out)
292 Returns: 0 on success, -1 on error
294 Reads the general purpose registers from the vcpu.
298 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
299 __u64 rax, rbx, rcx, rdx;
300 __u64 rsi, rdi, rsp, rbp;
301 __u64 r8, r9, r10, r11;
302 __u64 r12, r13, r14, r15;
308 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
319 Architectures: all except ARM, arm64
321 Parameters: struct kvm_regs (in)
322 Returns: 0 on success, -1 on error
324 Writes the general purpose registers into the vcpu.
326 See KVM_GET_REGS for the data structure.
332 Architectures: x86, ppc
334 Parameters: struct kvm_sregs (out)
335 Returns: 0 on success, -1 on error
337 Reads special registers from the vcpu.
341 struct kvm_segment cs, ds, es, fs, gs, ss;
342 struct kvm_segment tr, ldt;
343 struct kvm_dtable gdt, idt;
344 __u64 cr0, cr2, cr3, cr4, cr8;
347 __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64];
350 /* ppc -- see arch/powerpc/include/uapi/asm/kvm.h */
352 interrupt_bitmap is a bitmap of pending external interrupts. At most
353 one bit may be set. This interrupt has been acknowledged by the APIC
354 but not yet injected into the cpu core.
360 Architectures: x86, ppc
362 Parameters: struct kvm_sregs (in)
363 Returns: 0 on success, -1 on error
365 Writes special registers into the vcpu. See KVM_GET_SREGS for the
374 Parameters: struct kvm_translation (in/out)
375 Returns: 0 on success, -1 on error
377 Translates a virtual address according to the vcpu's current address
380 struct kvm_translation {
382 __u64 linear_address;
385 __u64 physical_address;
396 Architectures: x86, ppc, mips
398 Parameters: struct kvm_interrupt (in)
399 Returns: 0 on success, -1 on error
401 Queues a hardware interrupt vector to be injected. This is only
402 useful if in-kernel local APIC or equivalent is not used.
404 /* for KVM_INTERRUPT */
405 struct kvm_interrupt {
412 Note 'irq' is an interrupt vector, not an interrupt pin or line.
416 Queues an external interrupt to be injected. This ioctl is overleaded
417 with 3 different irq values:
421 This injects an edge type external interrupt into the guest once it's ready
422 to receive interrupts. When injected, the interrupt is done.
424 b) KVM_INTERRUPT_UNSET
426 This unsets any pending interrupt.
428 Only available with KVM_CAP_PPC_UNSET_IRQ.
430 c) KVM_INTERRUPT_SET_LEVEL
432 This injects a level type external interrupt into the guest context. The
433 interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET
436 Only available with KVM_CAP_PPC_IRQ_LEVEL.
438 Note that any value for 'irq' other than the ones stated above is invalid
439 and incurs unexpected behavior.
443 Queues an external interrupt to be injected into the virtual CPU. A negative
444 interrupt number dequeues the interrupt.
455 Support for this has been removed. Use KVM_SET_GUEST_DEBUG instead.
463 Parameters: struct kvm_msrs (in/out)
464 Returns: 0 on success, -1 on error
466 Reads model-specific registers from the vcpu. Supported msr indices can
467 be obtained using KVM_GET_MSR_INDEX_LIST.
470 __u32 nmsrs; /* number of msrs in entries */
473 struct kvm_msr_entry entries[0];
476 struct kvm_msr_entry {
482 Application code should set the 'nmsrs' member (which indicates the
483 size of the entries array) and the 'index' member of each array entry.
484 kvm will fill in the 'data' member.
492 Parameters: struct kvm_msrs (in)
493 Returns: 0 on success, -1 on error
495 Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the
498 Application code should set the 'nmsrs' member (which indicates the
499 size of the entries array), and the 'index' and 'data' members of each
508 Parameters: struct kvm_cpuid (in)
509 Returns: 0 on success, -1 on error
511 Defines the vcpu responses to the cpuid instruction. Applications
512 should use the KVM_SET_CPUID2 ioctl if available.
515 struct kvm_cpuid_entry {
524 /* for KVM_SET_CPUID */
528 struct kvm_cpuid_entry entries[0];
532 4.21 KVM_SET_SIGNAL_MASK
537 Parameters: struct kvm_signal_mask (in)
538 Returns: 0 on success, -1 on error
540 Defines which signals are blocked during execution of KVM_RUN. This
541 signal mask temporarily overrides the threads signal mask. Any
542 unblocked signal received (except SIGKILL and SIGSTOP, which retain
543 their traditional behaviour) will cause KVM_RUN to return with -EINTR.
545 Note the signal will only be delivered if not blocked by the original
548 /* for KVM_SET_SIGNAL_MASK */
549 struct kvm_signal_mask {
560 Parameters: struct kvm_fpu (out)
561 Returns: 0 on success, -1 on error
563 Reads the floating point state from the vcpu.
565 /* for KVM_GET_FPU and KVM_SET_FPU */
570 __u8 ftwx; /* in fxsave format */
586 Parameters: struct kvm_fpu (in)
587 Returns: 0 on success, -1 on error
589 Writes the floating point state to the vcpu.
591 /* for KVM_GET_FPU and KVM_SET_FPU */
596 __u8 ftwx; /* in fxsave format */
607 4.24 KVM_CREATE_IRQCHIP
609 Capability: KVM_CAP_IRQCHIP, KVM_CAP_S390_IRQCHIP (s390)
610 Architectures: x86, ARM, arm64, s390
613 Returns: 0 on success, -1 on error
615 Creates an interrupt controller model in the kernel. On x86, creates a virtual
616 ioapic, a virtual PIC (two PICs, nested), and sets up future vcpus to have a
617 local APIC. IRQ routing for GSIs 0-15 is set to both PIC and IOAPIC; GSI 16-23
618 only go to the IOAPIC. On ARM/arm64, a GIC is
619 created. On s390, a dummy irq routing table is created.
621 Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled
622 before KVM_CREATE_IRQCHIP can be used.
627 Capability: KVM_CAP_IRQCHIP
628 Architectures: x86, arm, arm64
630 Parameters: struct kvm_irq_level
631 Returns: 0 on success, -1 on error
633 Sets the level of a GSI input to the interrupt controller model in the kernel.
634 On some architectures it is required that an interrupt controller model has
635 been previously created with KVM_CREATE_IRQCHIP. Note that edge-triggered
636 interrupts require the level to be set to 1 and then back to 0.
638 On real hardware, interrupt pins can be active-low or active-high. This
639 does not matter for the level field of struct kvm_irq_level: 1 always
640 means active (asserted), 0 means inactive (deasserted).
642 x86 allows the operating system to program the interrupt polarity
643 (active-low/active-high) for level-triggered interrupts, and KVM used
644 to consider the polarity. However, due to bitrot in the handling of
645 active-low interrupts, the above convention is now valid on x86 too.
646 This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED. Userspace
647 should not present interrupts to the guest as active-low unless this
648 capability is present (or unless it is not using the in-kernel irqchip,
652 ARM/arm64 can signal an interrupt either at the CPU level, or at the
653 in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to
654 use PPIs designated for specific cpus. The irq field is interpreted
657 Â bits: | 31 ... 24 | 23 ... 16 | 15 ... 0 |
658 field: | irq_type | vcpu_index | irq_id |
660 The irq_type field has the following values:
661 - irq_type[0]: out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ
662 - irq_type[1]: in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.)
663 (the vcpu_index field is ignored)
664 - irq_type[2]: in-kernel GIC: PPI, irq_id between 16 and 31 (incl.)
666 (The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs)
668 In both cases, level is used to assert/deassert the line.
670 struct kvm_irq_level {
673 __s32 status; /* not used for KVM_IRQ_LEVEL */
675 __u32 level; /* 0 or 1 */
681 Capability: KVM_CAP_IRQCHIP
684 Parameters: struct kvm_irqchip (in/out)
685 Returns: 0 on success, -1 on error
687 Reads the state of a kernel interrupt controller created with
688 KVM_CREATE_IRQCHIP into a buffer provided by the caller.
691 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
694 char dummy[512]; /* reserving space */
695 struct kvm_pic_state pic;
696 struct kvm_ioapic_state ioapic;
703 Capability: KVM_CAP_IRQCHIP
706 Parameters: struct kvm_irqchip (in)
707 Returns: 0 on success, -1 on error
709 Sets the state of a kernel interrupt controller created with
710 KVM_CREATE_IRQCHIP from a buffer provided by the caller.
713 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
716 char dummy[512]; /* reserving space */
717 struct kvm_pic_state pic;
718 struct kvm_ioapic_state ioapic;
723 4.28 KVM_XEN_HVM_CONFIG
725 Capability: KVM_CAP_XEN_HVM
728 Parameters: struct kvm_xen_hvm_config (in)
729 Returns: 0 on success, -1 on error
731 Sets the MSR that the Xen HVM guest uses to initialize its hypercall
732 page, and provides the starting address and size of the hypercall
733 blobs in userspace. When the guest writes the MSR, kvm copies one
734 page of a blob (32- or 64-bit, depending on the vcpu mode) to guest
737 struct kvm_xen_hvm_config {
750 Capability: KVM_CAP_ADJUST_CLOCK
753 Parameters: struct kvm_clock_data (out)
754 Returns: 0 on success, -1 on error
756 Gets the current timestamp of kvmclock as seen by the current guest. In
757 conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios
760 struct kvm_clock_data {
761 __u64 clock; /* kvmclock current value */
769 Capability: KVM_CAP_ADJUST_CLOCK
772 Parameters: struct kvm_clock_data (in)
773 Returns: 0 on success, -1 on error
775 Sets the current timestamp of kvmclock to the value specified in its parameter.
776 In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios
779 struct kvm_clock_data {
780 __u64 clock; /* kvmclock current value */
786 4.31 KVM_GET_VCPU_EVENTS
788 Capability: KVM_CAP_VCPU_EVENTS
789 Extended by: KVM_CAP_INTR_SHADOW
792 Parameters: struct kvm_vcpu_event (out)
793 Returns: 0 on success, -1 on error
795 Gets currently pending exceptions, interrupts, and NMIs as well as related
798 struct kvm_vcpu_events {
822 KVM_VCPUEVENT_VALID_SHADOW may be set in the flags field to signal that
823 interrupt.shadow contains a valid state. Otherwise, this field is undefined.
826 4.32 KVM_SET_VCPU_EVENTS
828 Capability: KVM_CAP_VCPU_EVENTS
829 Extended by: KVM_CAP_INTR_SHADOW
832 Parameters: struct kvm_vcpu_event (in)
833 Returns: 0 on success, -1 on error
835 Set pending exceptions, interrupts, and NMIs as well as related states of the
838 See KVM_GET_VCPU_EVENTS for the data structure.
840 Fields that may be modified asynchronously by running VCPUs can be excluded
841 from the update. These fields are nmi.pending and sipi_vector. Keep the
842 corresponding bits in the flags field cleared to suppress overwriting the
843 current in-kernel state. The bits are:
845 KVM_VCPUEVENT_VALID_NMI_PENDING - transfer nmi.pending to the kernel
846 KVM_VCPUEVENT_VALID_SIPI_VECTOR - transfer sipi_vector
848 If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in
849 the flags field to signal that interrupt.shadow contains a valid state and
850 shall be written into the VCPU.
853 4.33 KVM_GET_DEBUGREGS
855 Capability: KVM_CAP_DEBUGREGS
858 Parameters: struct kvm_debugregs (out)
859 Returns: 0 on success, -1 on error
861 Reads debug registers from the vcpu.
863 struct kvm_debugregs {
872 4.34 KVM_SET_DEBUGREGS
874 Capability: KVM_CAP_DEBUGREGS
877 Parameters: struct kvm_debugregs (in)
878 Returns: 0 on success, -1 on error
880 Writes debug registers into the vcpu.
882 See KVM_GET_DEBUGREGS for the data structure. The flags field is unused
883 yet and must be cleared on entry.
886 4.35 KVM_SET_USER_MEMORY_REGION
888 Capability: KVM_CAP_USER_MEM
891 Parameters: struct kvm_userspace_memory_region (in)
892 Returns: 0 on success, -1 on error
894 struct kvm_userspace_memory_region {
897 __u64 guest_phys_addr;
898 __u64 memory_size; /* bytes */
899 __u64 userspace_addr; /* start of the userspace allocated memory */
902 /* for kvm_memory_region::flags */
903 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0)
904 #define KVM_MEM_READONLY (1UL << 1)
906 This ioctl allows the user to create or modify a guest physical memory
907 slot. When changing an existing slot, it may be moved in the guest
908 physical memory space, or its flags may be modified. It may not be
909 resized. Slots may not overlap in guest physical address space.
911 Memory for the region is taken starting at the address denoted by the
912 field userspace_addr, which must point at user addressable memory for
913 the entire memory slot size. Any object may back this memory, including
914 anonymous memory, ordinary files, and hugetlbfs.
916 It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr
917 be identical. This allows large pages in the guest to be backed by large
920 The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and
921 KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of
922 writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to
923 use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it,
924 to make a new slot read-only. In this case, writes to this memory will be
925 posted to userspace as KVM_EXIT_MMIO exits.
927 When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of
928 the memory region are automatically reflected into the guest. For example, an
929 mmap() that affects the region will be made visible immediately. Another
930 example is madvise(MADV_DROP).
932 It is recommended to use this API instead of the KVM_SET_MEMORY_REGION ioctl.
933 The KVM_SET_MEMORY_REGION does not allow fine grained control over memory
934 allocation and is deprecated.
937 4.36 KVM_SET_TSS_ADDR
939 Capability: KVM_CAP_SET_TSS_ADDR
942 Parameters: unsigned long tss_address (in)
943 Returns: 0 on success, -1 on error
945 This ioctl defines the physical address of a three-page region in the guest
946 physical address space. The region must be within the first 4GB of the
947 guest physical address space and must not conflict with any memory slot
948 or any mmio address. The guest may malfunction if it accesses this memory
951 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
952 because of a quirk in the virtualization implementation (see the internals
953 documentation when it pops into existence).
958 Capability: KVM_CAP_ENABLE_CAP, KVM_CAP_ENABLE_CAP_VM
959 Architectures: ppc, s390
960 Type: vcpu ioctl, vm ioctl (with KVM_CAP_ENABLE_CAP_VM)
961 Parameters: struct kvm_enable_cap (in)
962 Returns: 0 on success; -1 on error
964 +Not all extensions are enabled by default. Using this ioctl the application
965 can enable an extension, making it available to the guest.
967 On systems that do not support this ioctl, it always fails. On systems that
968 do support it, it only works for extensions that are supported for enablement.
970 To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should
973 struct kvm_enable_cap {
977 The capability that is supposed to get enabled.
981 A bitfield indicating future enhancements. Has to be 0 for now.
985 Arguments for enabling a feature. If a feature needs initial values to
986 function properly, this is the place to put them.
991 The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl
992 for vm-wide capabilities.
994 4.38 KVM_GET_MP_STATE
996 Capability: KVM_CAP_MP_STATE
997 Architectures: x86, s390
999 Parameters: struct kvm_mp_state (out)
1000 Returns: 0 on success; -1 on error
1002 struct kvm_mp_state {
1006 Returns the vcpu's current "multiprocessing state" (though also valid on
1007 uniprocessor guests).
1009 Possible values are:
1011 - KVM_MP_STATE_RUNNABLE: the vcpu is currently running [x86]
1012 - KVM_MP_STATE_UNINITIALIZED: the vcpu is an application processor (AP)
1013 which has not yet received an INIT signal [x86]
1014 - KVM_MP_STATE_INIT_RECEIVED: the vcpu has received an INIT signal, and is
1015 now ready for a SIPI [x86]
1016 - KVM_MP_STATE_HALTED: the vcpu has executed a HLT instruction and
1017 is waiting for an interrupt [x86]
1018 - KVM_MP_STATE_SIPI_RECEIVED: the vcpu has just received a SIPI (vector
1019 accessible via KVM_GET_VCPU_EVENTS) [x86]
1020 - KVM_MP_STATE_STOPPED: the vcpu is stopped [s390]
1021 - KVM_MP_STATE_CHECK_STOP: the vcpu is in a special error state [s390]
1022 - KVM_MP_STATE_OPERATING: the vcpu is operating (running or halted)
1024 - KVM_MP_STATE_LOAD: the vcpu is in a special load/startup state
1027 On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an
1028 in-kernel irqchip, the multiprocessing state must be maintained by userspace on
1029 these architectures.
1032 4.39 KVM_SET_MP_STATE
1034 Capability: KVM_CAP_MP_STATE
1035 Architectures: x86, s390
1037 Parameters: struct kvm_mp_state (in)
1038 Returns: 0 on success; -1 on error
1040 Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for
1043 On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an
1044 in-kernel irqchip, the multiprocessing state must be maintained by userspace on
1045 these architectures.
1048 4.40 KVM_SET_IDENTITY_MAP_ADDR
1050 Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR
1053 Parameters: unsigned long identity (in)
1054 Returns: 0 on success, -1 on error
1056 This ioctl defines the physical address of a one-page region in the guest
1057 physical address space. The region must be within the first 4GB of the
1058 guest physical address space and must not conflict with any memory slot
1059 or any mmio address. The guest may malfunction if it accesses this memory
1062 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
1063 because of a quirk in the virtualization implementation (see the internals
1064 documentation when it pops into existence).
1067 4.41 KVM_SET_BOOT_CPU_ID
1069 Capability: KVM_CAP_SET_BOOT_CPU_ID
1072 Parameters: unsigned long vcpu_id
1073 Returns: 0 on success, -1 on error
1075 Define which vcpu is the Bootstrap Processor (BSP). Values are the same
1076 as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default
1082 Capability: KVM_CAP_XSAVE
1085 Parameters: struct kvm_xsave (out)
1086 Returns: 0 on success, -1 on error
1092 This ioctl would copy current vcpu's xsave struct to the userspace.
1097 Capability: KVM_CAP_XSAVE
1100 Parameters: struct kvm_xsave (in)
1101 Returns: 0 on success, -1 on error
1107 This ioctl would copy userspace's xsave struct to the kernel.
1112 Capability: KVM_CAP_XCRS
1115 Parameters: struct kvm_xcrs (out)
1116 Returns: 0 on success, -1 on error
1127 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1131 This ioctl would copy current vcpu's xcrs to the userspace.
1136 Capability: KVM_CAP_XCRS
1139 Parameters: struct kvm_xcrs (in)
1140 Returns: 0 on success, -1 on error
1151 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1155 This ioctl would set vcpu's xcr to the value userspace specified.
1158 4.46 KVM_GET_SUPPORTED_CPUID
1160 Capability: KVM_CAP_EXT_CPUID
1163 Parameters: struct kvm_cpuid2 (in/out)
1164 Returns: 0 on success, -1 on error
1169 struct kvm_cpuid_entry2 entries[0];
1172 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
1173 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
1174 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
1176 struct kvm_cpuid_entry2 {
1187 This ioctl returns x86 cpuid features which are supported by both the hardware
1188 and kvm. Userspace can use the information returned by this ioctl to
1189 construct cpuid information (for KVM_SET_CPUID2) that is consistent with
1190 hardware, kernel, and userspace capabilities, and with user requirements (for
1191 example, the user may wish to constrain cpuid to emulate older hardware,
1192 or for feature consistency across a cluster).
1194 Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure
1195 with the 'nent' field indicating the number of entries in the variable-size
1196 array 'entries'. If the number of entries is too low to describe the cpu
1197 capabilities, an error (E2BIG) is returned. If the number is too high,
1198 the 'nent' field is adjusted and an error (ENOMEM) is returned. If the
1199 number is just right, the 'nent' field is adjusted to the number of valid
1200 entries in the 'entries' array, which is then filled.
1202 The entries returned are the host cpuid as returned by the cpuid instruction,
1203 with unknown or unsupported features masked out. Some features (for example,
1204 x2apic), may not be present in the host cpu, but are exposed by kvm if it can
1205 emulate them efficiently. The fields in each entry are defined as follows:
1207 function: the eax value used to obtain the entry
1208 index: the ecx value used to obtain the entry (for entries that are
1210 flags: an OR of zero or more of the following:
1211 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
1212 if the index field is valid
1213 KVM_CPUID_FLAG_STATEFUL_FUNC:
1214 if cpuid for this function returns different values for successive
1215 invocations; there will be several entries with the same function,
1216 all with this flag set
1217 KVM_CPUID_FLAG_STATE_READ_NEXT:
1218 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
1219 the first entry to be read by a cpu
1220 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
1221 this function/index combination
1223 The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned
1224 as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC
1225 support. Instead it is reported via
1227 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER)
1229 if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the
1230 feature in userspace, then you can enable the feature for KVM_SET_CPUID2.
1233 4.47 KVM_PPC_GET_PVINFO
1235 Capability: KVM_CAP_PPC_GET_PVINFO
1238 Parameters: struct kvm_ppc_pvinfo (out)
1239 Returns: 0 on success, !0 on error
1241 struct kvm_ppc_pvinfo {
1247 This ioctl fetches PV specific information that need to be passed to the guest
1248 using the device tree or other means from vm context.
1250 The hcall array defines 4 instructions that make up a hypercall.
1252 If any additional field gets added to this structure later on, a bit for that
1253 additional piece of information will be set in the flags bitmap.
1255 The flags bitmap is defined as:
1257 /* the host supports the ePAPR idle hcall
1258 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0)
1260 4.48 KVM_ASSIGN_PCI_DEVICE
1265 Parameters: struct kvm_assigned_pci_dev (in)
1266 Returns: 0 on success, -1 on error
1268 Assigns a host PCI device to the VM.
1270 struct kvm_assigned_pci_dev {
1271 __u32 assigned_dev_id;
1281 The PCI device is specified by the triple segnr, busnr, and devfn.
1282 Identification in succeeding service requests is done via assigned_dev_id. The
1283 following flags are specified:
1285 /* Depends on KVM_CAP_IOMMU */
1286 #define KVM_DEV_ASSIGN_ENABLE_IOMMU (1 << 0)
1287 /* The following two depend on KVM_CAP_PCI_2_3 */
1288 #define KVM_DEV_ASSIGN_PCI_2_3 (1 << 1)
1289 #define KVM_DEV_ASSIGN_MASK_INTX (1 << 2)
1291 If KVM_DEV_ASSIGN_PCI_2_3 is set, the kernel will manage legacy INTx interrupts
1292 via the PCI-2.3-compliant device-level mask, thus enable IRQ sharing with other
1293 assigned devices or host devices. KVM_DEV_ASSIGN_MASK_INTX specifies the
1294 guest's view on the INTx mask, see KVM_ASSIGN_SET_INTX_MASK for details.
1296 The KVM_DEV_ASSIGN_ENABLE_IOMMU flag is a mandatory option to ensure
1297 isolation of the device. Usages not specifying this flag are deprecated.
1299 Only PCI header type 0 devices with PCI BAR resources are supported by
1300 device assignment. The user requesting this ioctl must have read/write
1301 access to the PCI sysfs resource files associated with the device.
1304 ENOTTY: kernel does not support this ioctl
1306 Other error conditions may be defined by individual device types or
1307 have their standard meanings.
1310 4.49 KVM_DEASSIGN_PCI_DEVICE
1315 Parameters: struct kvm_assigned_pci_dev (in)
1316 Returns: 0 on success, -1 on error
1318 Ends PCI device assignment, releasing all associated resources.
1320 See KVM_ASSIGN_PCI_DEVICE for the data structure. Only assigned_dev_id is
1321 used in kvm_assigned_pci_dev to identify the device.
1324 ENOTTY: kernel does not support this ioctl
1326 Other error conditions may be defined by individual device types or
1327 have their standard meanings.
1329 4.50 KVM_ASSIGN_DEV_IRQ
1331 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1334 Parameters: struct kvm_assigned_irq (in)
1335 Returns: 0 on success, -1 on error
1337 Assigns an IRQ to a passed-through device.
1339 struct kvm_assigned_irq {
1340 __u32 assigned_dev_id;
1341 __u32 host_irq; /* ignored (legacy field) */
1349 The following flags are defined:
1351 #define KVM_DEV_IRQ_HOST_INTX (1 << 0)
1352 #define KVM_DEV_IRQ_HOST_MSI (1 << 1)
1353 #define KVM_DEV_IRQ_HOST_MSIX (1 << 2)
1355 #define KVM_DEV_IRQ_GUEST_INTX (1 << 8)
1356 #define KVM_DEV_IRQ_GUEST_MSI (1 << 9)
1357 #define KVM_DEV_IRQ_GUEST_MSIX (1 << 10)
1359 It is not valid to specify multiple types per host or guest IRQ. However, the
1360 IRQ type of host and guest can differ or can even be null.
1363 ENOTTY: kernel does not support this ioctl
1365 Other error conditions may be defined by individual device types or
1366 have their standard meanings.
1369 4.51 KVM_DEASSIGN_DEV_IRQ
1371 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1374 Parameters: struct kvm_assigned_irq (in)
1375 Returns: 0 on success, -1 on error
1377 Ends an IRQ assignment to a passed-through device.
1379 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1380 by assigned_dev_id, flags must correspond to the IRQ type specified on
1381 KVM_ASSIGN_DEV_IRQ. Partial deassignment of host or guest IRQ is allowed.
1384 4.52 KVM_SET_GSI_ROUTING
1386 Capability: KVM_CAP_IRQ_ROUTING
1387 Architectures: x86 s390
1389 Parameters: struct kvm_irq_routing (in)
1390 Returns: 0 on success, -1 on error
1392 Sets the GSI routing table entries, overwriting any previously set entries.
1394 struct kvm_irq_routing {
1397 struct kvm_irq_routing_entry entries[0];
1400 No flags are specified so far, the corresponding field must be set to zero.
1402 struct kvm_irq_routing_entry {
1408 struct kvm_irq_routing_irqchip irqchip;
1409 struct kvm_irq_routing_msi msi;
1410 struct kvm_irq_routing_s390_adapter adapter;
1415 /* gsi routing entry types */
1416 #define KVM_IRQ_ROUTING_IRQCHIP 1
1417 #define KVM_IRQ_ROUTING_MSI 2
1418 #define KVM_IRQ_ROUTING_S390_ADAPTER 3
1420 No flags are specified so far, the corresponding field must be set to zero.
1422 struct kvm_irq_routing_irqchip {
1427 struct kvm_irq_routing_msi {
1434 struct kvm_irq_routing_s390_adapter {
1438 __u32 summary_offset;
1443 4.53 KVM_ASSIGN_SET_MSIX_NR
1448 Parameters: struct kvm_assigned_msix_nr (in)
1449 Returns: 0 on success, -1 on error
1451 Set the number of MSI-X interrupts for an assigned device. The number is
1452 reset again by terminating the MSI-X assignment of the device via
1453 KVM_DEASSIGN_DEV_IRQ. Calling this service more than once at any earlier
1456 struct kvm_assigned_msix_nr {
1457 __u32 assigned_dev_id;
1462 #define KVM_MAX_MSIX_PER_DEV 256
1465 4.54 KVM_ASSIGN_SET_MSIX_ENTRY
1470 Parameters: struct kvm_assigned_msix_entry (in)
1471 Returns: 0 on success, -1 on error
1473 Specifies the routing of an MSI-X assigned device interrupt to a GSI. Setting
1474 the GSI vector to zero means disabling the interrupt.
1476 struct kvm_assigned_msix_entry {
1477 __u32 assigned_dev_id;
1479 __u16 entry; /* The index of entry in the MSI-X table */
1484 ENOTTY: kernel does not support this ioctl
1486 Other error conditions may be defined by individual device types or
1487 have their standard meanings.
1490 4.55 KVM_SET_TSC_KHZ
1492 Capability: KVM_CAP_TSC_CONTROL
1495 Parameters: virtual tsc_khz
1496 Returns: 0 on success, -1 on error
1498 Specifies the tsc frequency for the virtual machine. The unit of the
1502 4.56 KVM_GET_TSC_KHZ
1504 Capability: KVM_CAP_GET_TSC_KHZ
1508 Returns: virtual tsc-khz on success, negative value on error
1510 Returns the tsc frequency of the guest. The unit of the return value is
1511 KHz. If the host has unstable tsc this ioctl returns -EIO instead as an
1517 Capability: KVM_CAP_IRQCHIP
1520 Parameters: struct kvm_lapic_state (out)
1521 Returns: 0 on success, -1 on error
1523 #define KVM_APIC_REG_SIZE 0x400
1524 struct kvm_lapic_state {
1525 char regs[KVM_APIC_REG_SIZE];
1528 Reads the Local APIC registers and copies them into the input argument. The
1529 data format and layout are the same as documented in the architecture manual.
1534 Capability: KVM_CAP_IRQCHIP
1537 Parameters: struct kvm_lapic_state (in)
1538 Returns: 0 on success, -1 on error
1540 #define KVM_APIC_REG_SIZE 0x400
1541 struct kvm_lapic_state {
1542 char regs[KVM_APIC_REG_SIZE];
1545 Copies the input argument into the Local APIC registers. The data format
1546 and layout are the same as documented in the architecture manual.
1551 Capability: KVM_CAP_IOEVENTFD
1554 Parameters: struct kvm_ioeventfd (in)
1555 Returns: 0 on success, !0 on error
1557 This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address
1558 within the guest. A guest write in the registered address will signal the
1559 provided event instead of triggering an exit.
1561 struct kvm_ioeventfd {
1563 __u64 addr; /* legal pio/mmio address */
1564 __u32 len; /* 1, 2, 4, or 8 bytes */
1570 For the special case of virtio-ccw devices on s390, the ioevent is matched
1571 to a subchannel/virtqueue tuple instead.
1573 The following flags are defined:
1575 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch)
1576 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio)
1577 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign)
1578 #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \
1579 (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify)
1581 If datamatch flag is set, the event will be signaled only if the written value
1582 to the registered address is equal to datamatch in struct kvm_ioeventfd.
1584 For virtio-ccw devices, addr contains the subchannel id and datamatch the
1590 Capability: KVM_CAP_SW_TLB
1593 Parameters: struct kvm_dirty_tlb (in)
1594 Returns: 0 on success, -1 on error
1596 struct kvm_dirty_tlb {
1601 This must be called whenever userspace has changed an entry in the shared
1602 TLB, prior to calling KVM_RUN on the associated vcpu.
1604 The "bitmap" field is the userspace address of an array. This array
1605 consists of a number of bits, equal to the total number of TLB entries as
1606 determined by the last successful call to KVM_CONFIG_TLB, rounded up to the
1607 nearest multiple of 64.
1609 Each bit corresponds to one TLB entry, ordered the same as in the shared TLB
1612 The array is little-endian: the bit 0 is the least significant bit of the
1613 first byte, bit 8 is the least significant bit of the second byte, etc.
1614 This avoids any complications with differing word sizes.
1616 The "num_dirty" field is a performance hint for KVM to determine whether it
1617 should skip processing the bitmap and just invalidate everything. It must
1618 be set to the number of set bits in the bitmap.
1621 4.61 KVM_ASSIGN_SET_INTX_MASK
1623 Capability: KVM_CAP_PCI_2_3
1626 Parameters: struct kvm_assigned_pci_dev (in)
1627 Returns: 0 on success, -1 on error
1629 Allows userspace to mask PCI INTx interrupts from the assigned device. The
1630 kernel will not deliver INTx interrupts to the guest between setting and
1631 clearing of KVM_ASSIGN_SET_INTX_MASK via this interface. This enables use of
1632 and emulation of PCI 2.3 INTx disable command register behavior.
1634 This may be used for both PCI 2.3 devices supporting INTx disable natively and
1635 older devices lacking this support. Userspace is responsible for emulating the
1636 read value of the INTx disable bit in the guest visible PCI command register.
1637 When modifying the INTx disable state, userspace should precede updating the
1638 physical device command register by calling this ioctl to inform the kernel of
1639 the new intended INTx mask state.
1641 Note that the kernel uses the device INTx disable bit to internally manage the
1642 device interrupt state for PCI 2.3 devices. Reads of this register may
1643 therefore not match the expected value. Writes should always use the guest
1644 intended INTx disable value rather than attempting to read-copy-update the
1645 current physical device state. Races between user and kernel updates to the
1646 INTx disable bit are handled lazily in the kernel. It's possible the device
1647 may generate unintended interrupts, but they will not be injected into the
1650 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1651 by assigned_dev_id. In the flags field, only KVM_DEV_ASSIGN_MASK_INTX is
1655 4.62 KVM_CREATE_SPAPR_TCE
1657 Capability: KVM_CAP_SPAPR_TCE
1658 Architectures: powerpc
1660 Parameters: struct kvm_create_spapr_tce (in)
1661 Returns: file descriptor for manipulating the created TCE table
1663 This creates a virtual TCE (translation control entry) table, which
1664 is an IOMMU for PAPR-style virtual I/O. It is used to translate
1665 logical addresses used in virtual I/O into guest physical addresses,
1666 and provides a scatter/gather capability for PAPR virtual I/O.
1668 /* for KVM_CAP_SPAPR_TCE */
1669 struct kvm_create_spapr_tce {
1674 The liobn field gives the logical IO bus number for which to create a
1675 TCE table. The window_size field specifies the size of the DMA window
1676 which this TCE table will translate - the table will contain one 64
1677 bit TCE entry for every 4kiB of the DMA window.
1679 When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE
1680 table has been created using this ioctl(), the kernel will handle it
1681 in real mode, updating the TCE table. H_PUT_TCE calls for other
1682 liobns will cause a vm exit and must be handled by userspace.
1684 The return value is a file descriptor which can be passed to mmap(2)
1685 to map the created TCE table into userspace. This lets userspace read
1686 the entries written by kernel-handled H_PUT_TCE calls, and also lets
1687 userspace update the TCE table directly which is useful in some
1691 4.63 KVM_ALLOCATE_RMA
1693 Capability: KVM_CAP_PPC_RMA
1694 Architectures: powerpc
1696 Parameters: struct kvm_allocate_rma (out)
1697 Returns: file descriptor for mapping the allocated RMA
1699 This allocates a Real Mode Area (RMA) from the pool allocated at boot
1700 time by the kernel. An RMA is a physically-contiguous, aligned region
1701 of memory used on older POWER processors to provide the memory which
1702 will be accessed by real-mode (MMU off) accesses in a KVM guest.
1703 POWER processors support a set of sizes for the RMA that usually
1704 includes 64MB, 128MB, 256MB and some larger powers of two.
1706 /* for KVM_ALLOCATE_RMA */
1707 struct kvm_allocate_rma {
1711 The return value is a file descriptor which can be passed to mmap(2)
1712 to map the allocated RMA into userspace. The mapped area can then be
1713 passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the
1714 RMA for a virtual machine. The size of the RMA in bytes (which is
1715 fixed at host kernel boot time) is returned in the rma_size field of
1716 the argument structure.
1718 The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl
1719 is supported; 2 if the processor requires all virtual machines to have
1720 an RMA, or 1 if the processor can use an RMA but doesn't require it,
1721 because it supports the Virtual RMA (VRMA) facility.
1726 Capability: KVM_CAP_USER_NMI
1730 Returns: 0 on success, -1 on error
1732 Queues an NMI on the thread's vcpu. Note this is well defined only
1733 when KVM_CREATE_IRQCHIP has not been called, since this is an interface
1734 between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP
1735 has been called, this interface is completely emulated within the kernel.
1737 To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the
1738 following algorithm:
1741 - read the local APIC's state (KVM_GET_LAPIC)
1742 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1)
1743 - if so, issue KVM_NMI
1746 Some guests configure the LINT1 NMI input to cause a panic, aiding in
1750 4.65 KVM_S390_UCAS_MAP
1752 Capability: KVM_CAP_S390_UCONTROL
1755 Parameters: struct kvm_s390_ucas_mapping (in)
1756 Returns: 0 in case of success
1758 The parameter is defined like this:
1759 struct kvm_s390_ucas_mapping {
1765 This ioctl maps the memory at "user_addr" with the length "length" to
1766 the vcpu's address space starting at "vcpu_addr". All parameters need to
1767 be aligned by 1 megabyte.
1770 4.66 KVM_S390_UCAS_UNMAP
1772 Capability: KVM_CAP_S390_UCONTROL
1775 Parameters: struct kvm_s390_ucas_mapping (in)
1776 Returns: 0 in case of success
1778 The parameter is defined like this:
1779 struct kvm_s390_ucas_mapping {
1785 This ioctl unmaps the memory in the vcpu's address space starting at
1786 "vcpu_addr" with the length "length". The field "user_addr" is ignored.
1787 All parameters need to be aligned by 1 megabyte.
1790 4.67 KVM_S390_VCPU_FAULT
1792 Capability: KVM_CAP_S390_UCONTROL
1795 Parameters: vcpu absolute address (in)
1796 Returns: 0 in case of success
1798 This call creates a page table entry on the virtual cpu's address space
1799 (for user controlled virtual machines) or the virtual machine's address
1800 space (for regular virtual machines). This only works for minor faults,
1801 thus it's recommended to access subject memory page via the user page
1802 table upfront. This is useful to handle validity intercepts for user
1803 controlled virtual machines to fault in the virtual cpu's lowcore pages
1804 prior to calling the KVM_RUN ioctl.
1807 4.68 KVM_SET_ONE_REG
1809 Capability: KVM_CAP_ONE_REG
1812 Parameters: struct kvm_one_reg (in)
1813 Returns: 0 on success, negative value on failure
1815 struct kvm_one_reg {
1820 Using this ioctl, a single vcpu register can be set to a specific value
1821 defined by user space with the passed in struct kvm_one_reg, where id
1822 refers to the register identifier as described below and addr is a pointer
1823 to a variable with the respective size. There can be architecture agnostic
1824 and architecture specific registers. Each have their own range of operation
1825 and their own constants and width. To keep track of the implemented
1826 registers, find a list below:
1828 Arch | Register | Width (bits)
1830 PPC | KVM_REG_PPC_HIOR | 64
1831 PPC | KVM_REG_PPC_IAC1 | 64
1832 PPC | KVM_REG_PPC_IAC2 | 64
1833 PPC | KVM_REG_PPC_IAC3 | 64
1834 PPC | KVM_REG_PPC_IAC4 | 64
1835 PPC | KVM_REG_PPC_DAC1 | 64
1836 PPC | KVM_REG_PPC_DAC2 | 64
1837 PPC | KVM_REG_PPC_DABR | 64
1838 PPC | KVM_REG_PPC_DSCR | 64
1839 PPC | KVM_REG_PPC_PURR | 64
1840 PPC | KVM_REG_PPC_SPURR | 64
1841 PPC | KVM_REG_PPC_DAR | 64
1842 PPC | KVM_REG_PPC_DSISR | 32
1843 PPC | KVM_REG_PPC_AMR | 64
1844 PPC | KVM_REG_PPC_UAMOR | 64
1845 PPC | KVM_REG_PPC_MMCR0 | 64
1846 PPC | KVM_REG_PPC_MMCR1 | 64
1847 PPC | KVM_REG_PPC_MMCRA | 64
1848 PPC | KVM_REG_PPC_MMCR2 | 64
1849 PPC | KVM_REG_PPC_MMCRS | 64
1850 PPC | KVM_REG_PPC_SIAR | 64
1851 PPC | KVM_REG_PPC_SDAR | 64
1852 PPC | KVM_REG_PPC_SIER | 64
1853 PPC | KVM_REG_PPC_PMC1 | 32
1854 PPC | KVM_REG_PPC_PMC2 | 32
1855 PPC | KVM_REG_PPC_PMC3 | 32
1856 PPC | KVM_REG_PPC_PMC4 | 32
1857 PPC | KVM_REG_PPC_PMC5 | 32
1858 PPC | KVM_REG_PPC_PMC6 | 32
1859 PPC | KVM_REG_PPC_PMC7 | 32
1860 PPC | KVM_REG_PPC_PMC8 | 32
1861 PPC | KVM_REG_PPC_FPR0 | 64
1863 PPC | KVM_REG_PPC_FPR31 | 64
1864 PPC | KVM_REG_PPC_VR0 | 128
1866 PPC | KVM_REG_PPC_VR31 | 128
1867 PPC | KVM_REG_PPC_VSR0 | 128
1869 PPC | KVM_REG_PPC_VSR31 | 128
1870 PPC | KVM_REG_PPC_FPSCR | 64
1871 PPC | KVM_REG_PPC_VSCR | 32
1872 PPC | KVM_REG_PPC_VPA_ADDR | 64
1873 PPC | KVM_REG_PPC_VPA_SLB | 128
1874 PPC | KVM_REG_PPC_VPA_DTL | 128
1875 PPC | KVM_REG_PPC_EPCR | 32
1876 PPC | KVM_REG_PPC_EPR | 32
1877 PPC | KVM_REG_PPC_TCR | 32
1878 PPC | KVM_REG_PPC_TSR | 32
1879 PPC | KVM_REG_PPC_OR_TSR | 32
1880 PPC | KVM_REG_PPC_CLEAR_TSR | 32
1881 PPC | KVM_REG_PPC_MAS0 | 32
1882 PPC | KVM_REG_PPC_MAS1 | 32
1883 PPC | KVM_REG_PPC_MAS2 | 64
1884 PPC | KVM_REG_PPC_MAS7_3 | 64
1885 PPC | KVM_REG_PPC_MAS4 | 32
1886 PPC | KVM_REG_PPC_MAS6 | 32
1887 PPC | KVM_REG_PPC_MMUCFG | 32
1888 PPC | KVM_REG_PPC_TLB0CFG | 32
1889 PPC | KVM_REG_PPC_TLB1CFG | 32
1890 PPC | KVM_REG_PPC_TLB2CFG | 32
1891 PPC | KVM_REG_PPC_TLB3CFG | 32
1892 PPC | KVM_REG_PPC_TLB0PS | 32
1893 PPC | KVM_REG_PPC_TLB1PS | 32
1894 PPC | KVM_REG_PPC_TLB2PS | 32
1895 PPC | KVM_REG_PPC_TLB3PS | 32
1896 PPC | KVM_REG_PPC_EPTCFG | 32
1897 PPC | KVM_REG_PPC_ICP_STATE | 64
1898 PPC | KVM_REG_PPC_TB_OFFSET | 64
1899 PPC | KVM_REG_PPC_SPMC1 | 32
1900 PPC | KVM_REG_PPC_SPMC2 | 32
1901 PPC | KVM_REG_PPC_IAMR | 64
1902 PPC | KVM_REG_PPC_TFHAR | 64
1903 PPC | KVM_REG_PPC_TFIAR | 64
1904 PPC | KVM_REG_PPC_TEXASR | 64
1905 PPC | KVM_REG_PPC_FSCR | 64
1906 PPC | KVM_REG_PPC_PSPB | 32
1907 PPC | KVM_REG_PPC_EBBHR | 64
1908 PPC | KVM_REG_PPC_EBBRR | 64
1909 PPC | KVM_REG_PPC_BESCR | 64
1910 PPC | KVM_REG_PPC_TAR | 64
1911 PPC | KVM_REG_PPC_DPDES | 64
1912 PPC | KVM_REG_PPC_DAWR | 64
1913 PPC | KVM_REG_PPC_DAWRX | 64
1914 PPC | KVM_REG_PPC_CIABR | 64
1915 PPC | KVM_REG_PPC_IC | 64
1916 PPC | KVM_REG_PPC_VTB | 64
1917 PPC | KVM_REG_PPC_CSIGR | 64
1918 PPC | KVM_REG_PPC_TACR | 64
1919 PPC | KVM_REG_PPC_TCSCR | 64
1920 PPC | KVM_REG_PPC_PID | 64
1921 PPC | KVM_REG_PPC_ACOP | 64
1922 PPC | KVM_REG_PPC_VRSAVE | 32
1923 PPC | KVM_REG_PPC_LPCR | 32
1924 PPC | KVM_REG_PPC_LPCR_64 | 64
1925 PPC | KVM_REG_PPC_PPR | 64
1926 PPC | KVM_REG_PPC_ARCH_COMPAT | 32
1927 PPC | KVM_REG_PPC_DABRX | 32
1928 PPC | KVM_REG_PPC_WORT | 64
1929 PPC | KVM_REG_PPC_SPRG9 | 64
1930 PPC | KVM_REG_PPC_DBSR | 32
1931 PPC | KVM_REG_PPC_TM_GPR0 | 64
1933 PPC | KVM_REG_PPC_TM_GPR31 | 64
1934 PPC | KVM_REG_PPC_TM_VSR0 | 128
1936 PPC | KVM_REG_PPC_TM_VSR63 | 128
1937 PPC | KVM_REG_PPC_TM_CR | 64
1938 PPC | KVM_REG_PPC_TM_LR | 64
1939 PPC | KVM_REG_PPC_TM_CTR | 64
1940 PPC | KVM_REG_PPC_TM_FPSCR | 64
1941 PPC | KVM_REG_PPC_TM_AMR | 64
1942 PPC | KVM_REG_PPC_TM_PPR | 64
1943 PPC | KVM_REG_PPC_TM_VRSAVE | 64
1944 PPC | KVM_REG_PPC_TM_VSCR | 32
1945 PPC | KVM_REG_PPC_TM_DSCR | 64
1946 PPC | KVM_REG_PPC_TM_TAR | 64
1948 MIPS | KVM_REG_MIPS_R0 | 64
1950 MIPS | KVM_REG_MIPS_R31 | 64
1951 MIPS | KVM_REG_MIPS_HI | 64
1952 MIPS | KVM_REG_MIPS_LO | 64
1953 MIPS | KVM_REG_MIPS_PC | 64
1954 MIPS | KVM_REG_MIPS_CP0_INDEX | 32
1955 MIPS | KVM_REG_MIPS_CP0_CONTEXT | 64
1956 MIPS | KVM_REG_MIPS_CP0_USERLOCAL | 64
1957 MIPS | KVM_REG_MIPS_CP0_PAGEMASK | 32
1958 MIPS | KVM_REG_MIPS_CP0_WIRED | 32
1959 MIPS | KVM_REG_MIPS_CP0_HWRENA | 32
1960 MIPS | KVM_REG_MIPS_CP0_BADVADDR | 64
1961 MIPS | KVM_REG_MIPS_CP0_COUNT | 32
1962 MIPS | KVM_REG_MIPS_CP0_ENTRYHI | 64
1963 MIPS | KVM_REG_MIPS_CP0_COMPARE | 32
1964 MIPS | KVM_REG_MIPS_CP0_STATUS | 32
1965 MIPS | KVM_REG_MIPS_CP0_CAUSE | 32
1966 MIPS | KVM_REG_MIPS_CP0_EPC | 64
1967 MIPS | KVM_REG_MIPS_CP0_CONFIG | 32
1968 MIPS | KVM_REG_MIPS_CP0_CONFIG1 | 32
1969 MIPS | KVM_REG_MIPS_CP0_CONFIG2 | 32
1970 MIPS | KVM_REG_MIPS_CP0_CONFIG3 | 32
1971 MIPS | KVM_REG_MIPS_CP0_CONFIG7 | 32
1972 MIPS | KVM_REG_MIPS_CP0_ERROREPC | 64
1973 MIPS | KVM_REG_MIPS_COUNT_CTL | 64
1974 MIPS | KVM_REG_MIPS_COUNT_RESUME | 64
1975 MIPS | KVM_REG_MIPS_COUNT_HZ | 64
1977 ARM registers are mapped using the lower 32 bits. The upper 16 of that
1978 is the register group type, or coprocessor number:
1980 ARM core registers have the following id bit patterns:
1981 0x4020 0000 0010 <index into the kvm_regs struct:16>
1983 ARM 32-bit CP15 registers have the following id bit patterns:
1984 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3>
1986 ARM 64-bit CP15 registers have the following id bit patterns:
1987 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3>
1989 ARM CCSIDR registers are demultiplexed by CSSELR value:
1990 0x4020 0000 0011 00 <csselr:8>
1992 ARM 32-bit VFP control registers have the following id bit patterns:
1993 0x4020 0000 0012 1 <regno:12>
1995 ARM 64-bit FP registers have the following id bit patterns:
1996 0x4030 0000 0012 0 <regno:12>
1999 arm64 registers are mapped using the lower 32 bits. The upper 16 of
2000 that is the register group type, or coprocessor number:
2002 arm64 core/FP-SIMD registers have the following id bit patterns. Note
2003 that the size of the access is variable, as the kvm_regs structure
2004 contains elements ranging from 32 to 128 bits. The index is a 32bit
2005 value in the kvm_regs structure seen as a 32bit array.
2006 0x60x0 0000 0010 <index into the kvm_regs struct:16>
2008 arm64 CCSIDR registers are demultiplexed by CSSELR value:
2009 0x6020 0000 0011 00 <csselr:8>
2011 arm64 system registers have the following id bit patterns:
2012 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3>
2015 MIPS registers are mapped using the lower 32 bits. The upper 16 of that is
2016 the register group type:
2018 MIPS core registers (see above) have the following id bit patterns:
2019 0x7030 0000 0000 <reg:16>
2021 MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit
2022 patterns depending on whether they're 32-bit or 64-bit registers:
2023 0x7020 0000 0001 00 <reg:5> <sel:3> (32-bit)
2024 0x7030 0000 0001 00 <reg:5> <sel:3> (64-bit)
2026 MIPS KVM control registers (see above) have the following id bit patterns:
2027 0x7030 0000 0002 <reg:16>
2030 4.69 KVM_GET_ONE_REG
2032 Capability: KVM_CAP_ONE_REG
2035 Parameters: struct kvm_one_reg (in and out)
2036 Returns: 0 on success, negative value on failure
2038 This ioctl allows to receive the value of a single register implemented
2039 in a vcpu. The register to read is indicated by the "id" field of the
2040 kvm_one_reg struct passed in. On success, the register value can be found
2041 at the memory location pointed to by "addr".
2043 The list of registers accessible using this interface is identical to the
2047 4.70 KVM_KVMCLOCK_CTRL
2049 Capability: KVM_CAP_KVMCLOCK_CTRL
2050 Architectures: Any that implement pvclocks (currently x86 only)
2053 Returns: 0 on success, -1 on error
2055 This signals to the host kernel that the specified guest is being paused by
2056 userspace. The host will set a flag in the pvclock structure that is checked
2057 from the soft lockup watchdog. The flag is part of the pvclock structure that
2058 is shared between guest and host, specifically the second bit of the flags
2059 field of the pvclock_vcpu_time_info structure. It will be set exclusively by
2060 the host and read/cleared exclusively by the guest. The guest operation of
2061 checking and clearing the flag must an atomic operation so
2062 load-link/store-conditional, or equivalent must be used. There are two cases
2063 where the guest will clear the flag: when the soft lockup watchdog timer resets
2064 itself or when a soft lockup is detected. This ioctl can be called any time
2065 after pausing the vcpu, but before it is resumed.
2070 Capability: KVM_CAP_SIGNAL_MSI
2073 Parameters: struct kvm_msi (in)
2074 Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error
2076 Directly inject a MSI message. Only valid with in-kernel irqchip that handles
2087 No flags are defined so far. The corresponding field must be 0.
2090 4.71 KVM_CREATE_PIT2
2092 Capability: KVM_CAP_PIT2
2095 Parameters: struct kvm_pit_config (in)
2096 Returns: 0 on success, -1 on error
2098 Creates an in-kernel device model for the i8254 PIT. This call is only valid
2099 after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following
2100 parameters have to be passed:
2102 struct kvm_pit_config {
2109 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */
2111 PIT timer interrupts may use a per-VM kernel thread for injection. If it
2112 exists, this thread will have a name of the following pattern:
2114 kvm-pit/<owner-process-pid>
2116 When running a guest with elevated priorities, the scheduling parameters of
2117 this thread may have to be adjusted accordingly.
2119 This IOCTL replaces the obsolete KVM_CREATE_PIT.
2124 Capability: KVM_CAP_PIT_STATE2
2127 Parameters: struct kvm_pit_state2 (out)
2128 Returns: 0 on success, -1 on error
2130 Retrieves the state of the in-kernel PIT model. Only valid after
2131 KVM_CREATE_PIT2. The state is returned in the following structure:
2133 struct kvm_pit_state2 {
2134 struct kvm_pit_channel_state channels[3];
2141 /* disable PIT in HPET legacy mode */
2142 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001
2144 This IOCTL replaces the obsolete KVM_GET_PIT.
2149 Capability: KVM_CAP_PIT_STATE2
2152 Parameters: struct kvm_pit_state2 (in)
2153 Returns: 0 on success, -1 on error
2155 Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2.
2156 See KVM_GET_PIT2 for details on struct kvm_pit_state2.
2158 This IOCTL replaces the obsolete KVM_SET_PIT.
2161 4.74 KVM_PPC_GET_SMMU_INFO
2163 Capability: KVM_CAP_PPC_GET_SMMU_INFO
2164 Architectures: powerpc
2167 Returns: 0 on success, -1 on error
2169 This populates and returns a structure describing the features of
2170 the "Server" class MMU emulation supported by KVM.
2171 This can in turn be used by userspace to generate the appropriate
2172 device-tree properties for the guest operating system.
2174 The structure contains some global information, followed by an
2175 array of supported segment page sizes:
2177 struct kvm_ppc_smmu_info {
2181 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ];
2184 The supported flags are:
2186 - KVM_PPC_PAGE_SIZES_REAL:
2187 When that flag is set, guest page sizes must "fit" the backing
2188 store page sizes. When not set, any page size in the list can
2189 be used regardless of how they are backed by userspace.
2191 - KVM_PPC_1T_SEGMENTS
2192 The emulated MMU supports 1T segments in addition to the
2195 The "slb_size" field indicates how many SLB entries are supported
2197 The "sps" array contains 8 entries indicating the supported base
2198 page sizes for a segment in increasing order. Each entry is defined
2201 struct kvm_ppc_one_seg_page_size {
2202 __u32 page_shift; /* Base page shift of segment (or 0) */
2203 __u32 slb_enc; /* SLB encoding for BookS */
2204 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ];
2207 An entry with a "page_shift" of 0 is unused. Because the array is
2208 organized in increasing order, a lookup can stop when encoutering
2211 The "slb_enc" field provides the encoding to use in the SLB for the
2212 page size. The bits are in positions such as the value can directly
2213 be OR'ed into the "vsid" argument of the slbmte instruction.
2215 The "enc" array is a list which for each of those segment base page
2216 size provides the list of supported actual page sizes (which can be
2217 only larger or equal to the base page size), along with the
2218 corresponding encoding in the hash PTE. Similarly, the array is
2219 8 entries sorted by increasing sizes and an entry with a "0" shift
2220 is an empty entry and a terminator:
2222 struct kvm_ppc_one_page_size {
2223 __u32 page_shift; /* Page shift (or 0) */
2224 __u32 pte_enc; /* Encoding in the HPTE (>>12) */
2227 The "pte_enc" field provides a value that can OR'ed into the hash
2228 PTE's RPN field (ie, it needs to be shifted left by 12 to OR it
2229 into the hash PTE second double word).
2233 Capability: KVM_CAP_IRQFD
2234 Architectures: x86 s390
2236 Parameters: struct kvm_irqfd (in)
2237 Returns: 0 on success, -1 on error
2239 Allows setting an eventfd to directly trigger a guest interrupt.
2240 kvm_irqfd.fd specifies the file descriptor to use as the eventfd and
2241 kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When
2242 an event is triggered on the eventfd, an interrupt is injected into
2243 the guest using the specified gsi pin. The irqfd is removed using
2244 the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd
2247 With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify
2248 mechanism allowing emulation of level-triggered, irqfd-based
2249 interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an
2250 additional eventfd in the kvm_irqfd.resamplefd field. When operating
2251 in resample mode, posting of an interrupt through kvm_irq.fd asserts
2252 the specified gsi in the irqchip. When the irqchip is resampled, such
2253 as from an EOI, the gsi is de-asserted and the user is notified via
2254 kvm_irqfd.resamplefd. It is the user's responsibility to re-queue
2255 the interrupt if the device making use of it still requires service.
2256 Note that closing the resamplefd is not sufficient to disable the
2257 irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment
2258 and need not be specified with KVM_IRQFD_FLAG_DEASSIGN.
2260 4.76 KVM_PPC_ALLOCATE_HTAB
2262 Capability: KVM_CAP_PPC_ALLOC_HTAB
2263 Architectures: powerpc
2265 Parameters: Pointer to u32 containing hash table order (in/out)
2266 Returns: 0 on success, -1 on error
2268 This requests the host kernel to allocate an MMU hash table for a
2269 guest using the PAPR paravirtualization interface. This only does
2270 anything if the kernel is configured to use the Book 3S HV style of
2271 virtualization. Otherwise the capability doesn't exist and the ioctl
2272 returns an ENOTTY error. The rest of this description assumes Book 3S
2275 There must be no vcpus running when this ioctl is called; if there
2276 are, it will do nothing and return an EBUSY error.
2278 The parameter is a pointer to a 32-bit unsigned integer variable
2279 containing the order (log base 2) of the desired size of the hash
2280 table, which must be between 18 and 46. On successful return from the
2281 ioctl, it will have been updated with the order of the hash table that
2284 If no hash table has been allocated when any vcpu is asked to run
2285 (with the KVM_RUN ioctl), the host kernel will allocate a
2286 default-sized hash table (16 MB).
2288 If this ioctl is called when a hash table has already been allocated,
2289 the kernel will clear out the existing hash table (zero all HPTEs) and
2290 return the hash table order in the parameter. (If the guest is using
2291 the virtualized real-mode area (VRMA) facility, the kernel will
2292 re-create the VMRA HPTEs on the next KVM_RUN of any vcpu.)
2294 4.77 KVM_S390_INTERRUPT
2298 Type: vm ioctl, vcpu ioctl
2299 Parameters: struct kvm_s390_interrupt (in)
2300 Returns: 0 on success, -1 on error
2302 Allows to inject an interrupt to the guest. Interrupts can be floating
2303 (vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type.
2305 Interrupt parameters are passed via kvm_s390_interrupt:
2307 struct kvm_s390_interrupt {
2313 type can be one of the following:
2315 KVM_S390_SIGP_STOP (vcpu) - sigp restart
2316 KVM_S390_PROGRAM_INT (vcpu) - program check; code in parm
2317 KVM_S390_SIGP_SET_PREFIX (vcpu) - sigp set prefix; prefix address in parm
2318 KVM_S390_RESTART (vcpu) - restart
2319 KVM_S390_INT_CLOCK_COMP (vcpu) - clock comparator interrupt
2320 KVM_S390_INT_CPU_TIMER (vcpu) - CPU timer interrupt
2321 KVM_S390_INT_VIRTIO (vm) - virtio external interrupt; external interrupt
2322 parameters in parm and parm64
2323 KVM_S390_INT_SERVICE (vm) - sclp external interrupt; sclp parameter in parm
2324 KVM_S390_INT_EMERGENCY (vcpu) - sigp emergency; source cpu in parm
2325 KVM_S390_INT_EXTERNAL_CALL (vcpu) - sigp external call; source cpu in parm
2326 KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) - compound value to indicate an
2327 I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel);
2328 I/O interruption parameters in parm (subchannel) and parm64 (intparm,
2329 interruption subclass)
2330 KVM_S390_MCHK (vm, vcpu) - machine check interrupt; cr 14 bits in parm,
2331 machine check interrupt code in parm64 (note that
2332 machine checks needing further payload are not
2333 supported by this ioctl)
2335 Note that the vcpu ioctl is asynchronous to vcpu execution.
2337 4.78 KVM_PPC_GET_HTAB_FD
2339 Capability: KVM_CAP_PPC_HTAB_FD
2340 Architectures: powerpc
2342 Parameters: Pointer to struct kvm_get_htab_fd (in)
2343 Returns: file descriptor number (>= 0) on success, -1 on error
2345 This returns a file descriptor that can be used either to read out the
2346 entries in the guest's hashed page table (HPT), or to write entries to
2347 initialize the HPT. The returned fd can only be written to if the
2348 KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and
2349 can only be read if that bit is clear. The argument struct looks like
2352 /* For KVM_PPC_GET_HTAB_FD */
2353 struct kvm_get_htab_fd {
2359 /* Values for kvm_get_htab_fd.flags */
2360 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1)
2361 #define KVM_GET_HTAB_WRITE ((__u64)0x2)
2363 The `start_index' field gives the index in the HPT of the entry at
2364 which to start reading. It is ignored when writing.
2366 Reads on the fd will initially supply information about all
2367 "interesting" HPT entries. Interesting entries are those with the
2368 bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise
2369 all entries. When the end of the HPT is reached, the read() will
2370 return. If read() is called again on the fd, it will start again from
2371 the beginning of the HPT, but will only return HPT entries that have
2372 changed since they were last read.
2374 Data read or written is structured as a header (8 bytes) followed by a
2375 series of valid HPT entries (16 bytes) each. The header indicates how
2376 many valid HPT entries there are and how many invalid entries follow
2377 the valid entries. The invalid entries are not represented explicitly
2378 in the stream. The header format is:
2380 struct kvm_get_htab_header {
2386 Writes to the fd create HPT entries starting at the index given in the
2387 header; first `n_valid' valid entries with contents from the data
2388 written, then `n_invalid' invalid entries, invalidating any previously
2389 valid entries found.
2391 4.79 KVM_CREATE_DEVICE
2393 Capability: KVM_CAP_DEVICE_CTRL
2395 Parameters: struct kvm_create_device (in/out)
2396 Returns: 0 on success, -1 on error
2398 ENODEV: The device type is unknown or unsupported
2399 EEXIST: Device already created, and this type of device may not
2400 be instantiated multiple times
2402 Other error conditions may be defined by individual device types or
2403 have their standard meanings.
2405 Creates an emulated device in the kernel. The file descriptor returned
2406 in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR.
2408 If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the
2409 device type is supported (not necessarily whether it can be created
2412 Individual devices should not define flags. Attributes should be used
2413 for specifying any behavior that is not implied by the device type
2416 struct kvm_create_device {
2417 __u32 type; /* in: KVM_DEV_TYPE_xxx */
2418 __u32 fd; /* out: device handle */
2419 __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */
2422 4.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR
2424 Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device
2425 Type: device ioctl, vm ioctl
2426 Parameters: struct kvm_device_attr
2427 Returns: 0 on success, -1 on error
2429 ENXIO: The group or attribute is unknown/unsupported for this device
2430 EPERM: The attribute cannot (currently) be accessed this way
2431 (e.g. read-only attribute, or attribute that only makes
2432 sense when the device is in a different state)
2434 Other error conditions may be defined by individual device types.
2436 Gets/sets a specified piece of device configuration and/or state. The
2437 semantics are device-specific. See individual device documentation in
2438 the "devices" directory. As with ONE_REG, the size of the data
2439 transferred is defined by the particular attribute.
2441 struct kvm_device_attr {
2442 __u32 flags; /* no flags currently defined */
2443 __u32 group; /* device-defined */
2444 __u64 attr; /* group-defined */
2445 __u64 addr; /* userspace address of attr data */
2448 4.81 KVM_HAS_DEVICE_ATTR
2450 Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device
2451 Type: device ioctl, vm ioctl
2452 Parameters: struct kvm_device_attr
2453 Returns: 0 on success, -1 on error
2455 ENXIO: The group or attribute is unknown/unsupported for this device
2457 Tests whether a device supports a particular attribute. A successful
2458 return indicates the attribute is implemented. It does not necessarily
2459 indicate that the attribute can be read or written in the device's
2460 current state. "addr" is ignored.
2462 4.82 KVM_ARM_VCPU_INIT
2465 Architectures: arm, arm64
2467 Parameters: struct kvm_vcpu_init (in)
2468 Returns: 0 on success; -1 on error
2470 Â EINVAL: Â Â Â the target is unknown, or the combination of features is invalid.
2471 Â ENOENT: Â Â Â a features bit specified is unknown.
2473 This tells KVM what type of CPU to present to the guest, and what
2474 optional features it should have. Â This will cause a reset of the cpu
2475 registers to their initial values. Â If this is not called, KVM_RUN will
2476 return ENOEXEC for that vcpu.
2478 Note that because some registers reflect machine topology, all vcpus
2479 should be created before this ioctl is invoked.
2481 Userspace can call this function multiple times for a given vcpu, including
2482 after the vcpu has been run. This will reset the vcpu to its initial
2483 state. All calls to this function after the initial call must use the same
2484 target and same set of feature flags, otherwise EINVAL will be returned.
2487 - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state.
2488 Depends on KVM_CAP_ARM_PSCI. If not set, the CPU will be powered on
2489 and execute guest code when KVM_RUN is called.
2490 - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode.
2491 Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only).
2492 - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 for the CPU.
2493 Depends on KVM_CAP_ARM_PSCI_0_2.
2496 4.83 KVM_ARM_PREFERRED_TARGET
2499 Architectures: arm, arm64
2501 Parameters: struct struct kvm_vcpu_init (out)
2502 Returns: 0 on success; -1 on error
2504 ENODEV: no preferred target available for the host
2506 This queries KVM for preferred CPU target type which can be emulated
2507 by KVM on underlying host.
2509 The ioctl returns struct kvm_vcpu_init instance containing information
2510 about preferred CPU target type and recommended features for it. The
2511 kvm_vcpu_init->features bitmap returned will have feature bits set if
2512 the preferred target recommends setting these features, but this is
2515 The information returned by this ioctl can be used to prepare an instance
2516 of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in
2517 in VCPU matching underlying host.
2520 4.84 KVM_GET_REG_LIST
2523 Architectures: arm, arm64, mips
2525 Parameters: struct kvm_reg_list (in/out)
2526 Returns: 0 on success; -1 on error
2528 Â E2BIG: Â Â Â Â the reg index list is too big to fit in the array specified by
2529 Â Â Â Â Â Â Â Â Â Â Â Â the user (the number required will be written into n).
2531 struct kvm_reg_list {
2532 __u64 n; /* number of registers in reg[] */
2536 This ioctl returns the guest registers that are supported for the
2537 KVM_GET_ONE_REG/KVM_SET_ONE_REG calls.
2540 4.85 KVM_ARM_SET_DEVICE_ADDR (deprecated)
2542 Capability: KVM_CAP_ARM_SET_DEVICE_ADDR
2543 Architectures: arm, arm64
2545 Parameters: struct kvm_arm_device_address (in)
2546 Returns: 0 on success, -1 on error
2548 ENODEV: The device id is unknown
2549 ENXIO: Device not supported on current system
2550 EEXIST: Address already set
2551 E2BIG: Address outside guest physical address space
2552 EBUSY: Address overlaps with other device range
2554 struct kvm_arm_device_addr {
2559 Specify a device address in the guest's physical address space where guests
2560 can access emulated or directly exposed devices, which the host kernel needs
2561 to know about. The id field is an architecture specific identifier for a
2564 ARM/arm64 divides the id field into two parts, a device id and an
2565 address type id specific to the individual device.
2567 Â bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 |
2568 field: | 0x00000000 | device id | addr type id |
2570 ARM/arm64 currently only require this when using the in-kernel GIC
2571 support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2
2572 as the device id. When setting the base address for the guest's
2573 mapping of the VGIC virtual CPU and distributor interface, the ioctl
2574 must be called after calling KVM_CREATE_IRQCHIP, but before calling
2575 KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the
2576 base addresses will return -EEXIST.
2578 Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API
2579 should be used instead.
2582 4.86 KVM_PPC_RTAS_DEFINE_TOKEN
2584 Capability: KVM_CAP_PPC_RTAS
2587 Parameters: struct kvm_rtas_token_args
2588 Returns: 0 on success, -1 on error
2590 Defines a token value for a RTAS (Run Time Abstraction Services)
2591 service in order to allow it to be handled in the kernel. The
2592 argument struct gives the name of the service, which must be the name
2593 of a service that has a kernel-side implementation. If the token
2594 value is non-zero, it will be associated with that service, and
2595 subsequent RTAS calls by the guest specifying that token will be
2596 handled by the kernel. If the token value is 0, then any token
2597 associated with the service will be forgotten, and subsequent RTAS
2598 calls by the guest for that service will be passed to userspace to be
2601 4.87 KVM_SET_GUEST_DEBUG
2603 Capability: KVM_CAP_SET_GUEST_DEBUG
2604 Architectures: x86, s390, ppc
2606 Parameters: struct kvm_guest_debug (in)
2607 Returns: 0 on success; -1 on error
2609 struct kvm_guest_debug {
2612 struct kvm_guest_debug_arch arch;
2615 Set up the processor specific debug registers and configure vcpu for
2616 handling guest debug events. There are two parts to the structure, the
2617 first a control bitfield indicates the type of debug events to handle
2618 when running. Common control bits are:
2620 - KVM_GUESTDBG_ENABLE: guest debugging is enabled
2621 - KVM_GUESTDBG_SINGLESTEP: the next run should single-step
2623 The top 16 bits of the control field are architecture specific control
2624 flags which can include the following:
2626 - KVM_GUESTDBG_USE_SW_BP: using software breakpoints [x86]
2627 - KVM_GUESTDBG_USE_HW_BP: using hardware breakpoints [x86, s390]
2628 - KVM_GUESTDBG_INJECT_DB: inject DB type exception [x86]
2629 - KVM_GUESTDBG_INJECT_BP: inject BP type exception [x86]
2630 - KVM_GUESTDBG_EXIT_PENDING: trigger an immediate guest exit [s390]
2632 For example KVM_GUESTDBG_USE_SW_BP indicates that software breakpoints
2633 are enabled in memory so we need to ensure breakpoint exceptions are
2634 correctly trapped and the KVM run loop exits at the breakpoint and not
2635 running off into the normal guest vector. For KVM_GUESTDBG_USE_HW_BP
2636 we need to ensure the guest vCPUs architecture specific registers are
2637 updated to the correct (supplied) values.
2639 The second part of the structure is architecture specific and
2640 typically contains a set of debug registers.
2642 When debug events exit the main run loop with the reason
2643 KVM_EXIT_DEBUG with the kvm_debug_exit_arch part of the kvm_run
2644 structure containing architecture specific debug information.
2646 4.88 KVM_GET_EMULATED_CPUID
2648 Capability: KVM_CAP_EXT_EMUL_CPUID
2651 Parameters: struct kvm_cpuid2 (in/out)
2652 Returns: 0 on success, -1 on error
2657 struct kvm_cpuid_entry2 entries[0];
2660 The member 'flags' is used for passing flags from userspace.
2662 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0)
2663 #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1)
2664 #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2)
2666 struct kvm_cpuid_entry2 {
2677 This ioctl returns x86 cpuid features which are emulated by
2678 kvm.Userspace can use the information returned by this ioctl to query
2679 which features are emulated by kvm instead of being present natively.
2681 Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2
2682 structure with the 'nent' field indicating the number of entries in
2683 the variable-size array 'entries'. If the number of entries is too low
2684 to describe the cpu capabilities, an error (E2BIG) is returned. If the
2685 number is too high, the 'nent' field is adjusted and an error (ENOMEM)
2686 is returned. If the number is just right, the 'nent' field is adjusted
2687 to the number of valid entries in the 'entries' array, which is then
2690 The entries returned are the set CPUID bits of the respective features
2691 which kvm emulates, as returned by the CPUID instruction, with unknown
2692 or unsupported feature bits cleared.
2694 Features like x2apic, for example, may not be present in the host cpu
2695 but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be
2696 emulated efficiently and thus not included here.
2698 The fields in each entry are defined as follows:
2700 function: the eax value used to obtain the entry
2701 index: the ecx value used to obtain the entry (for entries that are
2703 flags: an OR of zero or more of the following:
2704 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
2705 if the index field is valid
2706 KVM_CPUID_FLAG_STATEFUL_FUNC:
2707 if cpuid for this function returns different values for successive
2708 invocations; there will be several entries with the same function,
2709 all with this flag set
2710 KVM_CPUID_FLAG_STATE_READ_NEXT:
2711 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
2712 the first entry to be read by a cpu
2713 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
2714 this function/index combination
2716 5. The kvm_run structure
2717 ------------------------
2719 Application code obtains a pointer to the kvm_run structure by
2720 mmap()ing a vcpu fd. From that point, application code can control
2721 execution by changing fields in kvm_run prior to calling the KVM_RUN
2722 ioctl, and obtain information about the reason KVM_RUN returned by
2723 looking up structure members.
2727 __u8 request_interrupt_window;
2729 Request that KVM_RUN return when it becomes possible to inject external
2730 interrupts into the guest. Useful in conjunction with KVM_INTERRUPT.
2737 When KVM_RUN has returned successfully (return value 0), this informs
2738 application code why KVM_RUN has returned. Allowable values for this
2739 field are detailed below.
2741 __u8 ready_for_interrupt_injection;
2743 If request_interrupt_window has been specified, this field indicates
2744 an interrupt can be injected now with KVM_INTERRUPT.
2748 The value of the current interrupt flag. Only valid if in-kernel
2749 local APIC is not used.
2753 /* in (pre_kvm_run), out (post_kvm_run) */
2756 The value of the cr8 register. Only valid if in-kernel local APIC is
2757 not used. Both input and output.
2761 The value of the APIC BASE msr. Only valid if in-kernel local
2762 APIC is not used. Both input and output.
2765 /* KVM_EXIT_UNKNOWN */
2767 __u64 hardware_exit_reason;
2770 If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown
2771 reasons. Further architecture-specific information is available in
2772 hardware_exit_reason.
2774 /* KVM_EXIT_FAIL_ENTRY */
2776 __u64 hardware_entry_failure_reason;
2779 If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due
2780 to unknown reasons. Further architecture-specific information is
2781 available in hardware_entry_failure_reason.
2783 /* KVM_EXIT_EXCEPTION */
2793 #define KVM_EXIT_IO_IN 0
2794 #define KVM_EXIT_IO_OUT 1
2796 __u8 size; /* bytes */
2799 __u64 data_offset; /* relative to kvm_run start */
2802 If exit_reason is KVM_EXIT_IO, then the vcpu has
2803 executed a port I/O instruction which could not be satisfied by kvm.
2804 data_offset describes where the data is located (KVM_EXIT_IO_OUT) or
2805 where kvm expects application code to place the data for the next
2806 KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array.
2809 struct kvm_debug_exit_arch arch;
2822 If exit_reason is KVM_EXIT_MMIO, then the vcpu has
2823 executed a memory-mapped I/O instruction which could not be satisfied
2824 by kvm. The 'data' member contains the written data if 'is_write' is
2825 true, and should be filled by application code otherwise.
2827 The 'data' member contains, in its first 'len' bytes, the value as it would
2828 appear if the VCPU performed a load or store of the appropriate width directly
2831 NOTE: For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_PAPR and
2832 KVM_EXIT_EPR the corresponding
2833 operations are complete (and guest state is consistent) only after userspace
2834 has re-entered the kernel with KVM_RUN. The kernel side will first finish
2835 incomplete operations and then check for pending signals. Userspace
2836 can re-enter the guest with an unmasked signal pending to complete
2839 /* KVM_EXIT_HYPERCALL */
2848 Unused. This was once used for 'hypercall to userspace'. To implement
2849 such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390).
2850 Note KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO.
2852 /* KVM_EXIT_TPR_ACCESS */
2859 To be documented (KVM_TPR_ACCESS_REPORTING).
2861 /* KVM_EXIT_S390_SIEIC */
2864 __u64 mask; /* psw upper half */
2865 __u64 addr; /* psw lower half */
2872 /* KVM_EXIT_S390_RESET */
2873 #define KVM_S390_RESET_POR 1
2874 #define KVM_S390_RESET_CLEAR 2
2875 #define KVM_S390_RESET_SUBSYSTEM 4
2876 #define KVM_S390_RESET_CPU_INIT 8
2877 #define KVM_S390_RESET_IPL 16
2878 __u64 s390_reset_flags;
2882 /* KVM_EXIT_S390_UCONTROL */
2884 __u64 trans_exc_code;
2888 s390 specific. A page fault has occurred for a user controlled virtual
2889 machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be
2890 resolved by the kernel.
2891 The program code and the translation exception code that were placed
2892 in the cpu's lowcore are presented here as defined by the z Architecture
2893 Principles of Operation Book in the Chapter for Dynamic Address Translation
2903 Deprecated - was used for 440 KVM.
2910 MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch
2911 hypercalls and exit with this exit struct that contains all the guest gprs.
2913 If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall.
2914 Userspace can now handle the hypercall and when it's done modify the gprs as
2915 necessary. Upon guest entry all guest GPRs will then be replaced by the values
2918 /* KVM_EXIT_PAPR_HCALL */
2925 This is used on 64-bit PowerPC when emulating a pSeries partition,
2926 e.g. with the 'pseries' machine type in qemu. It occurs when the
2927 guest does a hypercall using the 'sc 1' instruction. The 'nr' field
2928 contains the hypercall number (from the guest R3), and 'args' contains
2929 the arguments (from the guest R4 - R12). Userspace should put the
2930 return code in 'ret' and any extra returned values in args[].
2931 The possible hypercalls are defined in the Power Architecture Platform
2932 Requirements (PAPR) document available from www.power.org (free
2933 developer registration required to access it).
2935 /* KVM_EXIT_S390_TSCH */
2937 __u16 subchannel_id;
2938 __u16 subchannel_nr;
2945 s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled
2946 and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O
2947 interrupt for the target subchannel has been dequeued and subchannel_id,
2948 subchannel_nr, io_int_parm and io_int_word contain the parameters for that
2949 interrupt. ipb is needed for instruction parameter decoding.
2956 On FSL BookE PowerPC chips, the interrupt controller has a fast patch
2957 interrupt acknowledge path to the core. When the core successfully
2958 delivers an interrupt, it automatically populates the EPR register with
2959 the interrupt vector number and acknowledges the interrupt inside
2960 the interrupt controller.
2962 In case the interrupt controller lives in user space, we need to do
2963 the interrupt acknowledge cycle through it to fetch the next to be
2964 delivered interrupt vector using this exit.
2966 It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an
2967 external interrupt has just been delivered into the guest. User space
2968 should put the acknowledged interrupt vector into the 'epr' field.
2970 /* KVM_EXIT_SYSTEM_EVENT */
2972 #define KVM_SYSTEM_EVENT_SHUTDOWN 1
2973 #define KVM_SYSTEM_EVENT_RESET 2
2978 If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered
2979 a system-level event using some architecture specific mechanism (hypercall
2980 or some special instruction). In case of ARM/ARM64, this is triggered using
2981 HVC instruction based PSCI call from the vcpu. The 'type' field describes
2982 the system-level event type. The 'flags' field describes architecture
2983 specific flags for the system-level event.
2985 Valid values for 'type' are:
2986 KVM_SYSTEM_EVENT_SHUTDOWN -- the guest has requested a shutdown of the
2987 VM. Userspace is not obliged to honour this, and if it does honour
2988 this does not need to destroy the VM synchronously (ie it may call
2989 KVM_RUN again before shutdown finally occurs).
2990 KVM_SYSTEM_EVENT_RESET -- the guest has requested a reset of the VM.
2991 As with SHUTDOWN, userspace can choose to ignore the request, or
2992 to schedule the reset to occur in the future and may call KVM_RUN again.
2994 /* Fix the size of the union. */
2999 * shared registers between kvm and userspace.
3000 * kvm_valid_regs specifies the register classes set by the host
3001 * kvm_dirty_regs specified the register classes dirtied by userspace
3002 * struct kvm_sync_regs is architecture specific, as well as the
3003 * bits for kvm_valid_regs and kvm_dirty_regs
3005 __u64 kvm_valid_regs;
3006 __u64 kvm_dirty_regs;
3008 struct kvm_sync_regs regs;
3012 If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access
3013 certain guest registers without having to call SET/GET_*REGS. Thus we can
3014 avoid some system call overhead if userspace has to handle the exit.
3015 Userspace can query the validity of the structure by checking
3016 kvm_valid_regs for specific bits. These bits are architecture specific
3017 and usually define the validity of a groups of registers. (e.g. one bit
3018 for general purpose registers)
3020 Please note that the kernel is allowed to use the kvm_run structure as the
3021 primary storage for certain register types. Therefore, the kernel may use the
3022 values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set.
3028 6. Capabilities that can be enabled on vCPUs
3029 --------------------------------------------
3031 There are certain capabilities that change the behavior of the virtual CPU or
3032 the virtual machine when enabled. To enable them, please see section 4.37.
3033 Below you can find a list of capabilities and what their effect on the vCPU or
3034 the virtual machine is when enabling them.
3036 The following information is provided along with the description:
3038 Architectures: which instruction set architectures provide this ioctl.
3039 x86 includes both i386 and x86_64.
3041 Target: whether this is a per-vcpu or per-vm capability.
3043 Parameters: what parameters are accepted by the capability.
3045 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
3046 are not detailed, but errors with specific meanings are.
3054 Returns: 0 on success; -1 on error
3056 This capability enables interception of OSI hypercalls that otherwise would
3057 be treated as normal system calls to be injected into the guest. OSI hypercalls
3058 were invented by Mac-on-Linux to have a standardized communication mechanism
3059 between the guest and the host.
3061 When this capability is enabled, KVM_EXIT_OSI can occur.
3064 6.2 KVM_CAP_PPC_PAPR
3069 Returns: 0 on success; -1 on error
3071 This capability enables interception of PAPR hypercalls. PAPR hypercalls are
3072 done using the hypercall instruction "sc 1".
3074 It also sets the guest privilege level to "supervisor" mode. Usually the guest
3075 runs in "hypervisor" privilege mode with a few missing features.
3077 In addition to the above, it changes the semantics of SDR1. In this mode, the
3078 HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the
3079 HTAB invisible to the guest.
3081 When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur.
3088 Parameters: args[0] is the address of a struct kvm_config_tlb
3089 Returns: 0 on success; -1 on error
3091 struct kvm_config_tlb {
3098 Configures the virtual CPU's TLB array, establishing a shared memory area
3099 between userspace and KVM. The "params" and "array" fields are userspace
3100 addresses of mmu-type-specific data structures. The "array_len" field is an
3101 safety mechanism, and should be set to the size in bytes of the memory that
3102 userspace has reserved for the array. It must be at least the size dictated
3103 by "mmu_type" and "params".
3105 While KVM_RUN is active, the shared region is under control of KVM. Its
3106 contents are undefined, and any modification by userspace results in
3107 boundedly undefined behavior.
3109 On return from KVM_RUN, the shared region will reflect the current state of
3110 the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB
3111 to tell KVM which entries have been changed, prior to calling KVM_RUN again
3114 For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV:
3115 - The "params" field is of type "struct kvm_book3e_206_tlb_params".
3116 - The "array" field points to an array of type "struct
3117 kvm_book3e_206_tlb_entry".
3118 - The array consists of all entries in the first TLB, followed by all
3119 entries in the second TLB.
3120 - Within a TLB, entries are ordered first by increasing set number. Within a
3121 set, entries are ordered by way (increasing ESEL).
3122 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1)
3123 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value.
3124 - The tsize field of mas1 shall be set to 4K on TLB0, even though the
3125 hardware ignores this value for TLB0.
3127 6.4 KVM_CAP_S390_CSS_SUPPORT
3132 Returns: 0 on success; -1 on error
3134 This capability enables support for handling of channel I/O instructions.
3136 TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are
3137 handled in-kernel, while the other I/O instructions are passed to userspace.
3139 When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST
3140 SUBCHANNEL intercepts.
3142 Note that even though this capability is enabled per-vcpu, the complete
3143 virtual machine is affected.
3149 Parameters: args[0] defines whether the proxy facility is active
3150 Returns: 0 on success; -1 on error
3152 This capability enables or disables the delivery of interrupts through the
3153 external proxy facility.
3155 When enabled (args[0] != 0), every time the guest gets an external interrupt
3156 delivered, it automatically exits into user space with a KVM_EXIT_EPR exit
3157 to receive the topmost interrupt vector.
3159 When disabled (args[0] == 0), behavior is as if this facility is unsupported.
3161 When this capability is enabled, KVM_EXIT_EPR can occur.
3163 6.6 KVM_CAP_IRQ_MPIC
3166 Parameters: args[0] is the MPIC device fd
3167 args[1] is the MPIC CPU number for this vcpu
3169 This capability connects the vcpu to an in-kernel MPIC device.
3171 6.7 KVM_CAP_IRQ_XICS
3175 Parameters: args[0] is the XICS device fd
3176 args[1] is the XICS CPU number (server ID) for this vcpu
3178 This capability connects the vcpu to an in-kernel XICS device.
3180 6.8 KVM_CAP_S390_IRQCHIP
3186 This capability enables the in-kernel irqchip for s390. Please refer to
3187 "4.24 KVM_CREATE_IRQCHIP" for details.
3189 7. Capabilities that can be enabled on VMs
3190 ------------------------------------------
3192 There are certain capabilities that change the behavior of the virtual
3193 machine when enabled. To enable them, please see section 4.37. Below
3194 you can find a list of capabilities and what their effect on the VM
3195 is when enabling them.
3197 The following information is provided along with the description:
3199 Architectures: which instruction set architectures provide this ioctl.
3200 x86 includes both i386 and x86_64.
3202 Parameters: what parameters are accepted by the capability.
3204 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
3205 are not detailed, but errors with specific meanings are.
3208 7.1 KVM_CAP_PPC_ENABLE_HCALL
3211 Parameters: args[0] is the sPAPR hcall number
3212 args[1] is 0 to disable, 1 to enable in-kernel handling
3214 This capability controls whether individual sPAPR hypercalls (hcalls)
3215 get handled by the kernel or not. Enabling or disabling in-kernel
3216 handling of an hcall is effective across the VM. On creation, an
3217 initial set of hcalls are enabled for in-kernel handling, which
3218 consists of those hcalls for which in-kernel handlers were implemented
3219 before this capability was implemented. If disabled, the kernel will
3220 not to attempt to handle the hcall, but will always exit to userspace
3221 to handle it. Note that it may not make sense to enable some and
3222 disable others of a group of related hcalls, but KVM does not prevent
3223 userspace from doing that.
3225 If the hcall number specified is not one that has an in-kernel
3226 implementation, the KVM_ENABLE_CAP ioctl will fail with an EINVAL