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 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), or a KVM_CAP_xyz constant, which
72 means availability needs to be checked with KVM_CHECK_EXTENSION
75 Architectures: which instruction set architectures provide this ioctl.
76 x86 includes both i386 and x86_64.
78 Type: system, vm, or vcpu.
80 Parameters: what parameters are accepted by the ioctl.
82 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
83 are not detailed, but errors with specific meanings are.
86 4.1 KVM_GET_API_VERSION
92 Returns: the constant KVM_API_VERSION (=12)
94 This identifies the API version as the stable kvm API. It is not
95 expected that this number will change. However, Linux 2.6.20 and
96 2.6.21 report earlier versions; these are not documented and not
97 supported. Applications should refuse to run if KVM_GET_API_VERSION
98 returns a value other than 12. If this check passes, all ioctls
99 described as 'basic' will be available.
107 Parameters: machine type identifier (KVM_VM_*)
108 Returns: a VM fd that can be used to control the new virtual machine.
110 The new VM has no virtual cpus and no memory. An mmap() of a VM fd
111 will access the virtual machine's physical address space; offset zero
112 corresponds to guest physical address zero. Use of mmap() on a VM fd
113 is discouraged if userspace memory allocation (KVM_CAP_USER_MEMORY) is
115 You most certainly want to use 0 as machine type.
117 In order to create user controlled virtual machines on S390, check
118 KVM_CAP_S390_UCONTROL and use the flag KVM_VM_S390_UCONTROL as
119 privileged user (CAP_SYS_ADMIN).
122 4.3 KVM_GET_MSR_INDEX_LIST
127 Parameters: struct kvm_msr_list (in/out)
128 Returns: 0 on success; -1 on error
130 E2BIG: the msr index list is to be to fit in the array specified by
133 struct kvm_msr_list {
134 __u32 nmsrs; /* number of msrs in entries */
138 This ioctl returns the guest msrs that are supported. The list varies
139 by kvm version and host processor, but does not change otherwise. The
140 user fills in the size of the indices array in nmsrs, and in return
141 kvm adjusts nmsrs to reflect the actual number of msrs and fills in
142 the indices array with their numbers.
144 Note: if kvm indicates supports MCE (KVM_CAP_MCE), then the MCE bank MSRs are
145 not returned in the MSR list, as different vcpus can have a different number
146 of banks, as set via the KVM_X86_SETUP_MCE ioctl.
149 4.4 KVM_CHECK_EXTENSION
154 Parameters: extension identifier (KVM_CAP_*)
155 Returns: 0 if unsupported; 1 (or some other positive integer) if supported
157 The API allows the application to query about extensions to the core
158 kvm API. Userspace passes an extension identifier (an integer) and
159 receives an integer that describes the extension availability.
160 Generally 0 means no and 1 means yes, but some extensions may report
161 additional information in the integer return value.
164 4.5 KVM_GET_VCPU_MMAP_SIZE
170 Returns: size of vcpu mmap area, in bytes
172 The KVM_RUN ioctl (cf.) communicates with userspace via a shared
173 memory region. This ioctl returns the size of that region. See the
174 KVM_RUN documentation for details.
177 4.6 KVM_SET_MEMORY_REGION
182 Parameters: struct kvm_memory_region (in)
183 Returns: 0 on success, -1 on error
185 This ioctl is obsolete and has been removed.
193 Parameters: vcpu id (apic id on x86)
194 Returns: vcpu fd on success, -1 on error
196 This API adds a vcpu to a virtual machine. The vcpu id is a small integer
197 in the range [0, max_vcpus).
199 The recommended max_vcpus value can be retrieved using the KVM_CAP_NR_VCPUS of
200 the KVM_CHECK_EXTENSION ioctl() at run-time.
201 The maximum possible value for max_vcpus can be retrieved using the
202 KVM_CAP_MAX_VCPUS of the KVM_CHECK_EXTENSION ioctl() at run-time.
204 If the KVM_CAP_NR_VCPUS does not exist, you should assume that max_vcpus is 4
206 If the KVM_CAP_MAX_VCPUS does not exist, you should assume that max_vcpus is
207 same as the value returned from KVM_CAP_NR_VCPUS.
209 On powerpc using book3s_hv mode, the vcpus are mapped onto virtual
210 threads in one or more virtual CPU cores. (This is because the
211 hardware requires all the hardware threads in a CPU core to be in the
212 same partition.) The KVM_CAP_PPC_SMT capability indicates the number
213 of vcpus per virtual core (vcore). The vcore id is obtained by
214 dividing the vcpu id by the number of vcpus per vcore. The vcpus in a
215 given vcore will always be in the same physical core as each other
216 (though that might be a different physical core from time to time).
217 Userspace can control the threading (SMT) mode of the guest by its
218 allocation of vcpu ids. For example, if userspace wants
219 single-threaded guest vcpus, it should make all vcpu ids be a multiple
220 of the number of vcpus per vcore.
222 On powerpc using book3s_hv mode, the vcpus are mapped onto virtual
223 threads in one or more virtual CPU cores. (This is because the
224 hardware requires all the hardware threads in a CPU core to be in the
225 same partition.) The KVM_CAP_PPC_SMT capability indicates the number
226 of vcpus per virtual core (vcore). The vcore id is obtained by
227 dividing the vcpu id by the number of vcpus per vcore. The vcpus in a
228 given vcore will always be in the same physical core as each other
229 (though that might be a different physical core from time to time).
230 Userspace can control the threading (SMT) mode of the guest by its
231 allocation of vcpu ids. For example, if userspace wants
232 single-threaded guest vcpus, it should make all vcpu ids be a multiple
233 of the number of vcpus per vcore.
235 For virtual cpus that have been created with S390 user controlled virtual
236 machines, the resulting vcpu fd can be memory mapped at page offset
237 KVM_S390_SIE_PAGE_OFFSET in order to obtain a memory map of the virtual
238 cpu's hardware control block.
241 4.8 KVM_GET_DIRTY_LOG (vm ioctl)
246 Parameters: struct kvm_dirty_log (in/out)
247 Returns: 0 on success, -1 on error
249 /* for KVM_GET_DIRTY_LOG */
250 struct kvm_dirty_log {
254 void __user *dirty_bitmap; /* one bit per page */
259 Given a memory slot, return a bitmap containing any pages dirtied
260 since the last call to this ioctl. Bit 0 is the first page in the
261 memory slot. Ensure the entire structure is cleared to avoid padding
265 4.9 KVM_SET_MEMORY_ALIAS
270 Parameters: struct kvm_memory_alias (in)
271 Returns: 0 (success), -1 (error)
273 This ioctl is obsolete and has been removed.
282 Returns: 0 on success, -1 on error
284 EINTR: an unmasked signal is pending
286 This ioctl is used to run a guest virtual cpu. While there are no
287 explicit parameters, there is an implicit parameter block that can be
288 obtained by mmap()ing the vcpu fd at offset 0, with the size given by
289 KVM_GET_VCPU_MMAP_SIZE. The parameter block is formatted as a 'struct
290 kvm_run' (see below).
296 Architectures: all except ARM
298 Parameters: struct kvm_regs (out)
299 Returns: 0 on success, -1 on error
301 Reads the general purpose registers from the vcpu.
305 /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */
306 __u64 rax, rbx, rcx, rdx;
307 __u64 rsi, rdi, rsp, rbp;
308 __u64 r8, r9, r10, r11;
309 __u64 r12, r13, r14, r15;
317 Architectures: all except ARM
319 Parameters: struct kvm_regs (in)
320 Returns: 0 on success, -1 on error
322 Writes the general purpose registers into the vcpu.
324 See KVM_GET_REGS for the data structure.
330 Architectures: x86, ppc
332 Parameters: struct kvm_sregs (out)
333 Returns: 0 on success, -1 on error
335 Reads special registers from the vcpu.
339 struct kvm_segment cs, ds, es, fs, gs, ss;
340 struct kvm_segment tr, ldt;
341 struct kvm_dtable gdt, idt;
342 __u64 cr0, cr2, cr3, cr4, cr8;
345 __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64];
348 /* ppc -- see arch/powerpc/include/asm/kvm.h */
350 interrupt_bitmap is a bitmap of pending external interrupts. At most
351 one bit may be set. This interrupt has been acknowledged by the APIC
352 but not yet injected into the cpu core.
358 Architectures: x86, ppc
360 Parameters: struct kvm_sregs (in)
361 Returns: 0 on success, -1 on error
363 Writes special registers into the vcpu. See KVM_GET_SREGS for the
372 Parameters: struct kvm_translation (in/out)
373 Returns: 0 on success, -1 on error
375 Translates a virtual address according to the vcpu's current address
378 struct kvm_translation {
380 __u64 linear_address;
383 __u64 physical_address;
394 Architectures: x86, ppc
396 Parameters: struct kvm_interrupt (in)
397 Returns: 0 on success, -1 on error
399 Queues a hardware interrupt vector to be injected. This is only
400 useful if in-kernel local APIC or equivalent is not used.
402 /* for KVM_INTERRUPT */
403 struct kvm_interrupt {
410 Note 'irq' is an interrupt vector, not an interrupt pin or line.
414 Queues an external interrupt to be injected. This ioctl is overleaded
415 with 3 different irq values:
419 This injects an edge type external interrupt into the guest once it's ready
420 to receive interrupts. When injected, the interrupt is done.
422 b) KVM_INTERRUPT_UNSET
424 This unsets any pending interrupt.
426 Only available with KVM_CAP_PPC_UNSET_IRQ.
428 c) KVM_INTERRUPT_SET_LEVEL
430 This injects a level type external interrupt into the guest context. The
431 interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET
434 Only available with KVM_CAP_PPC_IRQ_LEVEL.
436 Note that any value for 'irq' other than the ones stated above is invalid
437 and incurs unexpected behavior.
448 Support for this has been removed. Use KVM_SET_GUEST_DEBUG instead.
456 Parameters: struct kvm_msrs (in/out)
457 Returns: 0 on success, -1 on error
459 Reads model-specific registers from the vcpu. Supported msr indices can
460 be obtained using KVM_GET_MSR_INDEX_LIST.
463 __u32 nmsrs; /* number of msrs in entries */
466 struct kvm_msr_entry entries[0];
469 struct kvm_msr_entry {
475 Application code should set the 'nmsrs' member (which indicates the
476 size of the entries array) and the 'index' member of each array entry.
477 kvm will fill in the 'data' member.
485 Parameters: struct kvm_msrs (in)
486 Returns: 0 on success, -1 on error
488 Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the
491 Application code should set the 'nmsrs' member (which indicates the
492 size of the entries array), and the 'index' and 'data' members of each
501 Parameters: struct kvm_cpuid (in)
502 Returns: 0 on success, -1 on error
504 Defines the vcpu responses to the cpuid instruction. Applications
505 should use the KVM_SET_CPUID2 ioctl if available.
508 struct kvm_cpuid_entry {
517 /* for KVM_SET_CPUID */
521 struct kvm_cpuid_entry entries[0];
525 4.21 KVM_SET_SIGNAL_MASK
530 Parameters: struct kvm_signal_mask (in)
531 Returns: 0 on success, -1 on error
533 Defines which signals are blocked during execution of KVM_RUN. This
534 signal mask temporarily overrides the threads signal mask. Any
535 unblocked signal received (except SIGKILL and SIGSTOP, which retain
536 their traditional behaviour) will cause KVM_RUN to return with -EINTR.
538 Note the signal will only be delivered if not blocked by the original
541 /* for KVM_SET_SIGNAL_MASK */
542 struct kvm_signal_mask {
553 Parameters: struct kvm_fpu (out)
554 Returns: 0 on success, -1 on error
556 Reads the floating point state from the vcpu.
558 /* for KVM_GET_FPU and KVM_SET_FPU */
563 __u8 ftwx; /* in fxsave format */
579 Parameters: struct kvm_fpu (in)
580 Returns: 0 on success, -1 on error
582 Writes the floating point state to the vcpu.
584 /* for KVM_GET_FPU and KVM_SET_FPU */
589 __u8 ftwx; /* in fxsave format */
600 4.24 KVM_CREATE_IRQCHIP
602 Capability: KVM_CAP_IRQCHIP
603 Architectures: x86, ia64, ARM
606 Returns: 0 on success, -1 on error
608 Creates an interrupt controller model in the kernel. On x86, creates a virtual
609 ioapic, a virtual PIC (two PICs, nested), and sets up future vcpus to have a
610 local APIC. IRQ routing for GSIs 0-15 is set to both PIC and IOAPIC; GSI 16-23
611 only go to the IOAPIC. On ia64, a IOSAPIC is created. On ARM, a GIC is
617 Capability: KVM_CAP_IRQCHIP
618 Architectures: x86, ia64
620 Parameters: struct kvm_irq_level
621 Returns: 0 on success, -1 on error
623 Sets the level of a GSI input to the interrupt controller model in the kernel.
624 Requires that an interrupt controller model has been previously created with
625 KVM_CREATE_IRQCHIP. Note that edge-triggered interrupts require the level
626 to be set to 1 and then back to 0.
628 struct kvm_irq_level {
631 __s32 status; /* not used for KVM_IRQ_LEVEL */
633 __u32 level; /* 0 or 1 */
639 Capability: KVM_CAP_IRQCHIP
640 Architectures: x86, ia64
642 Parameters: struct kvm_irqchip (in/out)
643 Returns: 0 on success, -1 on error
645 Reads the state of a kernel interrupt controller created with
646 KVM_CREATE_IRQCHIP into a buffer provided by the caller.
649 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
652 char dummy[512]; /* reserving space */
653 struct kvm_pic_state pic;
654 struct kvm_ioapic_state ioapic;
661 Capability: KVM_CAP_IRQCHIP
662 Architectures: x86, ia64
664 Parameters: struct kvm_irqchip (in)
665 Returns: 0 on success, -1 on error
667 Sets the state of a kernel interrupt controller created with
668 KVM_CREATE_IRQCHIP from a buffer provided by the caller.
671 __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */
674 char dummy[512]; /* reserving space */
675 struct kvm_pic_state pic;
676 struct kvm_ioapic_state ioapic;
681 4.28 KVM_XEN_HVM_CONFIG
683 Capability: KVM_CAP_XEN_HVM
686 Parameters: struct kvm_xen_hvm_config (in)
687 Returns: 0 on success, -1 on error
689 Sets the MSR that the Xen HVM guest uses to initialize its hypercall
690 page, and provides the starting address and size of the hypercall
691 blobs in userspace. When the guest writes the MSR, kvm copies one
692 page of a blob (32- or 64-bit, depending on the vcpu mode) to guest
695 struct kvm_xen_hvm_config {
708 Capability: KVM_CAP_ADJUST_CLOCK
711 Parameters: struct kvm_clock_data (out)
712 Returns: 0 on success, -1 on error
714 Gets the current timestamp of kvmclock as seen by the current guest. In
715 conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios
718 struct kvm_clock_data {
719 __u64 clock; /* kvmclock current value */
727 Capability: KVM_CAP_ADJUST_CLOCK
730 Parameters: struct kvm_clock_data (in)
731 Returns: 0 on success, -1 on error
733 Sets the current timestamp of kvmclock to the value specified in its parameter.
734 In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios
737 struct kvm_clock_data {
738 __u64 clock; /* kvmclock current value */
744 4.31 KVM_GET_VCPU_EVENTS
746 Capability: KVM_CAP_VCPU_EVENTS
747 Extended by: KVM_CAP_INTR_SHADOW
750 Parameters: struct kvm_vcpu_event (out)
751 Returns: 0 on success, -1 on error
753 Gets currently pending exceptions, interrupts, and NMIs as well as related
756 struct kvm_vcpu_events {
780 KVM_VCPUEVENT_VALID_SHADOW may be set in the flags field to signal that
781 interrupt.shadow contains a valid state. Otherwise, this field is undefined.
784 4.32 KVM_SET_VCPU_EVENTS
786 Capability: KVM_CAP_VCPU_EVENTS
787 Extended by: KVM_CAP_INTR_SHADOW
790 Parameters: struct kvm_vcpu_event (in)
791 Returns: 0 on success, -1 on error
793 Set pending exceptions, interrupts, and NMIs as well as related states of the
796 See KVM_GET_VCPU_EVENTS for the data structure.
798 Fields that may be modified asynchronously by running VCPUs can be excluded
799 from the update. These fields are nmi.pending and sipi_vector. Keep the
800 corresponding bits in the flags field cleared to suppress overwriting the
801 current in-kernel state. The bits are:
803 KVM_VCPUEVENT_VALID_NMI_PENDING - transfer nmi.pending to the kernel
804 KVM_VCPUEVENT_VALID_SIPI_VECTOR - transfer sipi_vector
806 If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in
807 the flags field to signal that interrupt.shadow contains a valid state and
808 shall be written into the VCPU.
811 4.33 KVM_GET_DEBUGREGS
813 Capability: KVM_CAP_DEBUGREGS
816 Parameters: struct kvm_debugregs (out)
817 Returns: 0 on success, -1 on error
819 Reads debug registers from the vcpu.
821 struct kvm_debugregs {
830 4.34 KVM_SET_DEBUGREGS
832 Capability: KVM_CAP_DEBUGREGS
835 Parameters: struct kvm_debugregs (in)
836 Returns: 0 on success, -1 on error
838 Writes debug registers into the vcpu.
840 See KVM_GET_DEBUGREGS for the data structure. The flags field is unused
841 yet and must be cleared on entry.
844 4.35 KVM_SET_USER_MEMORY_REGION
846 Capability: KVM_CAP_USER_MEM
849 Parameters: struct kvm_userspace_memory_region (in)
850 Returns: 0 on success, -1 on error
852 struct kvm_userspace_memory_region {
855 __u64 guest_phys_addr;
856 __u64 memory_size; /* bytes */
857 __u64 userspace_addr; /* start of the userspace allocated memory */
860 /* for kvm_memory_region::flags */
861 #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0)
862 #define KVM_MEM_READONLY (1UL << 1)
864 This ioctl allows the user to create or modify a guest physical memory
865 slot. When changing an existing slot, it may be moved in the guest
866 physical memory space, or its flags may be modified. It may not be
867 resized. Slots may not overlap in guest physical address space.
869 Memory for the region is taken starting at the address denoted by the
870 field userspace_addr, which must point at user addressable memory for
871 the entire memory slot size. Any object may back this memory, including
872 anonymous memory, ordinary files, and hugetlbfs.
874 It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr
875 be identical. This allows large pages in the guest to be backed by large
878 The flags field supports two flag, KVM_MEM_LOG_DIRTY_PAGES, which instructs
879 kvm to keep track of writes to memory within the slot. See KVM_GET_DIRTY_LOG
880 ioctl. The KVM_CAP_READONLY_MEM capability indicates the availability of the
881 KVM_MEM_READONLY flag. When this flag is set for a memory region, KVM only
882 allows read accesses. Writes will be posted to userspace as KVM_EXIT_MMIO
885 When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of
886 the memory region are automatically reflected into the guest. For example, an
887 mmap() that affects the region will be made visible immediately. Another
888 example is madvise(MADV_DROP).
890 It is recommended to use this API instead of the KVM_SET_MEMORY_REGION ioctl.
891 The KVM_SET_MEMORY_REGION does not allow fine grained control over memory
892 allocation and is deprecated.
895 4.36 KVM_SET_TSS_ADDR
897 Capability: KVM_CAP_SET_TSS_ADDR
900 Parameters: unsigned long tss_address (in)
901 Returns: 0 on success, -1 on error
903 This ioctl defines the physical address of a three-page region in the guest
904 physical address space. The region must be within the first 4GB of the
905 guest physical address space and must not conflict with any memory slot
906 or any mmio address. The guest may malfunction if it accesses this memory
909 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
910 because of a quirk in the virtualization implementation (see the internals
911 documentation when it pops into existence).
916 Capability: KVM_CAP_ENABLE_CAP
919 Parameters: struct kvm_enable_cap (in)
920 Returns: 0 on success; -1 on error
922 +Not all extensions are enabled by default. Using this ioctl the application
923 can enable an extension, making it available to the guest.
925 On systems that do not support this ioctl, it always fails. On systems that
926 do support it, it only works for extensions that are supported for enablement.
928 To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should
931 struct kvm_enable_cap {
935 The capability that is supposed to get enabled.
939 A bitfield indicating future enhancements. Has to be 0 for now.
943 Arguments for enabling a feature. If a feature needs initial values to
944 function properly, this is the place to put them.
950 4.38 KVM_GET_MP_STATE
952 Capability: KVM_CAP_MP_STATE
953 Architectures: x86, ia64
955 Parameters: struct kvm_mp_state (out)
956 Returns: 0 on success; -1 on error
958 struct kvm_mp_state {
962 Returns the vcpu's current "multiprocessing state" (though also valid on
963 uniprocessor guests).
967 - KVM_MP_STATE_RUNNABLE: the vcpu is currently running
968 - KVM_MP_STATE_UNINITIALIZED: the vcpu is an application processor (AP)
969 which has not yet received an INIT signal
970 - KVM_MP_STATE_INIT_RECEIVED: the vcpu has received an INIT signal, and is
972 - KVM_MP_STATE_HALTED: the vcpu has executed a HLT instruction and
973 is waiting for an interrupt
974 - KVM_MP_STATE_SIPI_RECEIVED: the vcpu has just received a SIPI (vector
975 accessible via KVM_GET_VCPU_EVENTS)
977 This ioctl is only useful after KVM_CREATE_IRQCHIP. Without an in-kernel
978 irqchip, the multiprocessing state must be maintained by userspace.
981 4.39 KVM_SET_MP_STATE
983 Capability: KVM_CAP_MP_STATE
984 Architectures: x86, ia64
986 Parameters: struct kvm_mp_state (in)
987 Returns: 0 on success; -1 on error
989 Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for
992 This ioctl is only useful after KVM_CREATE_IRQCHIP. Without an in-kernel
993 irqchip, the multiprocessing state must be maintained by userspace.
996 4.40 KVM_SET_IDENTITY_MAP_ADDR
998 Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR
1001 Parameters: unsigned long identity (in)
1002 Returns: 0 on success, -1 on error
1004 This ioctl defines the physical address of a one-page region in the guest
1005 physical address space. The region must be within the first 4GB of the
1006 guest physical address space and must not conflict with any memory slot
1007 or any mmio address. The guest may malfunction if it accesses this memory
1010 This ioctl is required on Intel-based hosts. This is needed on Intel hardware
1011 because of a quirk in the virtualization implementation (see the internals
1012 documentation when it pops into existence).
1015 4.41 KVM_SET_BOOT_CPU_ID
1017 Capability: KVM_CAP_SET_BOOT_CPU_ID
1018 Architectures: x86, ia64
1020 Parameters: unsigned long vcpu_id
1021 Returns: 0 on success, -1 on error
1023 Define which vcpu is the Bootstrap Processor (BSP). Values are the same
1024 as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default
1030 Capability: KVM_CAP_XSAVE
1033 Parameters: struct kvm_xsave (out)
1034 Returns: 0 on success, -1 on error
1040 This ioctl would copy current vcpu's xsave struct to the userspace.
1045 Capability: KVM_CAP_XSAVE
1048 Parameters: struct kvm_xsave (in)
1049 Returns: 0 on success, -1 on error
1055 This ioctl would copy userspace's xsave struct to the kernel.
1060 Capability: KVM_CAP_XCRS
1063 Parameters: struct kvm_xcrs (out)
1064 Returns: 0 on success, -1 on error
1075 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1079 This ioctl would copy current vcpu's xcrs to the userspace.
1084 Capability: KVM_CAP_XCRS
1087 Parameters: struct kvm_xcrs (in)
1088 Returns: 0 on success, -1 on error
1099 struct kvm_xcr xcrs[KVM_MAX_XCRS];
1103 This ioctl would set vcpu's xcr to the value userspace specified.
1106 4.46 KVM_GET_SUPPORTED_CPUID
1108 Capability: KVM_CAP_EXT_CPUID
1111 Parameters: struct kvm_cpuid2 (in/out)
1112 Returns: 0 on success, -1 on error
1117 struct kvm_cpuid_entry2 entries[0];
1120 #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX 1
1121 #define KVM_CPUID_FLAG_STATEFUL_FUNC 2
1122 #define KVM_CPUID_FLAG_STATE_READ_NEXT 4
1124 struct kvm_cpuid_entry2 {
1135 This ioctl returns x86 cpuid features which are supported by both the hardware
1136 and kvm. Userspace can use the information returned by this ioctl to
1137 construct cpuid information (for KVM_SET_CPUID2) that is consistent with
1138 hardware, kernel, and userspace capabilities, and with user requirements (for
1139 example, the user may wish to constrain cpuid to emulate older hardware,
1140 or for feature consistency across a cluster).
1142 Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure
1143 with the 'nent' field indicating the number of entries in the variable-size
1144 array 'entries'. If the number of entries is too low to describe the cpu
1145 capabilities, an error (E2BIG) is returned. If the number is too high,
1146 the 'nent' field is adjusted and an error (ENOMEM) is returned. If the
1147 number is just right, the 'nent' field is adjusted to the number of valid
1148 entries in the 'entries' array, which is then filled.
1150 The entries returned are the host cpuid as returned by the cpuid instruction,
1151 with unknown or unsupported features masked out. Some features (for example,
1152 x2apic), may not be present in the host cpu, but are exposed by kvm if it can
1153 emulate them efficiently. The fields in each entry are defined as follows:
1155 function: the eax value used to obtain the entry
1156 index: the ecx value used to obtain the entry (for entries that are
1158 flags: an OR of zero or more of the following:
1159 KVM_CPUID_FLAG_SIGNIFCANT_INDEX:
1160 if the index field is valid
1161 KVM_CPUID_FLAG_STATEFUL_FUNC:
1162 if cpuid for this function returns different values for successive
1163 invocations; there will be several entries with the same function,
1164 all with this flag set
1165 KVM_CPUID_FLAG_STATE_READ_NEXT:
1166 for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is
1167 the first entry to be read by a cpu
1168 eax, ebx, ecx, edx: the values returned by the cpuid instruction for
1169 this function/index combination
1171 The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned
1172 as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC
1173 support. Instead it is reported via
1175 ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER)
1177 if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the
1178 feature in userspace, then you can enable the feature for KVM_SET_CPUID2.
1181 4.47 KVM_PPC_GET_PVINFO
1183 Capability: KVM_CAP_PPC_GET_PVINFO
1186 Parameters: struct kvm_ppc_pvinfo (out)
1187 Returns: 0 on success, !0 on error
1189 struct kvm_ppc_pvinfo {
1195 This ioctl fetches PV specific information that need to be passed to the guest
1196 using the device tree or other means from vm context.
1198 The hcall array defines 4 instructions that make up a hypercall.
1200 If any additional field gets added to this structure later on, a bit for that
1201 additional piece of information will be set in the flags bitmap.
1203 The flags bitmap is defined as:
1205 /* the host supports the ePAPR idle hcall
1206 #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0)
1208 4.48 KVM_ASSIGN_PCI_DEVICE
1210 Capability: KVM_CAP_DEVICE_ASSIGNMENT
1211 Architectures: x86 ia64
1213 Parameters: struct kvm_assigned_pci_dev (in)
1214 Returns: 0 on success, -1 on error
1216 Assigns a host PCI device to the VM.
1218 struct kvm_assigned_pci_dev {
1219 __u32 assigned_dev_id;
1229 The PCI device is specified by the triple segnr, busnr, and devfn.
1230 Identification in succeeding service requests is done via assigned_dev_id. The
1231 following flags are specified:
1233 /* Depends on KVM_CAP_IOMMU */
1234 #define KVM_DEV_ASSIGN_ENABLE_IOMMU (1 << 0)
1235 /* The following two depend on KVM_CAP_PCI_2_3 */
1236 #define KVM_DEV_ASSIGN_PCI_2_3 (1 << 1)
1237 #define KVM_DEV_ASSIGN_MASK_INTX (1 << 2)
1239 If KVM_DEV_ASSIGN_PCI_2_3 is set, the kernel will manage legacy INTx interrupts
1240 via the PCI-2.3-compliant device-level mask, thus enable IRQ sharing with other
1241 assigned devices or host devices. KVM_DEV_ASSIGN_MASK_INTX specifies the
1242 guest's view on the INTx mask, see KVM_ASSIGN_SET_INTX_MASK for details.
1244 The KVM_DEV_ASSIGN_ENABLE_IOMMU flag is a mandatory option to ensure
1245 isolation of the device. Usages not specifying this flag are deprecated.
1247 Only PCI header type 0 devices with PCI BAR resources are supported by
1248 device assignment. The user requesting this ioctl must have read/write
1249 access to the PCI sysfs resource files associated with the device.
1252 4.49 KVM_DEASSIGN_PCI_DEVICE
1254 Capability: KVM_CAP_DEVICE_DEASSIGNMENT
1255 Architectures: x86 ia64
1257 Parameters: struct kvm_assigned_pci_dev (in)
1258 Returns: 0 on success, -1 on error
1260 Ends PCI device assignment, releasing all associated resources.
1262 See KVM_CAP_DEVICE_ASSIGNMENT for the data structure. Only assigned_dev_id is
1263 used in kvm_assigned_pci_dev to identify the device.
1266 4.50 KVM_ASSIGN_DEV_IRQ
1268 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1269 Architectures: x86 ia64
1271 Parameters: struct kvm_assigned_irq (in)
1272 Returns: 0 on success, -1 on error
1274 Assigns an IRQ to a passed-through device.
1276 struct kvm_assigned_irq {
1277 __u32 assigned_dev_id;
1278 __u32 host_irq; /* ignored (legacy field) */
1286 The following flags are defined:
1288 #define KVM_DEV_IRQ_HOST_INTX (1 << 0)
1289 #define KVM_DEV_IRQ_HOST_MSI (1 << 1)
1290 #define KVM_DEV_IRQ_HOST_MSIX (1 << 2)
1292 #define KVM_DEV_IRQ_GUEST_INTX (1 << 8)
1293 #define KVM_DEV_IRQ_GUEST_MSI (1 << 9)
1294 #define KVM_DEV_IRQ_GUEST_MSIX (1 << 10)
1296 It is not valid to specify multiple types per host or guest IRQ. However, the
1297 IRQ type of host and guest can differ or can even be null.
1300 4.51 KVM_DEASSIGN_DEV_IRQ
1302 Capability: KVM_CAP_ASSIGN_DEV_IRQ
1303 Architectures: x86 ia64
1305 Parameters: struct kvm_assigned_irq (in)
1306 Returns: 0 on success, -1 on error
1308 Ends an IRQ assignment to a passed-through device.
1310 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1311 by assigned_dev_id, flags must correspond to the IRQ type specified on
1312 KVM_ASSIGN_DEV_IRQ. Partial deassignment of host or guest IRQ is allowed.
1315 4.52 KVM_SET_GSI_ROUTING
1317 Capability: KVM_CAP_IRQ_ROUTING
1318 Architectures: x86 ia64
1320 Parameters: struct kvm_irq_routing (in)
1321 Returns: 0 on success, -1 on error
1323 Sets the GSI routing table entries, overwriting any previously set entries.
1325 struct kvm_irq_routing {
1328 struct kvm_irq_routing_entry entries[0];
1331 No flags are specified so far, the corresponding field must be set to zero.
1333 struct kvm_irq_routing_entry {
1339 struct kvm_irq_routing_irqchip irqchip;
1340 struct kvm_irq_routing_msi msi;
1345 /* gsi routing entry types */
1346 #define KVM_IRQ_ROUTING_IRQCHIP 1
1347 #define KVM_IRQ_ROUTING_MSI 2
1349 No flags are specified so far, the corresponding field must be set to zero.
1351 struct kvm_irq_routing_irqchip {
1356 struct kvm_irq_routing_msi {
1364 4.53 KVM_ASSIGN_SET_MSIX_NR
1366 Capability: KVM_CAP_DEVICE_MSIX
1367 Architectures: x86 ia64
1369 Parameters: struct kvm_assigned_msix_nr (in)
1370 Returns: 0 on success, -1 on error
1372 Set the number of MSI-X interrupts for an assigned device. The number is
1373 reset again by terminating the MSI-X assignment of the device via
1374 KVM_DEASSIGN_DEV_IRQ. Calling this service more than once at any earlier
1377 struct kvm_assigned_msix_nr {
1378 __u32 assigned_dev_id;
1383 #define KVM_MAX_MSIX_PER_DEV 256
1386 4.54 KVM_ASSIGN_SET_MSIX_ENTRY
1388 Capability: KVM_CAP_DEVICE_MSIX
1389 Architectures: x86 ia64
1391 Parameters: struct kvm_assigned_msix_entry (in)
1392 Returns: 0 on success, -1 on error
1394 Specifies the routing of an MSI-X assigned device interrupt to a GSI. Setting
1395 the GSI vector to zero means disabling the interrupt.
1397 struct kvm_assigned_msix_entry {
1398 __u32 assigned_dev_id;
1400 __u16 entry; /* The index of entry in the MSI-X table */
1405 4.55 KVM_SET_TSC_KHZ
1407 Capability: KVM_CAP_TSC_CONTROL
1410 Parameters: virtual tsc_khz
1411 Returns: 0 on success, -1 on error
1413 Specifies the tsc frequency for the virtual machine. The unit of the
1417 4.56 KVM_GET_TSC_KHZ
1419 Capability: KVM_CAP_GET_TSC_KHZ
1423 Returns: virtual tsc-khz on success, negative value on error
1425 Returns the tsc frequency of the guest. The unit of the return value is
1426 KHz. If the host has unstable tsc this ioctl returns -EIO instead as an
1432 Capability: KVM_CAP_IRQCHIP
1435 Parameters: struct kvm_lapic_state (out)
1436 Returns: 0 on success, -1 on error
1438 #define KVM_APIC_REG_SIZE 0x400
1439 struct kvm_lapic_state {
1440 char regs[KVM_APIC_REG_SIZE];
1443 Reads the Local APIC registers and copies them into the input argument. The
1444 data format and layout are the same as documented in the architecture manual.
1449 Capability: KVM_CAP_IRQCHIP
1452 Parameters: struct kvm_lapic_state (in)
1453 Returns: 0 on success, -1 on error
1455 #define KVM_APIC_REG_SIZE 0x400
1456 struct kvm_lapic_state {
1457 char regs[KVM_APIC_REG_SIZE];
1460 Copies the input argument into the the Local APIC registers. The data format
1461 and layout are the same as documented in the architecture manual.
1466 Capability: KVM_CAP_IOEVENTFD
1469 Parameters: struct kvm_ioeventfd (in)
1470 Returns: 0 on success, !0 on error
1472 This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address
1473 within the guest. A guest write in the registered address will signal the
1474 provided event instead of triggering an exit.
1476 struct kvm_ioeventfd {
1478 __u64 addr; /* legal pio/mmio address */
1479 __u32 len; /* 1, 2, 4, or 8 bytes */
1485 The following flags are defined:
1487 #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch)
1488 #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio)
1489 #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign)
1491 If datamatch flag is set, the event will be signaled only if the written value
1492 to the registered address is equal to datamatch in struct kvm_ioeventfd.
1497 Capability: KVM_CAP_SW_TLB
1500 Parameters: struct kvm_dirty_tlb (in)
1501 Returns: 0 on success, -1 on error
1503 struct kvm_dirty_tlb {
1508 This must be called whenever userspace has changed an entry in the shared
1509 TLB, prior to calling KVM_RUN on the associated vcpu.
1511 The "bitmap" field is the userspace address of an array. This array
1512 consists of a number of bits, equal to the total number of TLB entries as
1513 determined by the last successful call to KVM_CONFIG_TLB, rounded up to the
1514 nearest multiple of 64.
1516 Each bit corresponds to one TLB entry, ordered the same as in the shared TLB
1519 The array is little-endian: the bit 0 is the least significant bit of the
1520 first byte, bit 8 is the least significant bit of the second byte, etc.
1521 This avoids any complications with differing word sizes.
1523 The "num_dirty" field is a performance hint for KVM to determine whether it
1524 should skip processing the bitmap and just invalidate everything. It must
1525 be set to the number of set bits in the bitmap.
1528 4.61 KVM_ASSIGN_SET_INTX_MASK
1530 Capability: KVM_CAP_PCI_2_3
1533 Parameters: struct kvm_assigned_pci_dev (in)
1534 Returns: 0 on success, -1 on error
1536 Allows userspace to mask PCI INTx interrupts from the assigned device. The
1537 kernel will not deliver INTx interrupts to the guest between setting and
1538 clearing of KVM_ASSIGN_SET_INTX_MASK via this interface. This enables use of
1539 and emulation of PCI 2.3 INTx disable command register behavior.
1541 This may be used for both PCI 2.3 devices supporting INTx disable natively and
1542 older devices lacking this support. Userspace is responsible for emulating the
1543 read value of the INTx disable bit in the guest visible PCI command register.
1544 When modifying the INTx disable state, userspace should precede updating the
1545 physical device command register by calling this ioctl to inform the kernel of
1546 the new intended INTx mask state.
1548 Note that the kernel uses the device INTx disable bit to internally manage the
1549 device interrupt state for PCI 2.3 devices. Reads of this register may
1550 therefore not match the expected value. Writes should always use the guest
1551 intended INTx disable value rather than attempting to read-copy-update the
1552 current physical device state. Races between user and kernel updates to the
1553 INTx disable bit are handled lazily in the kernel. It's possible the device
1554 may generate unintended interrupts, but they will not be injected into the
1557 See KVM_ASSIGN_DEV_IRQ for the data structure. The target device is specified
1558 by assigned_dev_id. In the flags field, only KVM_DEV_ASSIGN_MASK_INTX is
1562 4.62 KVM_CREATE_SPAPR_TCE
1564 Capability: KVM_CAP_SPAPR_TCE
1565 Architectures: powerpc
1567 Parameters: struct kvm_create_spapr_tce (in)
1568 Returns: file descriptor for manipulating the created TCE table
1570 This creates a virtual TCE (translation control entry) table, which
1571 is an IOMMU for PAPR-style virtual I/O. It is used to translate
1572 logical addresses used in virtual I/O into guest physical addresses,
1573 and provides a scatter/gather capability for PAPR virtual I/O.
1575 /* for KVM_CAP_SPAPR_TCE */
1576 struct kvm_create_spapr_tce {
1581 The liobn field gives the logical IO bus number for which to create a
1582 TCE table. The window_size field specifies the size of the DMA window
1583 which this TCE table will translate - the table will contain one 64
1584 bit TCE entry for every 4kiB of the DMA window.
1586 When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE
1587 table has been created using this ioctl(), the kernel will handle it
1588 in real mode, updating the TCE table. H_PUT_TCE calls for other
1589 liobns will cause a vm exit and must be handled by userspace.
1591 The return value is a file descriptor which can be passed to mmap(2)
1592 to map the created TCE table into userspace. This lets userspace read
1593 the entries written by kernel-handled H_PUT_TCE calls, and also lets
1594 userspace update the TCE table directly which is useful in some
1598 4.63 KVM_ALLOCATE_RMA
1600 Capability: KVM_CAP_PPC_RMA
1601 Architectures: powerpc
1603 Parameters: struct kvm_allocate_rma (out)
1604 Returns: file descriptor for mapping the allocated RMA
1606 This allocates a Real Mode Area (RMA) from the pool allocated at boot
1607 time by the kernel. An RMA is a physically-contiguous, aligned region
1608 of memory used on older POWER processors to provide the memory which
1609 will be accessed by real-mode (MMU off) accesses in a KVM guest.
1610 POWER processors support a set of sizes for the RMA that usually
1611 includes 64MB, 128MB, 256MB and some larger powers of two.
1613 /* for KVM_ALLOCATE_RMA */
1614 struct kvm_allocate_rma {
1618 The return value is a file descriptor which can be passed to mmap(2)
1619 to map the allocated RMA into userspace. The mapped area can then be
1620 passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the
1621 RMA for a virtual machine. The size of the RMA in bytes (which is
1622 fixed at host kernel boot time) is returned in the rma_size field of
1623 the argument structure.
1625 The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl
1626 is supported; 2 if the processor requires all virtual machines to have
1627 an RMA, or 1 if the processor can use an RMA but doesn't require it,
1628 because it supports the Virtual RMA (VRMA) facility.
1633 Capability: KVM_CAP_USER_NMI
1637 Returns: 0 on success, -1 on error
1639 Queues an NMI on the thread's vcpu. Note this is well defined only
1640 when KVM_CREATE_IRQCHIP has not been called, since this is an interface
1641 between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP
1642 has been called, this interface is completely emulated within the kernel.
1644 To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the
1645 following algorithm:
1648 - read the local APIC's state (KVM_GET_LAPIC)
1649 - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1)
1650 - if so, issue KVM_NMI
1653 Some guests configure the LINT1 NMI input to cause a panic, aiding in
1657 4.65 KVM_S390_UCAS_MAP
1659 Capability: KVM_CAP_S390_UCONTROL
1662 Parameters: struct kvm_s390_ucas_mapping (in)
1663 Returns: 0 in case of success
1665 The parameter is defined like this:
1666 struct kvm_s390_ucas_mapping {
1672 This ioctl maps the memory at "user_addr" with the length "length" to
1673 the vcpu's address space starting at "vcpu_addr". All parameters need to
1674 be alligned by 1 megabyte.
1677 4.66 KVM_S390_UCAS_UNMAP
1679 Capability: KVM_CAP_S390_UCONTROL
1682 Parameters: struct kvm_s390_ucas_mapping (in)
1683 Returns: 0 in case of success
1685 The parameter is defined like this:
1686 struct kvm_s390_ucas_mapping {
1692 This ioctl unmaps the memory in the vcpu's address space starting at
1693 "vcpu_addr" with the length "length". The field "user_addr" is ignored.
1694 All parameters need to be alligned by 1 megabyte.
1697 4.67 KVM_S390_VCPU_FAULT
1699 Capability: KVM_CAP_S390_UCONTROL
1702 Parameters: vcpu absolute address (in)
1703 Returns: 0 in case of success
1705 This call creates a page table entry on the virtual cpu's address space
1706 (for user controlled virtual machines) or the virtual machine's address
1707 space (for regular virtual machines). This only works for minor faults,
1708 thus it's recommended to access subject memory page via the user page
1709 table upfront. This is useful to handle validity intercepts for user
1710 controlled virtual machines to fault in the virtual cpu's lowcore pages
1711 prior to calling the KVM_RUN ioctl.
1714 4.68 KVM_SET_ONE_REG
1716 Capability: KVM_CAP_ONE_REG
1719 Parameters: struct kvm_one_reg (in)
1720 Returns: 0 on success, negative value on failure
1722 struct kvm_one_reg {
1727 Using this ioctl, a single vcpu register can be set to a specific value
1728 defined by user space with the passed in struct kvm_one_reg, where id
1729 refers to the register identifier as described below and addr is a pointer
1730 to a variable with the respective size. There can be architecture agnostic
1731 and architecture specific registers. Each have their own range of operation
1732 and their own constants and width. To keep track of the implemented
1733 registers, find a list below:
1735 Arch | Register | Width (bits)
1737 PPC | KVM_REG_PPC_HIOR | 64
1738 PPC | KVM_REG_PPC_IAC1 | 64
1739 PPC | KVM_REG_PPC_IAC2 | 64
1740 PPC | KVM_REG_PPC_IAC3 | 64
1741 PPC | KVM_REG_PPC_IAC4 | 64
1742 PPC | KVM_REG_PPC_DAC1 | 64
1743 PPC | KVM_REG_PPC_DAC2 | 64
1744 PPC | KVM_REG_PPC_DABR | 64
1745 PPC | KVM_REG_PPC_DSCR | 64
1746 PPC | KVM_REG_PPC_PURR | 64
1747 PPC | KVM_REG_PPC_SPURR | 64
1748 PPC | KVM_REG_PPC_DAR | 64
1749 PPC | KVM_REG_PPC_DSISR | 32
1750 PPC | KVM_REG_PPC_AMR | 64
1751 PPC | KVM_REG_PPC_UAMOR | 64
1752 PPC | KVM_REG_PPC_MMCR0 | 64
1753 PPC | KVM_REG_PPC_MMCR1 | 64
1754 PPC | KVM_REG_PPC_MMCRA | 64
1755 PPC | KVM_REG_PPC_PMC1 | 32
1756 PPC | KVM_REG_PPC_PMC2 | 32
1757 PPC | KVM_REG_PPC_PMC3 | 32
1758 PPC | KVM_REG_PPC_PMC4 | 32
1759 PPC | KVM_REG_PPC_PMC5 | 32
1760 PPC | KVM_REG_PPC_PMC6 | 32
1761 PPC | KVM_REG_PPC_PMC7 | 32
1762 PPC | KVM_REG_PPC_PMC8 | 32
1763 PPC | KVM_REG_PPC_FPR0 | 64
1765 PPC | KVM_REG_PPC_FPR31 | 64
1766 PPC | KVM_REG_PPC_VR0 | 128
1768 PPC | KVM_REG_PPC_VR31 | 128
1769 PPC | KVM_REG_PPC_VSR0 | 128
1771 PPC | KVM_REG_PPC_VSR31 | 128
1772 PPC | KVM_REG_PPC_FPSCR | 64
1773 PPC | KVM_REG_PPC_VSCR | 32
1774 PPC | KVM_REG_PPC_VPA_ADDR | 64
1775 PPC | KVM_REG_PPC_VPA_SLB | 128
1776 PPC | KVM_REG_PPC_VPA_DTL | 128
1777 PPC | KVM_REG_PPC_EPCR | 32
1779 ARM registers are mapped using the lower 32 bits. The upper 16 of that
1780 is the register group type, or coprocessor number:
1782 ARM core registers have the following id bit patterns:
1783 0x4002 0000 0010 <index into the kvm_regs struct:16>
1787 4.69 KVM_GET_ONE_REG
1789 Capability: KVM_CAP_ONE_REG
1792 Parameters: struct kvm_one_reg (in and out)
1793 Returns: 0 on success, negative value on failure
1795 This ioctl allows to receive the value of a single register implemented
1796 in a vcpu. The register to read is indicated by the "id" field of the
1797 kvm_one_reg struct passed in. On success, the register value can be found
1798 at the memory location pointed to by "addr".
1800 The list of registers accessible using this interface is identical to the
1804 4.70 KVM_KVMCLOCK_CTRL
1806 Capability: KVM_CAP_KVMCLOCK_CTRL
1807 Architectures: Any that implement pvclocks (currently x86 only)
1810 Returns: 0 on success, -1 on error
1812 This signals to the host kernel that the specified guest is being paused by
1813 userspace. The host will set a flag in the pvclock structure that is checked
1814 from the soft lockup watchdog. The flag is part of the pvclock structure that
1815 is shared between guest and host, specifically the second bit of the flags
1816 field of the pvclock_vcpu_time_info structure. It will be set exclusively by
1817 the host and read/cleared exclusively by the guest. The guest operation of
1818 checking and clearing the flag must an atomic operation so
1819 load-link/store-conditional, or equivalent must be used. There are two cases
1820 where the guest will clear the flag: when the soft lockup watchdog timer resets
1821 itself or when a soft lockup is detected. This ioctl can be called any time
1822 after pausing the vcpu, but before it is resumed.
1827 Capability: KVM_CAP_SIGNAL_MSI
1830 Parameters: struct kvm_msi (in)
1831 Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error
1833 Directly inject a MSI message. Only valid with in-kernel irqchip that handles
1844 No flags are defined so far. The corresponding field must be 0.
1847 4.71 KVM_CREATE_PIT2
1849 Capability: KVM_CAP_PIT2
1852 Parameters: struct kvm_pit_config (in)
1853 Returns: 0 on success, -1 on error
1855 Creates an in-kernel device model for the i8254 PIT. This call is only valid
1856 after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following
1857 parameters have to be passed:
1859 struct kvm_pit_config {
1866 #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */
1868 PIT timer interrupts may use a per-VM kernel thread for injection. If it
1869 exists, this thread will have a name of the following pattern:
1871 kvm-pit/<owner-process-pid>
1873 When running a guest with elevated priorities, the scheduling parameters of
1874 this thread may have to be adjusted accordingly.
1876 This IOCTL replaces the obsolete KVM_CREATE_PIT.
1881 Capability: KVM_CAP_PIT_STATE2
1884 Parameters: struct kvm_pit_state2 (out)
1885 Returns: 0 on success, -1 on error
1887 Retrieves the state of the in-kernel PIT model. Only valid after
1888 KVM_CREATE_PIT2. The state is returned in the following structure:
1890 struct kvm_pit_state2 {
1891 struct kvm_pit_channel_state channels[3];
1898 /* disable PIT in HPET legacy mode */
1899 #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001
1901 This IOCTL replaces the obsolete KVM_GET_PIT.
1906 Capability: KVM_CAP_PIT_STATE2
1909 Parameters: struct kvm_pit_state2 (in)
1910 Returns: 0 on success, -1 on error
1912 Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2.
1913 See KVM_GET_PIT2 for details on struct kvm_pit_state2.
1915 This IOCTL replaces the obsolete KVM_SET_PIT.
1918 4.74 KVM_PPC_GET_SMMU_INFO
1920 Capability: KVM_CAP_PPC_GET_SMMU_INFO
1921 Architectures: powerpc
1924 Returns: 0 on success, -1 on error
1926 This populates and returns a structure describing the features of
1927 the "Server" class MMU emulation supported by KVM.
1928 This can in turn be used by userspace to generate the appropariate
1929 device-tree properties for the guest operating system.
1931 The structure contains some global informations, followed by an
1932 array of supported segment page sizes:
1934 struct kvm_ppc_smmu_info {
1938 struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ];
1941 The supported flags are:
1943 - KVM_PPC_PAGE_SIZES_REAL:
1944 When that flag is set, guest page sizes must "fit" the backing
1945 store page sizes. When not set, any page size in the list can
1946 be used regardless of how they are backed by userspace.
1948 - KVM_PPC_1T_SEGMENTS
1949 The emulated MMU supports 1T segments in addition to the
1952 The "slb_size" field indicates how many SLB entries are supported
1954 The "sps" array contains 8 entries indicating the supported base
1955 page sizes for a segment in increasing order. Each entry is defined
1958 struct kvm_ppc_one_seg_page_size {
1959 __u32 page_shift; /* Base page shift of segment (or 0) */
1960 __u32 slb_enc; /* SLB encoding for BookS */
1961 struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ];
1964 An entry with a "page_shift" of 0 is unused. Because the array is
1965 organized in increasing order, a lookup can stop when encoutering
1968 The "slb_enc" field provides the encoding to use in the SLB for the
1969 page size. The bits are in positions such as the value can directly
1970 be OR'ed into the "vsid" argument of the slbmte instruction.
1972 The "enc" array is a list which for each of those segment base page
1973 size provides the list of supported actual page sizes (which can be
1974 only larger or equal to the base page size), along with the
1975 corresponding encoding in the hash PTE. Similarily, the array is
1976 8 entries sorted by increasing sizes and an entry with a "0" shift
1977 is an empty entry and a terminator:
1979 struct kvm_ppc_one_page_size {
1980 __u32 page_shift; /* Page shift (or 0) */
1981 __u32 pte_enc; /* Encoding in the HPTE (>>12) */
1984 The "pte_enc" field provides a value that can OR'ed into the hash
1985 PTE's RPN field (ie, it needs to be shifted left by 12 to OR it
1986 into the hash PTE second double word).
1990 Capability: KVM_CAP_IRQFD
1993 Parameters: struct kvm_irqfd (in)
1994 Returns: 0 on success, -1 on error
1996 Allows setting an eventfd to directly trigger a guest interrupt.
1997 kvm_irqfd.fd specifies the file descriptor to use as the eventfd and
1998 kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When
1999 an event is tiggered on the eventfd, an interrupt is injected into
2000 the guest using the specified gsi pin. The irqfd is removed using
2001 the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd
2004 With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify
2005 mechanism allowing emulation of level-triggered, irqfd-based
2006 interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an
2007 additional eventfd in the kvm_irqfd.resamplefd field. When operating
2008 in resample mode, posting of an interrupt through kvm_irq.fd asserts
2009 the specified gsi in the irqchip. When the irqchip is resampled, such
2010 as from an EOI, the gsi is de-asserted and the user is notifed via
2011 kvm_irqfd.resamplefd. It is the user's responsibility to re-queue
2012 the interrupt if the device making use of it still requires service.
2013 Note that closing the resamplefd is not sufficient to disable the
2014 irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment
2015 and need not be specified with KVM_IRQFD_FLAG_DEASSIGN.
2017 4.76 KVM_PPC_ALLOCATE_HTAB
2019 Capability: KVM_CAP_PPC_ALLOC_HTAB
2020 Architectures: powerpc
2022 Parameters: Pointer to u32 containing hash table order (in/out)
2023 Returns: 0 on success, -1 on error
2025 This requests the host kernel to allocate an MMU hash table for a
2026 guest using the PAPR paravirtualization interface. This only does
2027 anything if the kernel is configured to use the Book 3S HV style of
2028 virtualization. Otherwise the capability doesn't exist and the ioctl
2029 returns an ENOTTY error. The rest of this description assumes Book 3S
2032 There must be no vcpus running when this ioctl is called; if there
2033 are, it will do nothing and return an EBUSY error.
2035 The parameter is a pointer to a 32-bit unsigned integer variable
2036 containing the order (log base 2) of the desired size of the hash
2037 table, which must be between 18 and 46. On successful return from the
2038 ioctl, it will have been updated with the order of the hash table that
2041 If no hash table has been allocated when any vcpu is asked to run
2042 (with the KVM_RUN ioctl), the host kernel will allocate a
2043 default-sized hash table (16 MB).
2045 If this ioctl is called when a hash table has already been allocated,
2046 the kernel will clear out the existing hash table (zero all HPTEs) and
2047 return the hash table order in the parameter. (If the guest is using
2048 the virtualized real-mode area (VRMA) facility, the kernel will
2049 re-create the VMRA HPTEs on the next KVM_RUN of any vcpu.)
2051 4.77 KVM_S390_INTERRUPT
2055 Type: vm ioctl, vcpu ioctl
2056 Parameters: struct kvm_s390_interrupt (in)
2057 Returns: 0 on success, -1 on error
2059 Allows to inject an interrupt to the guest. Interrupts can be floating
2060 (vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type.
2062 Interrupt parameters are passed via kvm_s390_interrupt:
2064 struct kvm_s390_interrupt {
2070 type can be one of the following:
2072 KVM_S390_SIGP_STOP (vcpu) - sigp restart
2073 KVM_S390_PROGRAM_INT (vcpu) - program check; code in parm
2074 KVM_S390_SIGP_SET_PREFIX (vcpu) - sigp set prefix; prefix address in parm
2075 KVM_S390_RESTART (vcpu) - restart
2076 KVM_S390_INT_VIRTIO (vm) - virtio external interrupt; external interrupt
2077 parameters in parm and parm64
2078 KVM_S390_INT_SERVICE (vm) - sclp external interrupt; sclp parameter in parm
2079 KVM_S390_INT_EMERGENCY (vcpu) - sigp emergency; source cpu in parm
2080 KVM_S390_INT_EXTERNAL_CALL (vcpu) - sigp external call; source cpu in parm
2082 Note that the vcpu ioctl is asynchronous to vcpu execution.
2084 4.78 KVM_PPC_GET_HTAB_FD
2086 Capability: KVM_CAP_PPC_HTAB_FD
2087 Architectures: powerpc
2089 Parameters: Pointer to struct kvm_get_htab_fd (in)
2090 Returns: file descriptor number (>= 0) on success, -1 on error
2092 This returns a file descriptor that can be used either to read out the
2093 entries in the guest's hashed page table (HPT), or to write entries to
2094 initialize the HPT. The returned fd can only be written to if the
2095 KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and
2096 can only be read if that bit is clear. The argument struct looks like
2099 /* For KVM_PPC_GET_HTAB_FD */
2100 struct kvm_get_htab_fd {
2106 /* Values for kvm_get_htab_fd.flags */
2107 #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1)
2108 #define KVM_GET_HTAB_WRITE ((__u64)0x2)
2110 The `start_index' field gives the index in the HPT of the entry at
2111 which to start reading. It is ignored when writing.
2113 Reads on the fd will initially supply information about all
2114 "interesting" HPT entries. Interesting entries are those with the
2115 bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise
2116 all entries. When the end of the HPT is reached, the read() will
2117 return. If read() is called again on the fd, it will start again from
2118 the beginning of the HPT, but will only return HPT entries that have
2119 changed since they were last read.
2121 Data read or written is structured as a header (8 bytes) followed by a
2122 series of valid HPT entries (16 bytes) each. The header indicates how
2123 many valid HPT entries there are and how many invalid entries follow
2124 the valid entries. The invalid entries are not represented explicitly
2125 in the stream. The header format is:
2127 struct kvm_get_htab_header {
2133 Writes to the fd create HPT entries starting at the index given in the
2134 header; first `n_valid' valid entries with contents from the data
2135 written, then `n_invalid' invalid entries, invalidating any previously
2136 valid entries found.
2139 4.77 KVM_ARM_VCPU_INIT
2144 Parameters: struct struct kvm_vcpu_init (in)
2145 Returns: 0 on success; -1 on error
2147 Â EINVAL: Â Â Â the target is unknown, or the combination of features is invalid.
2148 Â ENOENT: Â Â Â a features bit specified is unknown.
2150 This tells KVM what type of CPU to present to the guest, and what
2151 optional features it should have. Â This will cause a reset of the cpu
2152 registers to their initial values. Â If this is not called, KVM_RUN will
2153 return ENOEXEC for that vcpu.
2155 Note that because some registers reflect machine topology, all vcpus
2156 should be created before this ioctl is invoked.
2159 4.78 KVM_GET_REG_LIST
2164 Parameters: struct kvm_reg_list (in/out)
2165 Returns: 0 on success; -1 on error
2167 Â E2BIG: Â Â Â Â the reg index list is too big to fit in the array specified by
2168 Â Â Â Â Â Â Â Â Â Â Â Â the user (the number required will be written into n).
2170 struct kvm_reg_list {
2171 __u64 n; /* number of registers in reg[] */
2175 This ioctl returns the guest registers that are supported for the
2176 KVM_GET_ONE_REG/KVM_SET_ONE_REG calls.
2179 5. The kvm_run structure
2180 ------------------------
2182 Application code obtains a pointer to the kvm_run structure by
2183 mmap()ing a vcpu fd. From that point, application code can control
2184 execution by changing fields in kvm_run prior to calling the KVM_RUN
2185 ioctl, and obtain information about the reason KVM_RUN returned by
2186 looking up structure members.
2190 __u8 request_interrupt_window;
2192 Request that KVM_RUN return when it becomes possible to inject external
2193 interrupts into the guest. Useful in conjunction with KVM_INTERRUPT.
2200 When KVM_RUN has returned successfully (return value 0), this informs
2201 application code why KVM_RUN has returned. Allowable values for this
2202 field are detailed below.
2204 __u8 ready_for_interrupt_injection;
2206 If request_interrupt_window has been specified, this field indicates
2207 an interrupt can be injected now with KVM_INTERRUPT.
2211 The value of the current interrupt flag. Only valid if in-kernel
2212 local APIC is not used.
2216 /* in (pre_kvm_run), out (post_kvm_run) */
2219 The value of the cr8 register. Only valid if in-kernel local APIC is
2220 not used. Both input and output.
2224 The value of the APIC BASE msr. Only valid if in-kernel local
2225 APIC is not used. Both input and output.
2228 /* KVM_EXIT_UNKNOWN */
2230 __u64 hardware_exit_reason;
2233 If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown
2234 reasons. Further architecture-specific information is available in
2235 hardware_exit_reason.
2237 /* KVM_EXIT_FAIL_ENTRY */
2239 __u64 hardware_entry_failure_reason;
2242 If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due
2243 to unknown reasons. Further architecture-specific information is
2244 available in hardware_entry_failure_reason.
2246 /* KVM_EXIT_EXCEPTION */
2256 #define KVM_EXIT_IO_IN 0
2257 #define KVM_EXIT_IO_OUT 1
2259 __u8 size; /* bytes */
2262 __u64 data_offset; /* relative to kvm_run start */
2265 If exit_reason is KVM_EXIT_IO, then the vcpu has
2266 executed a port I/O instruction which could not be satisfied by kvm.
2267 data_offset describes where the data is located (KVM_EXIT_IO_OUT) or
2268 where kvm expects application code to place the data for the next
2269 KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array.
2272 struct kvm_debug_exit_arch arch;
2285 If exit_reason is KVM_EXIT_MMIO, then the vcpu has
2286 executed a memory-mapped I/O instruction which could not be satisfied
2287 by kvm. The 'data' member contains the written data if 'is_write' is
2288 true, and should be filled by application code otherwise.
2290 NOTE: For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_DCR
2291 and KVM_EXIT_PAPR the corresponding
2292 operations are complete (and guest state is consistent) only after userspace
2293 has re-entered the kernel with KVM_RUN. The kernel side will first finish
2294 incomplete operations and then check for pending signals. Userspace
2295 can re-enter the guest with an unmasked signal pending to complete
2298 /* KVM_EXIT_HYPERCALL */
2307 Unused. This was once used for 'hypercall to userspace'. To implement
2308 such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390).
2309 Note KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO.
2311 /* KVM_EXIT_TPR_ACCESS */
2318 To be documented (KVM_TPR_ACCESS_REPORTING).
2320 /* KVM_EXIT_S390_SIEIC */
2323 __u64 mask; /* psw upper half */
2324 __u64 addr; /* psw lower half */
2331 /* KVM_EXIT_S390_RESET */
2332 #define KVM_S390_RESET_POR 1
2333 #define KVM_S390_RESET_CLEAR 2
2334 #define KVM_S390_RESET_SUBSYSTEM 4
2335 #define KVM_S390_RESET_CPU_INIT 8
2336 #define KVM_S390_RESET_IPL 16
2337 __u64 s390_reset_flags;
2341 /* KVM_EXIT_S390_UCONTROL */
2343 __u64 trans_exc_code;
2347 s390 specific. A page fault has occurred for a user controlled virtual
2348 machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be
2349 resolved by the kernel.
2350 The program code and the translation exception code that were placed
2351 in the cpu's lowcore are presented here as defined by the z Architecture
2352 Principles of Operation Book in the Chapter for Dynamic Address Translation
2369 MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch
2370 hypercalls and exit with this exit struct that contains all the guest gprs.
2372 If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall.
2373 Userspace can now handle the hypercall and when it's done modify the gprs as
2374 necessary. Upon guest entry all guest GPRs will then be replaced by the values
2377 /* KVM_EXIT_PAPR_HCALL */
2384 This is used on 64-bit PowerPC when emulating a pSeries partition,
2385 e.g. with the 'pseries' machine type in qemu. It occurs when the
2386 guest does a hypercall using the 'sc 1' instruction. The 'nr' field
2387 contains the hypercall number (from the guest R3), and 'args' contains
2388 the arguments (from the guest R4 - R12). Userspace should put the
2389 return code in 'ret' and any extra returned values in args[].
2390 The possible hypercalls are defined in the Power Architecture Platform
2391 Requirements (PAPR) document available from www.power.org (free
2392 developer registration required to access it).
2394 /* Fix the size of the union. */
2399 * shared registers between kvm and userspace.
2400 * kvm_valid_regs specifies the register classes set by the host
2401 * kvm_dirty_regs specified the register classes dirtied by userspace
2402 * struct kvm_sync_regs is architecture specific, as well as the
2403 * bits for kvm_valid_regs and kvm_dirty_regs
2405 __u64 kvm_valid_regs;
2406 __u64 kvm_dirty_regs;
2408 struct kvm_sync_regs regs;
2412 If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access
2413 certain guest registers without having to call SET/GET_*REGS. Thus we can
2414 avoid some system call overhead if userspace has to handle the exit.
2415 Userspace can query the validity of the structure by checking
2416 kvm_valid_regs for specific bits. These bits are architecture specific
2417 and usually define the validity of a groups of registers. (e.g. one bit
2418 for general purpose registers)
2423 6. Capabilities that can be enabled
2424 -----------------------------------
2426 There are certain capabilities that change the behavior of the virtual CPU when
2427 enabled. To enable them, please see section 4.37. Below you can find a list of
2428 capabilities and what their effect on the vCPU is when enabling them.
2430 The following information is provided along with the description:
2432 Architectures: which instruction set architectures provide this ioctl.
2433 x86 includes both i386 and x86_64.
2435 Parameters: what parameters are accepted by the capability.
2437 Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL)
2438 are not detailed, but errors with specific meanings are.
2445 Returns: 0 on success; -1 on error
2447 This capability enables interception of OSI hypercalls that otherwise would
2448 be treated as normal system calls to be injected into the guest. OSI hypercalls
2449 were invented by Mac-on-Linux to have a standardized communication mechanism
2450 between the guest and the host.
2452 When this capability is enabled, KVM_EXIT_OSI can occur.
2455 6.2 KVM_CAP_PPC_PAPR
2459 Returns: 0 on success; -1 on error
2461 This capability enables interception of PAPR hypercalls. PAPR hypercalls are
2462 done using the hypercall instruction "sc 1".
2464 It also sets the guest privilege level to "supervisor" mode. Usually the guest
2465 runs in "hypervisor" privilege mode with a few missing features.
2467 In addition to the above, it changes the semantics of SDR1. In this mode, the
2468 HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the
2469 HTAB invisible to the guest.
2471 When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur.
2477 Parameters: args[0] is the address of a struct kvm_config_tlb
2478 Returns: 0 on success; -1 on error
2480 struct kvm_config_tlb {
2487 Configures the virtual CPU's TLB array, establishing a shared memory area
2488 between userspace and KVM. The "params" and "array" fields are userspace
2489 addresses of mmu-type-specific data structures. The "array_len" field is an
2490 safety mechanism, and should be set to the size in bytes of the memory that
2491 userspace has reserved for the array. It must be at least the size dictated
2492 by "mmu_type" and "params".
2494 While KVM_RUN is active, the shared region is under control of KVM. Its
2495 contents are undefined, and any modification by userspace results in
2496 boundedly undefined behavior.
2498 On return from KVM_RUN, the shared region will reflect the current state of
2499 the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB
2500 to tell KVM which entries have been changed, prior to calling KVM_RUN again
2503 For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV:
2504 - The "params" field is of type "struct kvm_book3e_206_tlb_params".
2505 - The "array" field points to an array of type "struct
2506 kvm_book3e_206_tlb_entry".
2507 - The array consists of all entries in the first TLB, followed by all
2508 entries in the second TLB.
2509 - Within a TLB, entries are ordered first by increasing set number. Within a
2510 set, entries are ordered by way (increasing ESEL).
2511 - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1)
2512 where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value.
2513 - The tsize field of mas1 shall be set to 4K on TLB0, even though the
2514 hardware ignores this value for TLB0.