]>
Commit | Line | Data |
---|---|---|
1 | The Definitive KVM (Kernel-based Virtual Machine) API Documentation | |
2 | =================================================================== | |
3 | ||
4 | 1. General description | |
5 | ---------------------- | |
6 | ||
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 | |
9 | ||
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 | |
12 | virtual machines | |
13 | ||
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). | |
17 | ||
18 | Only run VM ioctls from the same process (address space) that was used | |
19 | to create the VM. | |
20 | ||
21 | - vcpu ioctls: These query and set attributes that control the operation | |
22 | of a single virtual cpu. | |
23 | ||
24 | Only run vcpu ioctls from the same thread that was used to create the | |
25 | vcpu. | |
26 | ||
27 | ||
28 | 2. File descriptors | |
29 | ------------------- | |
30 | ||
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. | |
39 | ||
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. | |
46 | ||
47 | ||
48 | 3. Extensions | |
49 | ------------- | |
50 | ||
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 | |
54 | queried and used. | |
55 | ||
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. | |
60 | ||
61 | ||
62 | 4. API description | |
63 | ------------------ | |
64 | ||
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 | |
67 | description: | |
68 | ||
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. | |
77 | ||
78 | Architectures: which instruction set architectures provide this ioctl. | |
79 | x86 includes both i386 and x86_64. | |
80 | ||
81 | Type: system, vm, or vcpu. | |
82 | ||
83 | Parameters: what parameters are accepted by the ioctl. | |
84 | ||
85 | Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL) | |
86 | are not detailed, but errors with specific meanings are. | |
87 | ||
88 | ||
89 | 4.1 KVM_GET_API_VERSION | |
90 | ||
91 | Capability: basic | |
92 | Architectures: all | |
93 | Type: system ioctl | |
94 | Parameters: none | |
95 | Returns: the constant KVM_API_VERSION (=12) | |
96 | ||
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. | |
103 | ||
104 | ||
105 | 4.2 KVM_CREATE_VM | |
106 | ||
107 | Capability: basic | |
108 | Architectures: all | |
109 | Type: system ioctl | |
110 | Parameters: machine type identifier (KVM_VM_*) | |
111 | Returns: a VM fd that can be used to control the new virtual machine. | |
112 | ||
113 | The new VM has no virtual cpus and no memory. | |
114 | You probably want to use 0 as machine type. | |
115 | ||
116 | In order to create user controlled virtual machines on S390, check | |
117 | KVM_CAP_S390_UCONTROL and use the flag KVM_VM_S390_UCONTROL as | |
118 | privileged user (CAP_SYS_ADMIN). | |
119 | ||
120 | To use hardware assisted virtualization on MIPS (VZ ASE) rather than | |
121 | the default trap & emulate implementation (which changes the virtual | |
122 | memory layout to fit in user mode), check KVM_CAP_MIPS_VZ and use the | |
123 | flag KVM_VM_MIPS_VZ. | |
124 | ||
125 | ||
126 | On arm64, the physical address size for a VM (IPA Size limit) is limited | |
127 | to 40bits by default. The limit can be configured if the host supports the | |
128 | extension KVM_CAP_ARM_VM_IPA_SIZE. When supported, use | |
129 | KVM_VM_TYPE_ARM_IPA_SIZE(IPA_Bits) to set the size in the machine type | |
130 | identifier, where IPA_Bits is the maximum width of any physical | |
131 | address used by the VM. The IPA_Bits is encoded in bits[7-0] of the | |
132 | machine type identifier. | |
133 | ||
134 | e.g, to configure a guest to use 48bit physical address size : | |
135 | ||
136 | vm_fd = ioctl(dev_fd, KVM_CREATE_VM, KVM_VM_TYPE_ARM_IPA_SIZE(48)); | |
137 | ||
138 | The requested size (IPA_Bits) must be : | |
139 | 0 - Implies default size, 40bits (for backward compatibility) | |
140 | ||
141 | or | |
142 | ||
143 | N - Implies N bits, where N is a positive integer such that, | |
144 | 32 <= N <= Host_IPA_Limit | |
145 | ||
146 | Host_IPA_Limit is the maximum possible value for IPA_Bits on the host and | |
147 | is dependent on the CPU capability and the kernel configuration. The limit can | |
148 | be retrieved using KVM_CAP_ARM_VM_IPA_SIZE of the KVM_CHECK_EXTENSION | |
149 | ioctl() at run-time. | |
150 | ||
151 | Please note that configuring the IPA size does not affect the capability | |
152 | exposed by the guest CPUs in ID_AA64MMFR0_EL1[PARange]. It only affects | |
153 | size of the address translated by the stage2 level (guest physical to | |
154 | host physical address translations). | |
155 | ||
156 | ||
157 | 4.3 KVM_GET_MSR_INDEX_LIST, KVM_GET_MSR_FEATURE_INDEX_LIST | |
158 | ||
159 | Capability: basic, KVM_CAP_GET_MSR_FEATURES for KVM_GET_MSR_FEATURE_INDEX_LIST | |
160 | Architectures: x86 | |
161 | Type: system ioctl | |
162 | Parameters: struct kvm_msr_list (in/out) | |
163 | Returns: 0 on success; -1 on error | |
164 | Errors: | |
165 | EFAULT: the msr index list cannot be read from or written to | |
166 | E2BIG: the msr index list is to be to fit in the array specified by | |
167 | the user. | |
168 | ||
169 | struct kvm_msr_list { | |
170 | __u32 nmsrs; /* number of msrs in entries */ | |
171 | __u32 indices[0]; | |
172 | }; | |
173 | ||
174 | The user fills in the size of the indices array in nmsrs, and in return | |
175 | kvm adjusts nmsrs to reflect the actual number of msrs and fills in the | |
176 | indices array with their numbers. | |
177 | ||
178 | KVM_GET_MSR_INDEX_LIST returns the guest msrs that are supported. The list | |
179 | varies by kvm version and host processor, but does not change otherwise. | |
180 | ||
181 | Note: if kvm indicates supports MCE (KVM_CAP_MCE), then the MCE bank MSRs are | |
182 | not returned in the MSR list, as different vcpus can have a different number | |
183 | of banks, as set via the KVM_X86_SETUP_MCE ioctl. | |
184 | ||
185 | KVM_GET_MSR_FEATURE_INDEX_LIST returns the list of MSRs that can be passed | |
186 | to the KVM_GET_MSRS system ioctl. This lets userspace probe host capabilities | |
187 | and processor features that are exposed via MSRs (e.g., VMX capabilities). | |
188 | This list also varies by kvm version and host processor, but does not change | |
189 | otherwise. | |
190 | ||
191 | ||
192 | 4.4 KVM_CHECK_EXTENSION | |
193 | ||
194 | Capability: basic, KVM_CAP_CHECK_EXTENSION_VM for vm ioctl | |
195 | Architectures: all | |
196 | Type: system ioctl, vm ioctl | |
197 | Parameters: extension identifier (KVM_CAP_*) | |
198 | Returns: 0 if unsupported; 1 (or some other positive integer) if supported | |
199 | ||
200 | The API allows the application to query about extensions to the core | |
201 | kvm API. Userspace passes an extension identifier (an integer) and | |
202 | receives an integer that describes the extension availability. | |
203 | Generally 0 means no and 1 means yes, but some extensions may report | |
204 | additional information in the integer return value. | |
205 | ||
206 | Based on their initialization different VMs may have different capabilities. | |
207 | It is thus encouraged to use the vm ioctl to query for capabilities (available | |
208 | with KVM_CAP_CHECK_EXTENSION_VM on the vm fd) | |
209 | ||
210 | 4.5 KVM_GET_VCPU_MMAP_SIZE | |
211 | ||
212 | Capability: basic | |
213 | Architectures: all | |
214 | Type: system ioctl | |
215 | Parameters: none | |
216 | Returns: size of vcpu mmap area, in bytes | |
217 | ||
218 | The KVM_RUN ioctl (cf.) communicates with userspace via a shared | |
219 | memory region. This ioctl returns the size of that region. See the | |
220 | KVM_RUN documentation for details. | |
221 | ||
222 | ||
223 | 4.6 KVM_SET_MEMORY_REGION | |
224 | ||
225 | Capability: basic | |
226 | Architectures: all | |
227 | Type: vm ioctl | |
228 | Parameters: struct kvm_memory_region (in) | |
229 | Returns: 0 on success, -1 on error | |
230 | ||
231 | This ioctl is obsolete and has been removed. | |
232 | ||
233 | ||
234 | 4.7 KVM_CREATE_VCPU | |
235 | ||
236 | Capability: basic | |
237 | Architectures: all | |
238 | Type: vm ioctl | |
239 | Parameters: vcpu id (apic id on x86) | |
240 | Returns: vcpu fd on success, -1 on error | |
241 | ||
242 | This API adds a vcpu to a virtual machine. No more than max_vcpus may be added. | |
243 | The vcpu id is an integer in the range [0, max_vcpu_id). | |
244 | ||
245 | The recommended max_vcpus value can be retrieved using the KVM_CAP_NR_VCPUS of | |
246 | the KVM_CHECK_EXTENSION ioctl() at run-time. | |
247 | The maximum possible value for max_vcpus can be retrieved using the | |
248 | KVM_CAP_MAX_VCPUS of the KVM_CHECK_EXTENSION ioctl() at run-time. | |
249 | ||
250 | If the KVM_CAP_NR_VCPUS does not exist, you should assume that max_vcpus is 4 | |
251 | cpus max. | |
252 | If the KVM_CAP_MAX_VCPUS does not exist, you should assume that max_vcpus is | |
253 | same as the value returned from KVM_CAP_NR_VCPUS. | |
254 | ||
255 | The maximum possible value for max_vcpu_id can be retrieved using the | |
256 | KVM_CAP_MAX_VCPU_ID of the KVM_CHECK_EXTENSION ioctl() at run-time. | |
257 | ||
258 | If the KVM_CAP_MAX_VCPU_ID does not exist, you should assume that max_vcpu_id | |
259 | is the same as the value returned from KVM_CAP_MAX_VCPUS. | |
260 | ||
261 | On powerpc using book3s_hv mode, the vcpus are mapped onto virtual | |
262 | threads in one or more virtual CPU cores. (This is because the | |
263 | hardware requires all the hardware threads in a CPU core to be in the | |
264 | same partition.) The KVM_CAP_PPC_SMT capability indicates the number | |
265 | of vcpus per virtual core (vcore). The vcore id is obtained by | |
266 | dividing the vcpu id by the number of vcpus per vcore. The vcpus in a | |
267 | given vcore will always be in the same physical core as each other | |
268 | (though that might be a different physical core from time to time). | |
269 | Userspace can control the threading (SMT) mode of the guest by its | |
270 | allocation of vcpu ids. For example, if userspace wants | |
271 | single-threaded guest vcpus, it should make all vcpu ids be a multiple | |
272 | of the number of vcpus per vcore. | |
273 | ||
274 | For virtual cpus that have been created with S390 user controlled virtual | |
275 | machines, the resulting vcpu fd can be memory mapped at page offset | |
276 | KVM_S390_SIE_PAGE_OFFSET in order to obtain a memory map of the virtual | |
277 | cpu's hardware control block. | |
278 | ||
279 | ||
280 | 4.8 KVM_GET_DIRTY_LOG (vm ioctl) | |
281 | ||
282 | Capability: basic | |
283 | Architectures: x86 | |
284 | Type: vm ioctl | |
285 | Parameters: struct kvm_dirty_log (in/out) | |
286 | Returns: 0 on success, -1 on error | |
287 | ||
288 | /* for KVM_GET_DIRTY_LOG */ | |
289 | struct kvm_dirty_log { | |
290 | __u32 slot; | |
291 | __u32 padding; | |
292 | union { | |
293 | void __user *dirty_bitmap; /* one bit per page */ | |
294 | __u64 padding; | |
295 | }; | |
296 | }; | |
297 | ||
298 | Given a memory slot, return a bitmap containing any pages dirtied | |
299 | since the last call to this ioctl. Bit 0 is the first page in the | |
300 | memory slot. Ensure the entire structure is cleared to avoid padding | |
301 | issues. | |
302 | ||
303 | If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 specifies | |
304 | the address space for which you want to return the dirty bitmap. | |
305 | They must be less than the value that KVM_CHECK_EXTENSION returns for | |
306 | the KVM_CAP_MULTI_ADDRESS_SPACE capability. | |
307 | ||
308 | The bits in the dirty bitmap are cleared before the ioctl returns, unless | |
309 | KVM_CAP_MANUAL_DIRTY_LOG_PROTECT is enabled. For more information, | |
310 | see the description of the capability. | |
311 | ||
312 | 4.9 KVM_SET_MEMORY_ALIAS | |
313 | ||
314 | Capability: basic | |
315 | Architectures: x86 | |
316 | Type: vm ioctl | |
317 | Parameters: struct kvm_memory_alias (in) | |
318 | Returns: 0 (success), -1 (error) | |
319 | ||
320 | This ioctl is obsolete and has been removed. | |
321 | ||
322 | ||
323 | 4.10 KVM_RUN | |
324 | ||
325 | Capability: basic | |
326 | Architectures: all | |
327 | Type: vcpu ioctl | |
328 | Parameters: none | |
329 | Returns: 0 on success, -1 on error | |
330 | Errors: | |
331 | EINTR: an unmasked signal is pending | |
332 | ||
333 | This ioctl is used to run a guest virtual cpu. While there are no | |
334 | explicit parameters, there is an implicit parameter block that can be | |
335 | obtained by mmap()ing the vcpu fd at offset 0, with the size given by | |
336 | KVM_GET_VCPU_MMAP_SIZE. The parameter block is formatted as a 'struct | |
337 | kvm_run' (see below). | |
338 | ||
339 | ||
340 | 4.11 KVM_GET_REGS | |
341 | ||
342 | Capability: basic | |
343 | Architectures: all except ARM, arm64 | |
344 | Type: vcpu ioctl | |
345 | Parameters: struct kvm_regs (out) | |
346 | Returns: 0 on success, -1 on error | |
347 | ||
348 | Reads the general purpose registers from the vcpu. | |
349 | ||
350 | /* x86 */ | |
351 | struct kvm_regs { | |
352 | /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */ | |
353 | __u64 rax, rbx, rcx, rdx; | |
354 | __u64 rsi, rdi, rsp, rbp; | |
355 | __u64 r8, r9, r10, r11; | |
356 | __u64 r12, r13, r14, r15; | |
357 | __u64 rip, rflags; | |
358 | }; | |
359 | ||
360 | /* mips */ | |
361 | struct kvm_regs { | |
362 | /* out (KVM_GET_REGS) / in (KVM_SET_REGS) */ | |
363 | __u64 gpr[32]; | |
364 | __u64 hi; | |
365 | __u64 lo; | |
366 | __u64 pc; | |
367 | }; | |
368 | ||
369 | ||
370 | 4.12 KVM_SET_REGS | |
371 | ||
372 | Capability: basic | |
373 | Architectures: all except ARM, arm64 | |
374 | Type: vcpu ioctl | |
375 | Parameters: struct kvm_regs (in) | |
376 | Returns: 0 on success, -1 on error | |
377 | ||
378 | Writes the general purpose registers into the vcpu. | |
379 | ||
380 | See KVM_GET_REGS for the data structure. | |
381 | ||
382 | ||
383 | 4.13 KVM_GET_SREGS | |
384 | ||
385 | Capability: basic | |
386 | Architectures: x86, ppc | |
387 | Type: vcpu ioctl | |
388 | Parameters: struct kvm_sregs (out) | |
389 | Returns: 0 on success, -1 on error | |
390 | ||
391 | Reads special registers from the vcpu. | |
392 | ||
393 | /* x86 */ | |
394 | struct kvm_sregs { | |
395 | struct kvm_segment cs, ds, es, fs, gs, ss; | |
396 | struct kvm_segment tr, ldt; | |
397 | struct kvm_dtable gdt, idt; | |
398 | __u64 cr0, cr2, cr3, cr4, cr8; | |
399 | __u64 efer; | |
400 | __u64 apic_base; | |
401 | __u64 interrupt_bitmap[(KVM_NR_INTERRUPTS + 63) / 64]; | |
402 | }; | |
403 | ||
404 | /* ppc -- see arch/powerpc/include/uapi/asm/kvm.h */ | |
405 | ||
406 | interrupt_bitmap is a bitmap of pending external interrupts. At most | |
407 | one bit may be set. This interrupt has been acknowledged by the APIC | |
408 | but not yet injected into the cpu core. | |
409 | ||
410 | ||
411 | 4.14 KVM_SET_SREGS | |
412 | ||
413 | Capability: basic | |
414 | Architectures: x86, ppc | |
415 | Type: vcpu ioctl | |
416 | Parameters: struct kvm_sregs (in) | |
417 | Returns: 0 on success, -1 on error | |
418 | ||
419 | Writes special registers into the vcpu. See KVM_GET_SREGS for the | |
420 | data structures. | |
421 | ||
422 | ||
423 | 4.15 KVM_TRANSLATE | |
424 | ||
425 | Capability: basic | |
426 | Architectures: x86 | |
427 | Type: vcpu ioctl | |
428 | Parameters: struct kvm_translation (in/out) | |
429 | Returns: 0 on success, -1 on error | |
430 | ||
431 | Translates a virtual address according to the vcpu's current address | |
432 | translation mode. | |
433 | ||
434 | struct kvm_translation { | |
435 | /* in */ | |
436 | __u64 linear_address; | |
437 | ||
438 | /* out */ | |
439 | __u64 physical_address; | |
440 | __u8 valid; | |
441 | __u8 writeable; | |
442 | __u8 usermode; | |
443 | __u8 pad[5]; | |
444 | }; | |
445 | ||
446 | ||
447 | 4.16 KVM_INTERRUPT | |
448 | ||
449 | Capability: basic | |
450 | Architectures: x86, ppc, mips | |
451 | Type: vcpu ioctl | |
452 | Parameters: struct kvm_interrupt (in) | |
453 | Returns: 0 on success, negative on failure. | |
454 | ||
455 | Queues a hardware interrupt vector to be injected. | |
456 | ||
457 | /* for KVM_INTERRUPT */ | |
458 | struct kvm_interrupt { | |
459 | /* in */ | |
460 | __u32 irq; | |
461 | }; | |
462 | ||
463 | X86: | |
464 | ||
465 | Returns: 0 on success, | |
466 | -EEXIST if an interrupt is already enqueued | |
467 | -EINVAL the the irq number is invalid | |
468 | -ENXIO if the PIC is in the kernel | |
469 | -EFAULT if the pointer is invalid | |
470 | ||
471 | Note 'irq' is an interrupt vector, not an interrupt pin or line. This | |
472 | ioctl is useful if the in-kernel PIC is not used. | |
473 | ||
474 | PPC: | |
475 | ||
476 | Queues an external interrupt to be injected. This ioctl is overleaded | |
477 | with 3 different irq values: | |
478 | ||
479 | a) KVM_INTERRUPT_SET | |
480 | ||
481 | This injects an edge type external interrupt into the guest once it's ready | |
482 | to receive interrupts. When injected, the interrupt is done. | |
483 | ||
484 | b) KVM_INTERRUPT_UNSET | |
485 | ||
486 | This unsets any pending interrupt. | |
487 | ||
488 | Only available with KVM_CAP_PPC_UNSET_IRQ. | |
489 | ||
490 | c) KVM_INTERRUPT_SET_LEVEL | |
491 | ||
492 | This injects a level type external interrupt into the guest context. The | |
493 | interrupt stays pending until a specific ioctl with KVM_INTERRUPT_UNSET | |
494 | is triggered. | |
495 | ||
496 | Only available with KVM_CAP_PPC_IRQ_LEVEL. | |
497 | ||
498 | Note that any value for 'irq' other than the ones stated above is invalid | |
499 | and incurs unexpected behavior. | |
500 | ||
501 | MIPS: | |
502 | ||
503 | Queues an external interrupt to be injected into the virtual CPU. A negative | |
504 | interrupt number dequeues the interrupt. | |
505 | ||
506 | ||
507 | 4.17 KVM_DEBUG_GUEST | |
508 | ||
509 | Capability: basic | |
510 | Architectures: none | |
511 | Type: vcpu ioctl | |
512 | Parameters: none) | |
513 | Returns: -1 on error | |
514 | ||
515 | Support for this has been removed. Use KVM_SET_GUEST_DEBUG instead. | |
516 | ||
517 | ||
518 | 4.18 KVM_GET_MSRS | |
519 | ||
520 | Capability: basic (vcpu), KVM_CAP_GET_MSR_FEATURES (system) | |
521 | Architectures: x86 | |
522 | Type: system ioctl, vcpu ioctl | |
523 | Parameters: struct kvm_msrs (in/out) | |
524 | Returns: number of msrs successfully returned; | |
525 | -1 on error | |
526 | ||
527 | When used as a system ioctl: | |
528 | Reads the values of MSR-based features that are available for the VM. This | |
529 | is similar to KVM_GET_SUPPORTED_CPUID, but it returns MSR indices and values. | |
530 | The list of msr-based features can be obtained using KVM_GET_MSR_FEATURE_INDEX_LIST | |
531 | in a system ioctl. | |
532 | ||
533 | When used as a vcpu ioctl: | |
534 | Reads model-specific registers from the vcpu. Supported msr indices can | |
535 | be obtained using KVM_GET_MSR_INDEX_LIST in a system ioctl. | |
536 | ||
537 | struct kvm_msrs { | |
538 | __u32 nmsrs; /* number of msrs in entries */ | |
539 | __u32 pad; | |
540 | ||
541 | struct kvm_msr_entry entries[0]; | |
542 | }; | |
543 | ||
544 | struct kvm_msr_entry { | |
545 | __u32 index; | |
546 | __u32 reserved; | |
547 | __u64 data; | |
548 | }; | |
549 | ||
550 | Application code should set the 'nmsrs' member (which indicates the | |
551 | size of the entries array) and the 'index' member of each array entry. | |
552 | kvm will fill in the 'data' member. | |
553 | ||
554 | ||
555 | 4.19 KVM_SET_MSRS | |
556 | ||
557 | Capability: basic | |
558 | Architectures: x86 | |
559 | Type: vcpu ioctl | |
560 | Parameters: struct kvm_msrs (in) | |
561 | Returns: 0 on success, -1 on error | |
562 | ||
563 | Writes model-specific registers to the vcpu. See KVM_GET_MSRS for the | |
564 | data structures. | |
565 | ||
566 | Application code should set the 'nmsrs' member (which indicates the | |
567 | size of the entries array), and the 'index' and 'data' members of each | |
568 | array entry. | |
569 | ||
570 | ||
571 | 4.20 KVM_SET_CPUID | |
572 | ||
573 | Capability: basic | |
574 | Architectures: x86 | |
575 | Type: vcpu ioctl | |
576 | Parameters: struct kvm_cpuid (in) | |
577 | Returns: 0 on success, -1 on error | |
578 | ||
579 | Defines the vcpu responses to the cpuid instruction. Applications | |
580 | should use the KVM_SET_CPUID2 ioctl if available. | |
581 | ||
582 | ||
583 | struct kvm_cpuid_entry { | |
584 | __u32 function; | |
585 | __u32 eax; | |
586 | __u32 ebx; | |
587 | __u32 ecx; | |
588 | __u32 edx; | |
589 | __u32 padding; | |
590 | }; | |
591 | ||
592 | /* for KVM_SET_CPUID */ | |
593 | struct kvm_cpuid { | |
594 | __u32 nent; | |
595 | __u32 padding; | |
596 | struct kvm_cpuid_entry entries[0]; | |
597 | }; | |
598 | ||
599 | ||
600 | 4.21 KVM_SET_SIGNAL_MASK | |
601 | ||
602 | Capability: basic | |
603 | Architectures: all | |
604 | Type: vcpu ioctl | |
605 | Parameters: struct kvm_signal_mask (in) | |
606 | Returns: 0 on success, -1 on error | |
607 | ||
608 | Defines which signals are blocked during execution of KVM_RUN. This | |
609 | signal mask temporarily overrides the threads signal mask. Any | |
610 | unblocked signal received (except SIGKILL and SIGSTOP, which retain | |
611 | their traditional behaviour) will cause KVM_RUN to return with -EINTR. | |
612 | ||
613 | Note the signal will only be delivered if not blocked by the original | |
614 | signal mask. | |
615 | ||
616 | /* for KVM_SET_SIGNAL_MASK */ | |
617 | struct kvm_signal_mask { | |
618 | __u32 len; | |
619 | __u8 sigset[0]; | |
620 | }; | |
621 | ||
622 | ||
623 | 4.22 KVM_GET_FPU | |
624 | ||
625 | Capability: basic | |
626 | Architectures: x86 | |
627 | Type: vcpu ioctl | |
628 | Parameters: struct kvm_fpu (out) | |
629 | Returns: 0 on success, -1 on error | |
630 | ||
631 | Reads the floating point state from the vcpu. | |
632 | ||
633 | /* for KVM_GET_FPU and KVM_SET_FPU */ | |
634 | struct kvm_fpu { | |
635 | __u8 fpr[8][16]; | |
636 | __u16 fcw; | |
637 | __u16 fsw; | |
638 | __u8 ftwx; /* in fxsave format */ | |
639 | __u8 pad1; | |
640 | __u16 last_opcode; | |
641 | __u64 last_ip; | |
642 | __u64 last_dp; | |
643 | __u8 xmm[16][16]; | |
644 | __u32 mxcsr; | |
645 | __u32 pad2; | |
646 | }; | |
647 | ||
648 | ||
649 | 4.23 KVM_SET_FPU | |
650 | ||
651 | Capability: basic | |
652 | Architectures: x86 | |
653 | Type: vcpu ioctl | |
654 | Parameters: struct kvm_fpu (in) | |
655 | Returns: 0 on success, -1 on error | |
656 | ||
657 | Writes the floating point state to the vcpu. | |
658 | ||
659 | /* for KVM_GET_FPU and KVM_SET_FPU */ | |
660 | struct kvm_fpu { | |
661 | __u8 fpr[8][16]; | |
662 | __u16 fcw; | |
663 | __u16 fsw; | |
664 | __u8 ftwx; /* in fxsave format */ | |
665 | __u8 pad1; | |
666 | __u16 last_opcode; | |
667 | __u64 last_ip; | |
668 | __u64 last_dp; | |
669 | __u8 xmm[16][16]; | |
670 | __u32 mxcsr; | |
671 | __u32 pad2; | |
672 | }; | |
673 | ||
674 | ||
675 | 4.24 KVM_CREATE_IRQCHIP | |
676 | ||
677 | Capability: KVM_CAP_IRQCHIP, KVM_CAP_S390_IRQCHIP (s390) | |
678 | Architectures: x86, ARM, arm64, s390 | |
679 | Type: vm ioctl | |
680 | Parameters: none | |
681 | Returns: 0 on success, -1 on error | |
682 | ||
683 | Creates an interrupt controller model in the kernel. | |
684 | On x86, creates a virtual ioapic, a virtual PIC (two PICs, nested), and sets up | |
685 | future vcpus to have a local APIC. IRQ routing for GSIs 0-15 is set to both | |
686 | PIC and IOAPIC; GSI 16-23 only go to the IOAPIC. | |
687 | On ARM/arm64, a GICv2 is created. Any other GIC versions require the usage of | |
688 | KVM_CREATE_DEVICE, which also supports creating a GICv2. Using | |
689 | KVM_CREATE_DEVICE is preferred over KVM_CREATE_IRQCHIP for GICv2. | |
690 | On s390, a dummy irq routing table is created. | |
691 | ||
692 | Note that on s390 the KVM_CAP_S390_IRQCHIP vm capability needs to be enabled | |
693 | before KVM_CREATE_IRQCHIP can be used. | |
694 | ||
695 | ||
696 | 4.25 KVM_IRQ_LINE | |
697 | ||
698 | Capability: KVM_CAP_IRQCHIP | |
699 | Architectures: x86, arm, arm64 | |
700 | Type: vm ioctl | |
701 | Parameters: struct kvm_irq_level | |
702 | Returns: 0 on success, -1 on error | |
703 | ||
704 | Sets the level of a GSI input to the interrupt controller model in the kernel. | |
705 | On some architectures it is required that an interrupt controller model has | |
706 | been previously created with KVM_CREATE_IRQCHIP. Note that edge-triggered | |
707 | interrupts require the level to be set to 1 and then back to 0. | |
708 | ||
709 | On real hardware, interrupt pins can be active-low or active-high. This | |
710 | does not matter for the level field of struct kvm_irq_level: 1 always | |
711 | means active (asserted), 0 means inactive (deasserted). | |
712 | ||
713 | x86 allows the operating system to program the interrupt polarity | |
714 | (active-low/active-high) for level-triggered interrupts, and KVM used | |
715 | to consider the polarity. However, due to bitrot in the handling of | |
716 | active-low interrupts, the above convention is now valid on x86 too. | |
717 | This is signaled by KVM_CAP_X86_IOAPIC_POLARITY_IGNORED. Userspace | |
718 | should not present interrupts to the guest as active-low unless this | |
719 | capability is present (or unless it is not using the in-kernel irqchip, | |
720 | of course). | |
721 | ||
722 | ||
723 | ARM/arm64 can signal an interrupt either at the CPU level, or at the | |
724 | in-kernel irqchip (GIC), and for in-kernel irqchip can tell the GIC to | |
725 | use PPIs designated for specific cpus. The irq field is interpreted | |
726 | like this: | |
727 | ||
728 | Â bits: | 31 ... 24 | 23 ... 16 | 15 ... 0 | | |
729 | field: | irq_type | vcpu_index | irq_id | | |
730 | ||
731 | The irq_type field has the following values: | |
732 | - irq_type[0]: out-of-kernel GIC: irq_id 0 is IRQ, irq_id 1 is FIQ | |
733 | - irq_type[1]: in-kernel GIC: SPI, irq_id between 32 and 1019 (incl.) | |
734 | (the vcpu_index field is ignored) | |
735 | - irq_type[2]: in-kernel GIC: PPI, irq_id between 16 and 31 (incl.) | |
736 | ||
737 | (The irq_id field thus corresponds nicely to the IRQ ID in the ARM GIC specs) | |
738 | ||
739 | In both cases, level is used to assert/deassert the line. | |
740 | ||
741 | struct kvm_irq_level { | |
742 | union { | |
743 | __u32 irq; /* GSI */ | |
744 | __s32 status; /* not used for KVM_IRQ_LEVEL */ | |
745 | }; | |
746 | __u32 level; /* 0 or 1 */ | |
747 | }; | |
748 | ||
749 | ||
750 | 4.26 KVM_GET_IRQCHIP | |
751 | ||
752 | Capability: KVM_CAP_IRQCHIP | |
753 | Architectures: x86 | |
754 | Type: vm ioctl | |
755 | Parameters: struct kvm_irqchip (in/out) | |
756 | Returns: 0 on success, -1 on error | |
757 | ||
758 | Reads the state of a kernel interrupt controller created with | |
759 | KVM_CREATE_IRQCHIP into a buffer provided by the caller. | |
760 | ||
761 | struct kvm_irqchip { | |
762 | __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */ | |
763 | __u32 pad; | |
764 | union { | |
765 | char dummy[512]; /* reserving space */ | |
766 | struct kvm_pic_state pic; | |
767 | struct kvm_ioapic_state ioapic; | |
768 | } chip; | |
769 | }; | |
770 | ||
771 | ||
772 | 4.27 KVM_SET_IRQCHIP | |
773 | ||
774 | Capability: KVM_CAP_IRQCHIP | |
775 | Architectures: x86 | |
776 | Type: vm ioctl | |
777 | Parameters: struct kvm_irqchip (in) | |
778 | Returns: 0 on success, -1 on error | |
779 | ||
780 | Sets the state of a kernel interrupt controller created with | |
781 | KVM_CREATE_IRQCHIP from a buffer provided by the caller. | |
782 | ||
783 | struct kvm_irqchip { | |
784 | __u32 chip_id; /* 0 = PIC1, 1 = PIC2, 2 = IOAPIC */ | |
785 | __u32 pad; | |
786 | union { | |
787 | char dummy[512]; /* reserving space */ | |
788 | struct kvm_pic_state pic; | |
789 | struct kvm_ioapic_state ioapic; | |
790 | } chip; | |
791 | }; | |
792 | ||
793 | ||
794 | 4.28 KVM_XEN_HVM_CONFIG | |
795 | ||
796 | Capability: KVM_CAP_XEN_HVM | |
797 | Architectures: x86 | |
798 | Type: vm ioctl | |
799 | Parameters: struct kvm_xen_hvm_config (in) | |
800 | Returns: 0 on success, -1 on error | |
801 | ||
802 | Sets the MSR that the Xen HVM guest uses to initialize its hypercall | |
803 | page, and provides the starting address and size of the hypercall | |
804 | blobs in userspace. When the guest writes the MSR, kvm copies one | |
805 | page of a blob (32- or 64-bit, depending on the vcpu mode) to guest | |
806 | memory. | |
807 | ||
808 | struct kvm_xen_hvm_config { | |
809 | __u32 flags; | |
810 | __u32 msr; | |
811 | __u64 blob_addr_32; | |
812 | __u64 blob_addr_64; | |
813 | __u8 blob_size_32; | |
814 | __u8 blob_size_64; | |
815 | __u8 pad2[30]; | |
816 | }; | |
817 | ||
818 | ||
819 | 4.29 KVM_GET_CLOCK | |
820 | ||
821 | Capability: KVM_CAP_ADJUST_CLOCK | |
822 | Architectures: x86 | |
823 | Type: vm ioctl | |
824 | Parameters: struct kvm_clock_data (out) | |
825 | Returns: 0 on success, -1 on error | |
826 | ||
827 | Gets the current timestamp of kvmclock as seen by the current guest. In | |
828 | conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios | |
829 | such as migration. | |
830 | ||
831 | When KVM_CAP_ADJUST_CLOCK is passed to KVM_CHECK_EXTENSION, it returns the | |
832 | set of bits that KVM can return in struct kvm_clock_data's flag member. | |
833 | ||
834 | The only flag defined now is KVM_CLOCK_TSC_STABLE. If set, the returned | |
835 | value is the exact kvmclock value seen by all VCPUs at the instant | |
836 | when KVM_GET_CLOCK was called. If clear, the returned value is simply | |
837 | CLOCK_MONOTONIC plus a constant offset; the offset can be modified | |
838 | with KVM_SET_CLOCK. KVM will try to make all VCPUs follow this clock, | |
839 | but the exact value read by each VCPU could differ, because the host | |
840 | TSC is not stable. | |
841 | ||
842 | struct kvm_clock_data { | |
843 | __u64 clock; /* kvmclock current value */ | |
844 | __u32 flags; | |
845 | __u32 pad[9]; | |
846 | }; | |
847 | ||
848 | ||
849 | 4.30 KVM_SET_CLOCK | |
850 | ||
851 | Capability: KVM_CAP_ADJUST_CLOCK | |
852 | Architectures: x86 | |
853 | Type: vm ioctl | |
854 | Parameters: struct kvm_clock_data (in) | |
855 | Returns: 0 on success, -1 on error | |
856 | ||
857 | Sets the current timestamp of kvmclock to the value specified in its parameter. | |
858 | In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios | |
859 | such as migration. | |
860 | ||
861 | struct kvm_clock_data { | |
862 | __u64 clock; /* kvmclock current value */ | |
863 | __u32 flags; | |
864 | __u32 pad[9]; | |
865 | }; | |
866 | ||
867 | ||
868 | 4.31 KVM_GET_VCPU_EVENTS | |
869 | ||
870 | Capability: KVM_CAP_VCPU_EVENTS | |
871 | Extended by: KVM_CAP_INTR_SHADOW | |
872 | Architectures: x86, arm, arm64 | |
873 | Type: vcpu ioctl | |
874 | Parameters: struct kvm_vcpu_event (out) | |
875 | Returns: 0 on success, -1 on error | |
876 | ||
877 | X86: | |
878 | ||
879 | Gets currently pending exceptions, interrupts, and NMIs as well as related | |
880 | states of the vcpu. | |
881 | ||
882 | struct kvm_vcpu_events { | |
883 | struct { | |
884 | __u8 injected; | |
885 | __u8 nr; | |
886 | __u8 has_error_code; | |
887 | __u8 pending; | |
888 | __u32 error_code; | |
889 | } exception; | |
890 | struct { | |
891 | __u8 injected; | |
892 | __u8 nr; | |
893 | __u8 soft; | |
894 | __u8 shadow; | |
895 | } interrupt; | |
896 | struct { | |
897 | __u8 injected; | |
898 | __u8 pending; | |
899 | __u8 masked; | |
900 | __u8 pad; | |
901 | } nmi; | |
902 | __u32 sipi_vector; | |
903 | __u32 flags; | |
904 | struct { | |
905 | __u8 smm; | |
906 | __u8 pending; | |
907 | __u8 smm_inside_nmi; | |
908 | __u8 latched_init; | |
909 | } smi; | |
910 | __u8 reserved[27]; | |
911 | __u8 exception_has_payload; | |
912 | __u64 exception_payload; | |
913 | }; | |
914 | ||
915 | The following bits are defined in the flags field: | |
916 | ||
917 | - KVM_VCPUEVENT_VALID_SHADOW may be set to signal that | |
918 | interrupt.shadow contains a valid state. | |
919 | ||
920 | - KVM_VCPUEVENT_VALID_SMM may be set to signal that smi contains a | |
921 | valid state. | |
922 | ||
923 | - KVM_VCPUEVENT_VALID_PAYLOAD may be set to signal that the | |
924 | exception_has_payload, exception_payload, and exception.pending | |
925 | fields contain a valid state. This bit will be set whenever | |
926 | KVM_CAP_EXCEPTION_PAYLOAD is enabled. | |
927 | ||
928 | ARM/ARM64: | |
929 | ||
930 | If the guest accesses a device that is being emulated by the host kernel in | |
931 | such a way that a real device would generate a physical SError, KVM may make | |
932 | a virtual SError pending for that VCPU. This system error interrupt remains | |
933 | pending until the guest takes the exception by unmasking PSTATE.A. | |
934 | ||
935 | Running the VCPU may cause it to take a pending SError, or make an access that | |
936 | causes an SError to become pending. The event's description is only valid while | |
937 | the VPCU is not running. | |
938 | ||
939 | This API provides a way to read and write the pending 'event' state that is not | |
940 | visible to the guest. To save, restore or migrate a VCPU the struct representing | |
941 | the state can be read then written using this GET/SET API, along with the other | |
942 | guest-visible registers. It is not possible to 'cancel' an SError that has been | |
943 | made pending. | |
944 | ||
945 | A device being emulated in user-space may also wish to generate an SError. To do | |
946 | this the events structure can be populated by user-space. The current state | |
947 | should be read first, to ensure no existing SError is pending. If an existing | |
948 | SError is pending, the architecture's 'Multiple SError interrupts' rules should | |
949 | be followed. (2.5.3 of DDI0587.a "ARM Reliability, Availability, and | |
950 | Serviceability (RAS) Specification"). | |
951 | ||
952 | SError exceptions always have an ESR value. Some CPUs have the ability to | |
953 | specify what the virtual SError's ESR value should be. These systems will | |
954 | advertise KVM_CAP_ARM_INJECT_SERROR_ESR. In this case exception.has_esr will | |
955 | always have a non-zero value when read, and the agent making an SError pending | |
956 | should specify the ISS field in the lower 24 bits of exception.serror_esr. If | |
957 | the system supports KVM_CAP_ARM_INJECT_SERROR_ESR, but user-space sets the events | |
958 | with exception.has_esr as zero, KVM will choose an ESR. | |
959 | ||
960 | Specifying exception.has_esr on a system that does not support it will return | |
961 | -EINVAL. Setting anything other than the lower 24bits of exception.serror_esr | |
962 | will return -EINVAL. | |
963 | ||
964 | struct kvm_vcpu_events { | |
965 | struct { | |
966 | __u8 serror_pending; | |
967 | __u8 serror_has_esr; | |
968 | /* Align it to 8 bytes */ | |
969 | __u8 pad[6]; | |
970 | __u64 serror_esr; | |
971 | } exception; | |
972 | __u32 reserved[12]; | |
973 | }; | |
974 | ||
975 | 4.32 KVM_SET_VCPU_EVENTS | |
976 | ||
977 | Capability: KVM_CAP_VCPU_EVENTS | |
978 | Extended by: KVM_CAP_INTR_SHADOW | |
979 | Architectures: x86, arm, arm64 | |
980 | Type: vcpu ioctl | |
981 | Parameters: struct kvm_vcpu_event (in) | |
982 | Returns: 0 on success, -1 on error | |
983 | ||
984 | X86: | |
985 | ||
986 | Set pending exceptions, interrupts, and NMIs as well as related states of the | |
987 | vcpu. | |
988 | ||
989 | See KVM_GET_VCPU_EVENTS for the data structure. | |
990 | ||
991 | Fields that may be modified asynchronously by running VCPUs can be excluded | |
992 | from the update. These fields are nmi.pending, sipi_vector, smi.smm, | |
993 | smi.pending. Keep the corresponding bits in the flags field cleared to | |
994 | suppress overwriting the current in-kernel state. The bits are: | |
995 | ||
996 | KVM_VCPUEVENT_VALID_NMI_PENDING - transfer nmi.pending to the kernel | |
997 | KVM_VCPUEVENT_VALID_SIPI_VECTOR - transfer sipi_vector | |
998 | KVM_VCPUEVENT_VALID_SMM - transfer the smi sub-struct. | |
999 | ||
1000 | If KVM_CAP_INTR_SHADOW is available, KVM_VCPUEVENT_VALID_SHADOW can be set in | |
1001 | the flags field to signal that interrupt.shadow contains a valid state and | |
1002 | shall be written into the VCPU. | |
1003 | ||
1004 | KVM_VCPUEVENT_VALID_SMM can only be set if KVM_CAP_X86_SMM is available. | |
1005 | ||
1006 | If KVM_CAP_EXCEPTION_PAYLOAD is enabled, KVM_VCPUEVENT_VALID_PAYLOAD | |
1007 | can be set in the flags field to signal that the | |
1008 | exception_has_payload, exception_payload, and exception.pending fields | |
1009 | contain a valid state and shall be written into the VCPU. | |
1010 | ||
1011 | ARM/ARM64: | |
1012 | ||
1013 | Set the pending SError exception state for this VCPU. It is not possible to | |
1014 | 'cancel' an Serror that has been made pending. | |
1015 | ||
1016 | See KVM_GET_VCPU_EVENTS for the data structure. | |
1017 | ||
1018 | ||
1019 | 4.33 KVM_GET_DEBUGREGS | |
1020 | ||
1021 | Capability: KVM_CAP_DEBUGREGS | |
1022 | Architectures: x86 | |
1023 | Type: vm ioctl | |
1024 | Parameters: struct kvm_debugregs (out) | |
1025 | Returns: 0 on success, -1 on error | |
1026 | ||
1027 | Reads debug registers from the vcpu. | |
1028 | ||
1029 | struct kvm_debugregs { | |
1030 | __u64 db[4]; | |
1031 | __u64 dr6; | |
1032 | __u64 dr7; | |
1033 | __u64 flags; | |
1034 | __u64 reserved[9]; | |
1035 | }; | |
1036 | ||
1037 | ||
1038 | 4.34 KVM_SET_DEBUGREGS | |
1039 | ||
1040 | Capability: KVM_CAP_DEBUGREGS | |
1041 | Architectures: x86 | |
1042 | Type: vm ioctl | |
1043 | Parameters: struct kvm_debugregs (in) | |
1044 | Returns: 0 on success, -1 on error | |
1045 | ||
1046 | Writes debug registers into the vcpu. | |
1047 | ||
1048 | See KVM_GET_DEBUGREGS for the data structure. The flags field is unused | |
1049 | yet and must be cleared on entry. | |
1050 | ||
1051 | ||
1052 | 4.35 KVM_SET_USER_MEMORY_REGION | |
1053 | ||
1054 | Capability: KVM_CAP_USER_MEM | |
1055 | Architectures: all | |
1056 | Type: vm ioctl | |
1057 | Parameters: struct kvm_userspace_memory_region (in) | |
1058 | Returns: 0 on success, -1 on error | |
1059 | ||
1060 | struct kvm_userspace_memory_region { | |
1061 | __u32 slot; | |
1062 | __u32 flags; | |
1063 | __u64 guest_phys_addr; | |
1064 | __u64 memory_size; /* bytes */ | |
1065 | __u64 userspace_addr; /* start of the userspace allocated memory */ | |
1066 | }; | |
1067 | ||
1068 | /* for kvm_memory_region::flags */ | |
1069 | #define KVM_MEM_LOG_DIRTY_PAGES (1UL << 0) | |
1070 | #define KVM_MEM_READONLY (1UL << 1) | |
1071 | ||
1072 | This ioctl allows the user to create or modify a guest physical memory | |
1073 | slot. When changing an existing slot, it may be moved in the guest | |
1074 | physical memory space, or its flags may be modified. It may not be | |
1075 | resized. Slots may not overlap in guest physical address space. | |
1076 | Bits 0-15 of "slot" specifies the slot id and this value should be | |
1077 | less than the maximum number of user memory slots supported per VM. | |
1078 | The maximum allowed slots can be queried using KVM_CAP_NR_MEMSLOTS, | |
1079 | if this capability is supported by the architecture. | |
1080 | ||
1081 | If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 of "slot" | |
1082 | specifies the address space which is being modified. They must be | |
1083 | less than the value that KVM_CHECK_EXTENSION returns for the | |
1084 | KVM_CAP_MULTI_ADDRESS_SPACE capability. Slots in separate address spaces | |
1085 | are unrelated; the restriction on overlapping slots only applies within | |
1086 | each address space. | |
1087 | ||
1088 | Memory for the region is taken starting at the address denoted by the | |
1089 | field userspace_addr, which must point at user addressable memory for | |
1090 | the entire memory slot size. Any object may back this memory, including | |
1091 | anonymous memory, ordinary files, and hugetlbfs. | |
1092 | ||
1093 | It is recommended that the lower 21 bits of guest_phys_addr and userspace_addr | |
1094 | be identical. This allows large pages in the guest to be backed by large | |
1095 | pages in the host. | |
1096 | ||
1097 | The flags field supports two flags: KVM_MEM_LOG_DIRTY_PAGES and | |
1098 | KVM_MEM_READONLY. The former can be set to instruct KVM to keep track of | |
1099 | writes to memory within the slot. See KVM_GET_DIRTY_LOG ioctl to know how to | |
1100 | use it. The latter can be set, if KVM_CAP_READONLY_MEM capability allows it, | |
1101 | to make a new slot read-only. In this case, writes to this memory will be | |
1102 | posted to userspace as KVM_EXIT_MMIO exits. | |
1103 | ||
1104 | When the KVM_CAP_SYNC_MMU capability is available, changes in the backing of | |
1105 | the memory region are automatically reflected into the guest. For example, an | |
1106 | mmap() that affects the region will be made visible immediately. Another | |
1107 | example is madvise(MADV_DROP). | |
1108 | ||
1109 | It is recommended to use this API instead of the KVM_SET_MEMORY_REGION ioctl. | |
1110 | The KVM_SET_MEMORY_REGION does not allow fine grained control over memory | |
1111 | allocation and is deprecated. | |
1112 | ||
1113 | ||
1114 | 4.36 KVM_SET_TSS_ADDR | |
1115 | ||
1116 | Capability: KVM_CAP_SET_TSS_ADDR | |
1117 | Architectures: x86 | |
1118 | Type: vm ioctl | |
1119 | Parameters: unsigned long tss_address (in) | |
1120 | Returns: 0 on success, -1 on error | |
1121 | ||
1122 | This ioctl defines the physical address of a three-page region in the guest | |
1123 | physical address space. The region must be within the first 4GB of the | |
1124 | guest physical address space and must not conflict with any memory slot | |
1125 | or any mmio address. The guest may malfunction if it accesses this memory | |
1126 | region. | |
1127 | ||
1128 | This ioctl is required on Intel-based hosts. This is needed on Intel hardware | |
1129 | because of a quirk in the virtualization implementation (see the internals | |
1130 | documentation when it pops into existence). | |
1131 | ||
1132 | ||
1133 | 4.37 KVM_ENABLE_CAP | |
1134 | ||
1135 | Capability: KVM_CAP_ENABLE_CAP | |
1136 | Architectures: mips, ppc, s390 | |
1137 | Type: vcpu ioctl | |
1138 | Parameters: struct kvm_enable_cap (in) | |
1139 | Returns: 0 on success; -1 on error | |
1140 | ||
1141 | Capability: KVM_CAP_ENABLE_CAP_VM | |
1142 | Architectures: all | |
1143 | Type: vcpu ioctl | |
1144 | Parameters: struct kvm_enable_cap (in) | |
1145 | Returns: 0 on success; -1 on error | |
1146 | ||
1147 | +Not all extensions are enabled by default. Using this ioctl the application | |
1148 | can enable an extension, making it available to the guest. | |
1149 | ||
1150 | On systems that do not support this ioctl, it always fails. On systems that | |
1151 | do support it, it only works for extensions that are supported for enablement. | |
1152 | ||
1153 | To check if a capability can be enabled, the KVM_CHECK_EXTENSION ioctl should | |
1154 | be used. | |
1155 | ||
1156 | struct kvm_enable_cap { | |
1157 | /* in */ | |
1158 | __u32 cap; | |
1159 | ||
1160 | The capability that is supposed to get enabled. | |
1161 | ||
1162 | __u32 flags; | |
1163 | ||
1164 | A bitfield indicating future enhancements. Has to be 0 for now. | |
1165 | ||
1166 | __u64 args[4]; | |
1167 | ||
1168 | Arguments for enabling a feature. If a feature needs initial values to | |
1169 | function properly, this is the place to put them. | |
1170 | ||
1171 | __u8 pad[64]; | |
1172 | }; | |
1173 | ||
1174 | The vcpu ioctl should be used for vcpu-specific capabilities, the vm ioctl | |
1175 | for vm-wide capabilities. | |
1176 | ||
1177 | 4.38 KVM_GET_MP_STATE | |
1178 | ||
1179 | Capability: KVM_CAP_MP_STATE | |
1180 | Architectures: x86, s390, arm, arm64 | |
1181 | Type: vcpu ioctl | |
1182 | Parameters: struct kvm_mp_state (out) | |
1183 | Returns: 0 on success; -1 on error | |
1184 | ||
1185 | struct kvm_mp_state { | |
1186 | __u32 mp_state; | |
1187 | }; | |
1188 | ||
1189 | Returns the vcpu's current "multiprocessing state" (though also valid on | |
1190 | uniprocessor guests). | |
1191 | ||
1192 | Possible values are: | |
1193 | ||
1194 | - KVM_MP_STATE_RUNNABLE: the vcpu is currently running [x86,arm/arm64] | |
1195 | - KVM_MP_STATE_UNINITIALIZED: the vcpu is an application processor (AP) | |
1196 | which has not yet received an INIT signal [x86] | |
1197 | - KVM_MP_STATE_INIT_RECEIVED: the vcpu has received an INIT signal, and is | |
1198 | now ready for a SIPI [x86] | |
1199 | - KVM_MP_STATE_HALTED: the vcpu has executed a HLT instruction and | |
1200 | is waiting for an interrupt [x86] | |
1201 | - KVM_MP_STATE_SIPI_RECEIVED: the vcpu has just received a SIPI (vector | |
1202 | accessible via KVM_GET_VCPU_EVENTS) [x86] | |
1203 | - KVM_MP_STATE_STOPPED: the vcpu is stopped [s390,arm/arm64] | |
1204 | - KVM_MP_STATE_CHECK_STOP: the vcpu is in a special error state [s390] | |
1205 | - KVM_MP_STATE_OPERATING: the vcpu is operating (running or halted) | |
1206 | [s390] | |
1207 | - KVM_MP_STATE_LOAD: the vcpu is in a special load/startup state | |
1208 | [s390] | |
1209 | ||
1210 | On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an | |
1211 | in-kernel irqchip, the multiprocessing state must be maintained by userspace on | |
1212 | these architectures. | |
1213 | ||
1214 | For arm/arm64: | |
1215 | ||
1216 | The only states that are valid are KVM_MP_STATE_STOPPED and | |
1217 | KVM_MP_STATE_RUNNABLE which reflect if the vcpu is paused or not. | |
1218 | ||
1219 | 4.39 KVM_SET_MP_STATE | |
1220 | ||
1221 | Capability: KVM_CAP_MP_STATE | |
1222 | Architectures: x86, s390, arm, arm64 | |
1223 | Type: vcpu ioctl | |
1224 | Parameters: struct kvm_mp_state (in) | |
1225 | Returns: 0 on success; -1 on error | |
1226 | ||
1227 | Sets the vcpu's current "multiprocessing state"; see KVM_GET_MP_STATE for | |
1228 | arguments. | |
1229 | ||
1230 | On x86, this ioctl is only useful after KVM_CREATE_IRQCHIP. Without an | |
1231 | in-kernel irqchip, the multiprocessing state must be maintained by userspace on | |
1232 | these architectures. | |
1233 | ||
1234 | For arm/arm64: | |
1235 | ||
1236 | The only states that are valid are KVM_MP_STATE_STOPPED and | |
1237 | KVM_MP_STATE_RUNNABLE which reflect if the vcpu should be paused or not. | |
1238 | ||
1239 | 4.40 KVM_SET_IDENTITY_MAP_ADDR | |
1240 | ||
1241 | Capability: KVM_CAP_SET_IDENTITY_MAP_ADDR | |
1242 | Architectures: x86 | |
1243 | Type: vm ioctl | |
1244 | Parameters: unsigned long identity (in) | |
1245 | Returns: 0 on success, -1 on error | |
1246 | ||
1247 | This ioctl defines the physical address of a one-page region in the guest | |
1248 | physical address space. The region must be within the first 4GB of the | |
1249 | guest physical address space and must not conflict with any memory slot | |
1250 | or any mmio address. The guest may malfunction if it accesses this memory | |
1251 | region. | |
1252 | ||
1253 | Setting the address to 0 will result in resetting the address to its default | |
1254 | (0xfffbc000). | |
1255 | ||
1256 | This ioctl is required on Intel-based hosts. This is needed on Intel hardware | |
1257 | because of a quirk in the virtualization implementation (see the internals | |
1258 | documentation when it pops into existence). | |
1259 | ||
1260 | Fails if any VCPU has already been created. | |
1261 | ||
1262 | 4.41 KVM_SET_BOOT_CPU_ID | |
1263 | ||
1264 | Capability: KVM_CAP_SET_BOOT_CPU_ID | |
1265 | Architectures: x86 | |
1266 | Type: vm ioctl | |
1267 | Parameters: unsigned long vcpu_id | |
1268 | Returns: 0 on success, -1 on error | |
1269 | ||
1270 | Define which vcpu is the Bootstrap Processor (BSP). Values are the same | |
1271 | as the vcpu id in KVM_CREATE_VCPU. If this ioctl is not called, the default | |
1272 | is vcpu 0. | |
1273 | ||
1274 | ||
1275 | 4.42 KVM_GET_XSAVE | |
1276 | ||
1277 | Capability: KVM_CAP_XSAVE | |
1278 | Architectures: x86 | |
1279 | Type: vcpu ioctl | |
1280 | Parameters: struct kvm_xsave (out) | |
1281 | Returns: 0 on success, -1 on error | |
1282 | ||
1283 | struct kvm_xsave { | |
1284 | __u32 region[1024]; | |
1285 | }; | |
1286 | ||
1287 | This ioctl would copy current vcpu's xsave struct to the userspace. | |
1288 | ||
1289 | ||
1290 | 4.43 KVM_SET_XSAVE | |
1291 | ||
1292 | Capability: KVM_CAP_XSAVE | |
1293 | Architectures: x86 | |
1294 | Type: vcpu ioctl | |
1295 | Parameters: struct kvm_xsave (in) | |
1296 | Returns: 0 on success, -1 on error | |
1297 | ||
1298 | struct kvm_xsave { | |
1299 | __u32 region[1024]; | |
1300 | }; | |
1301 | ||
1302 | This ioctl would copy userspace's xsave struct to the kernel. | |
1303 | ||
1304 | ||
1305 | 4.44 KVM_GET_XCRS | |
1306 | ||
1307 | Capability: KVM_CAP_XCRS | |
1308 | Architectures: x86 | |
1309 | Type: vcpu ioctl | |
1310 | Parameters: struct kvm_xcrs (out) | |
1311 | Returns: 0 on success, -1 on error | |
1312 | ||
1313 | struct kvm_xcr { | |
1314 | __u32 xcr; | |
1315 | __u32 reserved; | |
1316 | __u64 value; | |
1317 | }; | |
1318 | ||
1319 | struct kvm_xcrs { | |
1320 | __u32 nr_xcrs; | |
1321 | __u32 flags; | |
1322 | struct kvm_xcr xcrs[KVM_MAX_XCRS]; | |
1323 | __u64 padding[16]; | |
1324 | }; | |
1325 | ||
1326 | This ioctl would copy current vcpu's xcrs to the userspace. | |
1327 | ||
1328 | ||
1329 | 4.45 KVM_SET_XCRS | |
1330 | ||
1331 | Capability: KVM_CAP_XCRS | |
1332 | Architectures: x86 | |
1333 | Type: vcpu ioctl | |
1334 | Parameters: struct kvm_xcrs (in) | |
1335 | Returns: 0 on success, -1 on error | |
1336 | ||
1337 | struct kvm_xcr { | |
1338 | __u32 xcr; | |
1339 | __u32 reserved; | |
1340 | __u64 value; | |
1341 | }; | |
1342 | ||
1343 | struct kvm_xcrs { | |
1344 | __u32 nr_xcrs; | |
1345 | __u32 flags; | |
1346 | struct kvm_xcr xcrs[KVM_MAX_XCRS]; | |
1347 | __u64 padding[16]; | |
1348 | }; | |
1349 | ||
1350 | This ioctl would set vcpu's xcr to the value userspace specified. | |
1351 | ||
1352 | ||
1353 | 4.46 KVM_GET_SUPPORTED_CPUID | |
1354 | ||
1355 | Capability: KVM_CAP_EXT_CPUID | |
1356 | Architectures: x86 | |
1357 | Type: system ioctl | |
1358 | Parameters: struct kvm_cpuid2 (in/out) | |
1359 | Returns: 0 on success, -1 on error | |
1360 | ||
1361 | struct kvm_cpuid2 { | |
1362 | __u32 nent; | |
1363 | __u32 padding; | |
1364 | struct kvm_cpuid_entry2 entries[0]; | |
1365 | }; | |
1366 | ||
1367 | #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) | |
1368 | #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) | |
1369 | #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) | |
1370 | ||
1371 | struct kvm_cpuid_entry2 { | |
1372 | __u32 function; | |
1373 | __u32 index; | |
1374 | __u32 flags; | |
1375 | __u32 eax; | |
1376 | __u32 ebx; | |
1377 | __u32 ecx; | |
1378 | __u32 edx; | |
1379 | __u32 padding[3]; | |
1380 | }; | |
1381 | ||
1382 | This ioctl returns x86 cpuid features which are supported by both the | |
1383 | hardware and kvm in its default configuration. Userspace can use the | |
1384 | information returned by this ioctl to construct cpuid information (for | |
1385 | KVM_SET_CPUID2) that is consistent with hardware, kernel, and | |
1386 | userspace capabilities, and with user requirements (for example, the | |
1387 | user may wish to constrain cpuid to emulate older hardware, or for | |
1388 | feature consistency across a cluster). | |
1389 | ||
1390 | Note that certain capabilities, such as KVM_CAP_X86_DISABLE_EXITS, may | |
1391 | expose cpuid features (e.g. MONITOR) which are not supported by kvm in | |
1392 | its default configuration. If userspace enables such capabilities, it | |
1393 | is responsible for modifying the results of this ioctl appropriately. | |
1394 | ||
1395 | Userspace invokes KVM_GET_SUPPORTED_CPUID by passing a kvm_cpuid2 structure | |
1396 | with the 'nent' field indicating the number of entries in the variable-size | |
1397 | array 'entries'. If the number of entries is too low to describe the cpu | |
1398 | capabilities, an error (E2BIG) is returned. If the number is too high, | |
1399 | the 'nent' field is adjusted and an error (ENOMEM) is returned. If the | |
1400 | number is just right, the 'nent' field is adjusted to the number of valid | |
1401 | entries in the 'entries' array, which is then filled. | |
1402 | ||
1403 | The entries returned are the host cpuid as returned by the cpuid instruction, | |
1404 | with unknown or unsupported features masked out. Some features (for example, | |
1405 | x2apic), may not be present in the host cpu, but are exposed by kvm if it can | |
1406 | emulate them efficiently. The fields in each entry are defined as follows: | |
1407 | ||
1408 | function: the eax value used to obtain the entry | |
1409 | index: the ecx value used to obtain the entry (for entries that are | |
1410 | affected by ecx) | |
1411 | flags: an OR of zero or more of the following: | |
1412 | KVM_CPUID_FLAG_SIGNIFCANT_INDEX: | |
1413 | if the index field is valid | |
1414 | KVM_CPUID_FLAG_STATEFUL_FUNC: | |
1415 | if cpuid for this function returns different values for successive | |
1416 | invocations; there will be several entries with the same function, | |
1417 | all with this flag set | |
1418 | KVM_CPUID_FLAG_STATE_READ_NEXT: | |
1419 | for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is | |
1420 | the first entry to be read by a cpu | |
1421 | eax, ebx, ecx, edx: the values returned by the cpuid instruction for | |
1422 | this function/index combination | |
1423 | ||
1424 | The TSC deadline timer feature (CPUID leaf 1, ecx[24]) is always returned | |
1425 | as false, since the feature depends on KVM_CREATE_IRQCHIP for local APIC | |
1426 | support. Instead it is reported via | |
1427 | ||
1428 | ioctl(KVM_CHECK_EXTENSION, KVM_CAP_TSC_DEADLINE_TIMER) | |
1429 | ||
1430 | if that returns true and you use KVM_CREATE_IRQCHIP, or if you emulate the | |
1431 | feature in userspace, then you can enable the feature for KVM_SET_CPUID2. | |
1432 | ||
1433 | ||
1434 | 4.47 KVM_PPC_GET_PVINFO | |
1435 | ||
1436 | Capability: KVM_CAP_PPC_GET_PVINFO | |
1437 | Architectures: ppc | |
1438 | Type: vm ioctl | |
1439 | Parameters: struct kvm_ppc_pvinfo (out) | |
1440 | Returns: 0 on success, !0 on error | |
1441 | ||
1442 | struct kvm_ppc_pvinfo { | |
1443 | __u32 flags; | |
1444 | __u32 hcall[4]; | |
1445 | __u8 pad[108]; | |
1446 | }; | |
1447 | ||
1448 | This ioctl fetches PV specific information that need to be passed to the guest | |
1449 | using the device tree or other means from vm context. | |
1450 | ||
1451 | The hcall array defines 4 instructions that make up a hypercall. | |
1452 | ||
1453 | If any additional field gets added to this structure later on, a bit for that | |
1454 | additional piece of information will be set in the flags bitmap. | |
1455 | ||
1456 | The flags bitmap is defined as: | |
1457 | ||
1458 | /* the host supports the ePAPR idle hcall | |
1459 | #define KVM_PPC_PVINFO_FLAGS_EV_IDLE (1<<0) | |
1460 | ||
1461 | 4.52 KVM_SET_GSI_ROUTING | |
1462 | ||
1463 | Capability: KVM_CAP_IRQ_ROUTING | |
1464 | Architectures: x86 s390 arm arm64 | |
1465 | Type: vm ioctl | |
1466 | Parameters: struct kvm_irq_routing (in) | |
1467 | Returns: 0 on success, -1 on error | |
1468 | ||
1469 | Sets the GSI routing table entries, overwriting any previously set entries. | |
1470 | ||
1471 | On arm/arm64, GSI routing has the following limitation: | |
1472 | - GSI routing does not apply to KVM_IRQ_LINE but only to KVM_IRQFD. | |
1473 | ||
1474 | struct kvm_irq_routing { | |
1475 | __u32 nr; | |
1476 | __u32 flags; | |
1477 | struct kvm_irq_routing_entry entries[0]; | |
1478 | }; | |
1479 | ||
1480 | No flags are specified so far, the corresponding field must be set to zero. | |
1481 | ||
1482 | struct kvm_irq_routing_entry { | |
1483 | __u32 gsi; | |
1484 | __u32 type; | |
1485 | __u32 flags; | |
1486 | __u32 pad; | |
1487 | union { | |
1488 | struct kvm_irq_routing_irqchip irqchip; | |
1489 | struct kvm_irq_routing_msi msi; | |
1490 | struct kvm_irq_routing_s390_adapter adapter; | |
1491 | struct kvm_irq_routing_hv_sint hv_sint; | |
1492 | __u32 pad[8]; | |
1493 | } u; | |
1494 | }; | |
1495 | ||
1496 | /* gsi routing entry types */ | |
1497 | #define KVM_IRQ_ROUTING_IRQCHIP 1 | |
1498 | #define KVM_IRQ_ROUTING_MSI 2 | |
1499 | #define KVM_IRQ_ROUTING_S390_ADAPTER 3 | |
1500 | #define KVM_IRQ_ROUTING_HV_SINT 4 | |
1501 | ||
1502 | flags: | |
1503 | - KVM_MSI_VALID_DEVID: used along with KVM_IRQ_ROUTING_MSI routing entry | |
1504 | type, specifies that the devid field contains a valid value. The per-VM | |
1505 | KVM_CAP_MSI_DEVID capability advertises the requirement to provide | |
1506 | the device ID. If this capability is not available, userspace should | |
1507 | never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. | |
1508 | - zero otherwise | |
1509 | ||
1510 | struct kvm_irq_routing_irqchip { | |
1511 | __u32 irqchip; | |
1512 | __u32 pin; | |
1513 | }; | |
1514 | ||
1515 | struct kvm_irq_routing_msi { | |
1516 | __u32 address_lo; | |
1517 | __u32 address_hi; | |
1518 | __u32 data; | |
1519 | union { | |
1520 | __u32 pad; | |
1521 | __u32 devid; | |
1522 | }; | |
1523 | }; | |
1524 | ||
1525 | If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier | |
1526 | for the device that wrote the MSI message. For PCI, this is usually a | |
1527 | BFD identifier in the lower 16 bits. | |
1528 | ||
1529 | On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS | |
1530 | feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, | |
1531 | address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of | |
1532 | address_hi must be zero. | |
1533 | ||
1534 | struct kvm_irq_routing_s390_adapter { | |
1535 | __u64 ind_addr; | |
1536 | __u64 summary_addr; | |
1537 | __u64 ind_offset; | |
1538 | __u32 summary_offset; | |
1539 | __u32 adapter_id; | |
1540 | }; | |
1541 | ||
1542 | struct kvm_irq_routing_hv_sint { | |
1543 | __u32 vcpu; | |
1544 | __u32 sint; | |
1545 | }; | |
1546 | ||
1547 | ||
1548 | 4.55 KVM_SET_TSC_KHZ | |
1549 | ||
1550 | Capability: KVM_CAP_TSC_CONTROL | |
1551 | Architectures: x86 | |
1552 | Type: vcpu ioctl | |
1553 | Parameters: virtual tsc_khz | |
1554 | Returns: 0 on success, -1 on error | |
1555 | ||
1556 | Specifies the tsc frequency for the virtual machine. The unit of the | |
1557 | frequency is KHz. | |
1558 | ||
1559 | ||
1560 | 4.56 KVM_GET_TSC_KHZ | |
1561 | ||
1562 | Capability: KVM_CAP_GET_TSC_KHZ | |
1563 | Architectures: x86 | |
1564 | Type: vcpu ioctl | |
1565 | Parameters: none | |
1566 | Returns: virtual tsc-khz on success, negative value on error | |
1567 | ||
1568 | Returns the tsc frequency of the guest. The unit of the return value is | |
1569 | KHz. If the host has unstable tsc this ioctl returns -EIO instead as an | |
1570 | error. | |
1571 | ||
1572 | ||
1573 | 4.57 KVM_GET_LAPIC | |
1574 | ||
1575 | Capability: KVM_CAP_IRQCHIP | |
1576 | Architectures: x86 | |
1577 | Type: vcpu ioctl | |
1578 | Parameters: struct kvm_lapic_state (out) | |
1579 | Returns: 0 on success, -1 on error | |
1580 | ||
1581 | #define KVM_APIC_REG_SIZE 0x400 | |
1582 | struct kvm_lapic_state { | |
1583 | char regs[KVM_APIC_REG_SIZE]; | |
1584 | }; | |
1585 | ||
1586 | Reads the Local APIC registers and copies them into the input argument. The | |
1587 | data format and layout are the same as documented in the architecture manual. | |
1588 | ||
1589 | If KVM_X2APIC_API_USE_32BIT_IDS feature of KVM_CAP_X2APIC_API is | |
1590 | enabled, then the format of APIC_ID register depends on the APIC mode | |
1591 | (reported by MSR_IA32_APICBASE) of its VCPU. x2APIC stores APIC ID in | |
1592 | the APIC_ID register (bytes 32-35). xAPIC only allows an 8-bit APIC ID | |
1593 | which is stored in bits 31-24 of the APIC register, or equivalently in | |
1594 | byte 35 of struct kvm_lapic_state's regs field. KVM_GET_LAPIC must then | |
1595 | be called after MSR_IA32_APICBASE has been set with KVM_SET_MSR. | |
1596 | ||
1597 | If KVM_X2APIC_API_USE_32BIT_IDS feature is disabled, struct kvm_lapic_state | |
1598 | always uses xAPIC format. | |
1599 | ||
1600 | ||
1601 | 4.58 KVM_SET_LAPIC | |
1602 | ||
1603 | Capability: KVM_CAP_IRQCHIP | |
1604 | Architectures: x86 | |
1605 | Type: vcpu ioctl | |
1606 | Parameters: struct kvm_lapic_state (in) | |
1607 | Returns: 0 on success, -1 on error | |
1608 | ||
1609 | #define KVM_APIC_REG_SIZE 0x400 | |
1610 | struct kvm_lapic_state { | |
1611 | char regs[KVM_APIC_REG_SIZE]; | |
1612 | }; | |
1613 | ||
1614 | Copies the input argument into the Local APIC registers. The data format | |
1615 | and layout are the same as documented in the architecture manual. | |
1616 | ||
1617 | The format of the APIC ID register (bytes 32-35 of struct kvm_lapic_state's | |
1618 | regs field) depends on the state of the KVM_CAP_X2APIC_API capability. | |
1619 | See the note in KVM_GET_LAPIC. | |
1620 | ||
1621 | ||
1622 | 4.59 KVM_IOEVENTFD | |
1623 | ||
1624 | Capability: KVM_CAP_IOEVENTFD | |
1625 | Architectures: all | |
1626 | Type: vm ioctl | |
1627 | Parameters: struct kvm_ioeventfd (in) | |
1628 | Returns: 0 on success, !0 on error | |
1629 | ||
1630 | This ioctl attaches or detaches an ioeventfd to a legal pio/mmio address | |
1631 | within the guest. A guest write in the registered address will signal the | |
1632 | provided event instead of triggering an exit. | |
1633 | ||
1634 | struct kvm_ioeventfd { | |
1635 | __u64 datamatch; | |
1636 | __u64 addr; /* legal pio/mmio address */ | |
1637 | __u32 len; /* 0, 1, 2, 4, or 8 bytes */ | |
1638 | __s32 fd; | |
1639 | __u32 flags; | |
1640 | __u8 pad[36]; | |
1641 | }; | |
1642 | ||
1643 | For the special case of virtio-ccw devices on s390, the ioevent is matched | |
1644 | to a subchannel/virtqueue tuple instead. | |
1645 | ||
1646 | The following flags are defined: | |
1647 | ||
1648 | #define KVM_IOEVENTFD_FLAG_DATAMATCH (1 << kvm_ioeventfd_flag_nr_datamatch) | |
1649 | #define KVM_IOEVENTFD_FLAG_PIO (1 << kvm_ioeventfd_flag_nr_pio) | |
1650 | #define KVM_IOEVENTFD_FLAG_DEASSIGN (1 << kvm_ioeventfd_flag_nr_deassign) | |
1651 | #define KVM_IOEVENTFD_FLAG_VIRTIO_CCW_NOTIFY \ | |
1652 | (1 << kvm_ioeventfd_flag_nr_virtio_ccw_notify) | |
1653 | ||
1654 | If datamatch flag is set, the event will be signaled only if the written value | |
1655 | to the registered address is equal to datamatch in struct kvm_ioeventfd. | |
1656 | ||
1657 | For virtio-ccw devices, addr contains the subchannel id and datamatch the | |
1658 | virtqueue index. | |
1659 | ||
1660 | With KVM_CAP_IOEVENTFD_ANY_LENGTH, a zero length ioeventfd is allowed, and | |
1661 | the kernel will ignore the length of guest write and may get a faster vmexit. | |
1662 | The speedup may only apply to specific architectures, but the ioeventfd will | |
1663 | work anyway. | |
1664 | ||
1665 | 4.60 KVM_DIRTY_TLB | |
1666 | ||
1667 | Capability: KVM_CAP_SW_TLB | |
1668 | Architectures: ppc | |
1669 | Type: vcpu ioctl | |
1670 | Parameters: struct kvm_dirty_tlb (in) | |
1671 | Returns: 0 on success, -1 on error | |
1672 | ||
1673 | struct kvm_dirty_tlb { | |
1674 | __u64 bitmap; | |
1675 | __u32 num_dirty; | |
1676 | }; | |
1677 | ||
1678 | This must be called whenever userspace has changed an entry in the shared | |
1679 | TLB, prior to calling KVM_RUN on the associated vcpu. | |
1680 | ||
1681 | The "bitmap" field is the userspace address of an array. This array | |
1682 | consists of a number of bits, equal to the total number of TLB entries as | |
1683 | determined by the last successful call to KVM_CONFIG_TLB, rounded up to the | |
1684 | nearest multiple of 64. | |
1685 | ||
1686 | Each bit corresponds to one TLB entry, ordered the same as in the shared TLB | |
1687 | array. | |
1688 | ||
1689 | The array is little-endian: the bit 0 is the least significant bit of the | |
1690 | first byte, bit 8 is the least significant bit of the second byte, etc. | |
1691 | This avoids any complications with differing word sizes. | |
1692 | ||
1693 | The "num_dirty" field is a performance hint for KVM to determine whether it | |
1694 | should skip processing the bitmap and just invalidate everything. It must | |
1695 | be set to the number of set bits in the bitmap. | |
1696 | ||
1697 | ||
1698 | 4.62 KVM_CREATE_SPAPR_TCE | |
1699 | ||
1700 | Capability: KVM_CAP_SPAPR_TCE | |
1701 | Architectures: powerpc | |
1702 | Type: vm ioctl | |
1703 | Parameters: struct kvm_create_spapr_tce (in) | |
1704 | Returns: file descriptor for manipulating the created TCE table | |
1705 | ||
1706 | This creates a virtual TCE (translation control entry) table, which | |
1707 | is an IOMMU for PAPR-style virtual I/O. It is used to translate | |
1708 | logical addresses used in virtual I/O into guest physical addresses, | |
1709 | and provides a scatter/gather capability for PAPR virtual I/O. | |
1710 | ||
1711 | /* for KVM_CAP_SPAPR_TCE */ | |
1712 | struct kvm_create_spapr_tce { | |
1713 | __u64 liobn; | |
1714 | __u32 window_size; | |
1715 | }; | |
1716 | ||
1717 | The liobn field gives the logical IO bus number for which to create a | |
1718 | TCE table. The window_size field specifies the size of the DMA window | |
1719 | which this TCE table will translate - the table will contain one 64 | |
1720 | bit TCE entry for every 4kiB of the DMA window. | |
1721 | ||
1722 | When the guest issues an H_PUT_TCE hcall on a liobn for which a TCE | |
1723 | table has been created using this ioctl(), the kernel will handle it | |
1724 | in real mode, updating the TCE table. H_PUT_TCE calls for other | |
1725 | liobns will cause a vm exit and must be handled by userspace. | |
1726 | ||
1727 | The return value is a file descriptor which can be passed to mmap(2) | |
1728 | to map the created TCE table into userspace. This lets userspace read | |
1729 | the entries written by kernel-handled H_PUT_TCE calls, and also lets | |
1730 | userspace update the TCE table directly which is useful in some | |
1731 | circumstances. | |
1732 | ||
1733 | ||
1734 | 4.63 KVM_ALLOCATE_RMA | |
1735 | ||
1736 | Capability: KVM_CAP_PPC_RMA | |
1737 | Architectures: powerpc | |
1738 | Type: vm ioctl | |
1739 | Parameters: struct kvm_allocate_rma (out) | |
1740 | Returns: file descriptor for mapping the allocated RMA | |
1741 | ||
1742 | This allocates a Real Mode Area (RMA) from the pool allocated at boot | |
1743 | time by the kernel. An RMA is a physically-contiguous, aligned region | |
1744 | of memory used on older POWER processors to provide the memory which | |
1745 | will be accessed by real-mode (MMU off) accesses in a KVM guest. | |
1746 | POWER processors support a set of sizes for the RMA that usually | |
1747 | includes 64MB, 128MB, 256MB and some larger powers of two. | |
1748 | ||
1749 | /* for KVM_ALLOCATE_RMA */ | |
1750 | struct kvm_allocate_rma { | |
1751 | __u64 rma_size; | |
1752 | }; | |
1753 | ||
1754 | The return value is a file descriptor which can be passed to mmap(2) | |
1755 | to map the allocated RMA into userspace. The mapped area can then be | |
1756 | passed to the KVM_SET_USER_MEMORY_REGION ioctl to establish it as the | |
1757 | RMA for a virtual machine. The size of the RMA in bytes (which is | |
1758 | fixed at host kernel boot time) is returned in the rma_size field of | |
1759 | the argument structure. | |
1760 | ||
1761 | The KVM_CAP_PPC_RMA capability is 1 or 2 if the KVM_ALLOCATE_RMA ioctl | |
1762 | is supported; 2 if the processor requires all virtual machines to have | |
1763 | an RMA, or 1 if the processor can use an RMA but doesn't require it, | |
1764 | because it supports the Virtual RMA (VRMA) facility. | |
1765 | ||
1766 | ||
1767 | 4.64 KVM_NMI | |
1768 | ||
1769 | Capability: KVM_CAP_USER_NMI | |
1770 | Architectures: x86 | |
1771 | Type: vcpu ioctl | |
1772 | Parameters: none | |
1773 | Returns: 0 on success, -1 on error | |
1774 | ||
1775 | Queues an NMI on the thread's vcpu. Note this is well defined only | |
1776 | when KVM_CREATE_IRQCHIP has not been called, since this is an interface | |
1777 | between the virtual cpu core and virtual local APIC. After KVM_CREATE_IRQCHIP | |
1778 | has been called, this interface is completely emulated within the kernel. | |
1779 | ||
1780 | To use this to emulate the LINT1 input with KVM_CREATE_IRQCHIP, use the | |
1781 | following algorithm: | |
1782 | ||
1783 | - pause the vcpu | |
1784 | - read the local APIC's state (KVM_GET_LAPIC) | |
1785 | - check whether changing LINT1 will queue an NMI (see the LVT entry for LINT1) | |
1786 | - if so, issue KVM_NMI | |
1787 | - resume the vcpu | |
1788 | ||
1789 | Some guests configure the LINT1 NMI input to cause a panic, aiding in | |
1790 | debugging. | |
1791 | ||
1792 | ||
1793 | 4.65 KVM_S390_UCAS_MAP | |
1794 | ||
1795 | Capability: KVM_CAP_S390_UCONTROL | |
1796 | Architectures: s390 | |
1797 | Type: vcpu ioctl | |
1798 | Parameters: struct kvm_s390_ucas_mapping (in) | |
1799 | Returns: 0 in case of success | |
1800 | ||
1801 | The parameter is defined like this: | |
1802 | struct kvm_s390_ucas_mapping { | |
1803 | __u64 user_addr; | |
1804 | __u64 vcpu_addr; | |
1805 | __u64 length; | |
1806 | }; | |
1807 | ||
1808 | This ioctl maps the memory at "user_addr" with the length "length" to | |
1809 | the vcpu's address space starting at "vcpu_addr". All parameters need to | |
1810 | be aligned by 1 megabyte. | |
1811 | ||
1812 | ||
1813 | 4.66 KVM_S390_UCAS_UNMAP | |
1814 | ||
1815 | Capability: KVM_CAP_S390_UCONTROL | |
1816 | Architectures: s390 | |
1817 | Type: vcpu ioctl | |
1818 | Parameters: struct kvm_s390_ucas_mapping (in) | |
1819 | Returns: 0 in case of success | |
1820 | ||
1821 | The parameter is defined like this: | |
1822 | struct kvm_s390_ucas_mapping { | |
1823 | __u64 user_addr; | |
1824 | __u64 vcpu_addr; | |
1825 | __u64 length; | |
1826 | }; | |
1827 | ||
1828 | This ioctl unmaps the memory in the vcpu's address space starting at | |
1829 | "vcpu_addr" with the length "length". The field "user_addr" is ignored. | |
1830 | All parameters need to be aligned by 1 megabyte. | |
1831 | ||
1832 | ||
1833 | 4.67 KVM_S390_VCPU_FAULT | |
1834 | ||
1835 | Capability: KVM_CAP_S390_UCONTROL | |
1836 | Architectures: s390 | |
1837 | Type: vcpu ioctl | |
1838 | Parameters: vcpu absolute address (in) | |
1839 | Returns: 0 in case of success | |
1840 | ||
1841 | This call creates a page table entry on the virtual cpu's address space | |
1842 | (for user controlled virtual machines) or the virtual machine's address | |
1843 | space (for regular virtual machines). This only works for minor faults, | |
1844 | thus it's recommended to access subject memory page via the user page | |
1845 | table upfront. This is useful to handle validity intercepts for user | |
1846 | controlled virtual machines to fault in the virtual cpu's lowcore pages | |
1847 | prior to calling the KVM_RUN ioctl. | |
1848 | ||
1849 | ||
1850 | 4.68 KVM_SET_ONE_REG | |
1851 | ||
1852 | Capability: KVM_CAP_ONE_REG | |
1853 | Architectures: all | |
1854 | Type: vcpu ioctl | |
1855 | Parameters: struct kvm_one_reg (in) | |
1856 | Returns: 0 on success, negative value on failure | |
1857 | ||
1858 | struct kvm_one_reg { | |
1859 | __u64 id; | |
1860 | __u64 addr; | |
1861 | }; | |
1862 | ||
1863 | Using this ioctl, a single vcpu register can be set to a specific value | |
1864 | defined by user space with the passed in struct kvm_one_reg, where id | |
1865 | refers to the register identifier as described below and addr is a pointer | |
1866 | to a variable with the respective size. There can be architecture agnostic | |
1867 | and architecture specific registers. Each have their own range of operation | |
1868 | and their own constants and width. To keep track of the implemented | |
1869 | registers, find a list below: | |
1870 | ||
1871 | Arch | Register | Width (bits) | |
1872 | | | | |
1873 | PPC | KVM_REG_PPC_HIOR | 64 | |
1874 | PPC | KVM_REG_PPC_IAC1 | 64 | |
1875 | PPC | KVM_REG_PPC_IAC2 | 64 | |
1876 | PPC | KVM_REG_PPC_IAC3 | 64 | |
1877 | PPC | KVM_REG_PPC_IAC4 | 64 | |
1878 | PPC | KVM_REG_PPC_DAC1 | 64 | |
1879 | PPC | KVM_REG_PPC_DAC2 | 64 | |
1880 | PPC | KVM_REG_PPC_DABR | 64 | |
1881 | PPC | KVM_REG_PPC_DSCR | 64 | |
1882 | PPC | KVM_REG_PPC_PURR | 64 | |
1883 | PPC | KVM_REG_PPC_SPURR | 64 | |
1884 | PPC | KVM_REG_PPC_DAR | 64 | |
1885 | PPC | KVM_REG_PPC_DSISR | 32 | |
1886 | PPC | KVM_REG_PPC_AMR | 64 | |
1887 | PPC | KVM_REG_PPC_UAMOR | 64 | |
1888 | PPC | KVM_REG_PPC_MMCR0 | 64 | |
1889 | PPC | KVM_REG_PPC_MMCR1 | 64 | |
1890 | PPC | KVM_REG_PPC_MMCRA | 64 | |
1891 | PPC | KVM_REG_PPC_MMCR2 | 64 | |
1892 | PPC | KVM_REG_PPC_MMCRS | 64 | |
1893 | PPC | KVM_REG_PPC_SIAR | 64 | |
1894 | PPC | KVM_REG_PPC_SDAR | 64 | |
1895 | PPC | KVM_REG_PPC_SIER | 64 | |
1896 | PPC | KVM_REG_PPC_PMC1 | 32 | |
1897 | PPC | KVM_REG_PPC_PMC2 | 32 | |
1898 | PPC | KVM_REG_PPC_PMC3 | 32 | |
1899 | PPC | KVM_REG_PPC_PMC4 | 32 | |
1900 | PPC | KVM_REG_PPC_PMC5 | 32 | |
1901 | PPC | KVM_REG_PPC_PMC6 | 32 | |
1902 | PPC | KVM_REG_PPC_PMC7 | 32 | |
1903 | PPC | KVM_REG_PPC_PMC8 | 32 | |
1904 | PPC | KVM_REG_PPC_FPR0 | 64 | |
1905 | ... | |
1906 | PPC | KVM_REG_PPC_FPR31 | 64 | |
1907 | PPC | KVM_REG_PPC_VR0 | 128 | |
1908 | ... | |
1909 | PPC | KVM_REG_PPC_VR31 | 128 | |
1910 | PPC | KVM_REG_PPC_VSR0 | 128 | |
1911 | ... | |
1912 | PPC | KVM_REG_PPC_VSR31 | 128 | |
1913 | PPC | KVM_REG_PPC_FPSCR | 64 | |
1914 | PPC | KVM_REG_PPC_VSCR | 32 | |
1915 | PPC | KVM_REG_PPC_VPA_ADDR | 64 | |
1916 | PPC | KVM_REG_PPC_VPA_SLB | 128 | |
1917 | PPC | KVM_REG_PPC_VPA_DTL | 128 | |
1918 | PPC | KVM_REG_PPC_EPCR | 32 | |
1919 | PPC | KVM_REG_PPC_EPR | 32 | |
1920 | PPC | KVM_REG_PPC_TCR | 32 | |
1921 | PPC | KVM_REG_PPC_TSR | 32 | |
1922 | PPC | KVM_REG_PPC_OR_TSR | 32 | |
1923 | PPC | KVM_REG_PPC_CLEAR_TSR | 32 | |
1924 | PPC | KVM_REG_PPC_MAS0 | 32 | |
1925 | PPC | KVM_REG_PPC_MAS1 | 32 | |
1926 | PPC | KVM_REG_PPC_MAS2 | 64 | |
1927 | PPC | KVM_REG_PPC_MAS7_3 | 64 | |
1928 | PPC | KVM_REG_PPC_MAS4 | 32 | |
1929 | PPC | KVM_REG_PPC_MAS6 | 32 | |
1930 | PPC | KVM_REG_PPC_MMUCFG | 32 | |
1931 | PPC | KVM_REG_PPC_TLB0CFG | 32 | |
1932 | PPC | KVM_REG_PPC_TLB1CFG | 32 | |
1933 | PPC | KVM_REG_PPC_TLB2CFG | 32 | |
1934 | PPC | KVM_REG_PPC_TLB3CFG | 32 | |
1935 | PPC | KVM_REG_PPC_TLB0PS | 32 | |
1936 | PPC | KVM_REG_PPC_TLB1PS | 32 | |
1937 | PPC | KVM_REG_PPC_TLB2PS | 32 | |
1938 | PPC | KVM_REG_PPC_TLB3PS | 32 | |
1939 | PPC | KVM_REG_PPC_EPTCFG | 32 | |
1940 | PPC | KVM_REG_PPC_ICP_STATE | 64 | |
1941 | PPC | KVM_REG_PPC_TB_OFFSET | 64 | |
1942 | PPC | KVM_REG_PPC_SPMC1 | 32 | |
1943 | PPC | KVM_REG_PPC_SPMC2 | 32 | |
1944 | PPC | KVM_REG_PPC_IAMR | 64 | |
1945 | PPC | KVM_REG_PPC_TFHAR | 64 | |
1946 | PPC | KVM_REG_PPC_TFIAR | 64 | |
1947 | PPC | KVM_REG_PPC_TEXASR | 64 | |
1948 | PPC | KVM_REG_PPC_FSCR | 64 | |
1949 | PPC | KVM_REG_PPC_PSPB | 32 | |
1950 | PPC | KVM_REG_PPC_EBBHR | 64 | |
1951 | PPC | KVM_REG_PPC_EBBRR | 64 | |
1952 | PPC | KVM_REG_PPC_BESCR | 64 | |
1953 | PPC | KVM_REG_PPC_TAR | 64 | |
1954 | PPC | KVM_REG_PPC_DPDES | 64 | |
1955 | PPC | KVM_REG_PPC_DAWR | 64 | |
1956 | PPC | KVM_REG_PPC_DAWRX | 64 | |
1957 | PPC | KVM_REG_PPC_CIABR | 64 | |
1958 | PPC | KVM_REG_PPC_IC | 64 | |
1959 | PPC | KVM_REG_PPC_VTB | 64 | |
1960 | PPC | KVM_REG_PPC_CSIGR | 64 | |
1961 | PPC | KVM_REG_PPC_TACR | 64 | |
1962 | PPC | KVM_REG_PPC_TCSCR | 64 | |
1963 | PPC | KVM_REG_PPC_PID | 64 | |
1964 | PPC | KVM_REG_PPC_ACOP | 64 | |
1965 | PPC | KVM_REG_PPC_VRSAVE | 32 | |
1966 | PPC | KVM_REG_PPC_LPCR | 32 | |
1967 | PPC | KVM_REG_PPC_LPCR_64 | 64 | |
1968 | PPC | KVM_REG_PPC_PPR | 64 | |
1969 | PPC | KVM_REG_PPC_ARCH_COMPAT | 32 | |
1970 | PPC | KVM_REG_PPC_DABRX | 32 | |
1971 | PPC | KVM_REG_PPC_WORT | 64 | |
1972 | PPC | KVM_REG_PPC_SPRG9 | 64 | |
1973 | PPC | KVM_REG_PPC_DBSR | 32 | |
1974 | PPC | KVM_REG_PPC_TIDR | 64 | |
1975 | PPC | KVM_REG_PPC_PSSCR | 64 | |
1976 | PPC | KVM_REG_PPC_DEC_EXPIRY | 64 | |
1977 | PPC | KVM_REG_PPC_PTCR | 64 | |
1978 | PPC | KVM_REG_PPC_TM_GPR0 | 64 | |
1979 | ... | |
1980 | PPC | KVM_REG_PPC_TM_GPR31 | 64 | |
1981 | PPC | KVM_REG_PPC_TM_VSR0 | 128 | |
1982 | ... | |
1983 | PPC | KVM_REG_PPC_TM_VSR63 | 128 | |
1984 | PPC | KVM_REG_PPC_TM_CR | 64 | |
1985 | PPC | KVM_REG_PPC_TM_LR | 64 | |
1986 | PPC | KVM_REG_PPC_TM_CTR | 64 | |
1987 | PPC | KVM_REG_PPC_TM_FPSCR | 64 | |
1988 | PPC | KVM_REG_PPC_TM_AMR | 64 | |
1989 | PPC | KVM_REG_PPC_TM_PPR | 64 | |
1990 | PPC | KVM_REG_PPC_TM_VRSAVE | 64 | |
1991 | PPC | KVM_REG_PPC_TM_VSCR | 32 | |
1992 | PPC | KVM_REG_PPC_TM_DSCR | 64 | |
1993 | PPC | KVM_REG_PPC_TM_TAR | 64 | |
1994 | PPC | KVM_REG_PPC_TM_XER | 64 | |
1995 | | | | |
1996 | MIPS | KVM_REG_MIPS_R0 | 64 | |
1997 | ... | |
1998 | MIPS | KVM_REG_MIPS_R31 | 64 | |
1999 | MIPS | KVM_REG_MIPS_HI | 64 | |
2000 | MIPS | KVM_REG_MIPS_LO | 64 | |
2001 | MIPS | KVM_REG_MIPS_PC | 64 | |
2002 | MIPS | KVM_REG_MIPS_CP0_INDEX | 32 | |
2003 | MIPS | KVM_REG_MIPS_CP0_ENTRYLO0 | 64 | |
2004 | MIPS | KVM_REG_MIPS_CP0_ENTRYLO1 | 64 | |
2005 | MIPS | KVM_REG_MIPS_CP0_CONTEXT | 64 | |
2006 | MIPS | KVM_REG_MIPS_CP0_CONTEXTCONFIG| 32 | |
2007 | MIPS | KVM_REG_MIPS_CP0_USERLOCAL | 64 | |
2008 | MIPS | KVM_REG_MIPS_CP0_XCONTEXTCONFIG| 64 | |
2009 | MIPS | KVM_REG_MIPS_CP0_PAGEMASK | 32 | |
2010 | MIPS | KVM_REG_MIPS_CP0_PAGEGRAIN | 32 | |
2011 | MIPS | KVM_REG_MIPS_CP0_SEGCTL0 | 64 | |
2012 | MIPS | KVM_REG_MIPS_CP0_SEGCTL1 | 64 | |
2013 | MIPS | KVM_REG_MIPS_CP0_SEGCTL2 | 64 | |
2014 | MIPS | KVM_REG_MIPS_CP0_PWBASE | 64 | |
2015 | MIPS | KVM_REG_MIPS_CP0_PWFIELD | 64 | |
2016 | MIPS | KVM_REG_MIPS_CP0_PWSIZE | 64 | |
2017 | MIPS | KVM_REG_MIPS_CP0_WIRED | 32 | |
2018 | MIPS | KVM_REG_MIPS_CP0_PWCTL | 32 | |
2019 | MIPS | KVM_REG_MIPS_CP0_HWRENA | 32 | |
2020 | MIPS | KVM_REG_MIPS_CP0_BADVADDR | 64 | |
2021 | MIPS | KVM_REG_MIPS_CP0_BADINSTR | 32 | |
2022 | MIPS | KVM_REG_MIPS_CP0_BADINSTRP | 32 | |
2023 | MIPS | KVM_REG_MIPS_CP0_COUNT | 32 | |
2024 | MIPS | KVM_REG_MIPS_CP0_ENTRYHI | 64 | |
2025 | MIPS | KVM_REG_MIPS_CP0_COMPARE | 32 | |
2026 | MIPS | KVM_REG_MIPS_CP0_STATUS | 32 | |
2027 | MIPS | KVM_REG_MIPS_CP0_INTCTL | 32 | |
2028 | MIPS | KVM_REG_MIPS_CP0_CAUSE | 32 | |
2029 | MIPS | KVM_REG_MIPS_CP0_EPC | 64 | |
2030 | MIPS | KVM_REG_MIPS_CP0_PRID | 32 | |
2031 | MIPS | KVM_REG_MIPS_CP0_EBASE | 64 | |
2032 | MIPS | KVM_REG_MIPS_CP0_CONFIG | 32 | |
2033 | MIPS | KVM_REG_MIPS_CP0_CONFIG1 | 32 | |
2034 | MIPS | KVM_REG_MIPS_CP0_CONFIG2 | 32 | |
2035 | MIPS | KVM_REG_MIPS_CP0_CONFIG3 | 32 | |
2036 | MIPS | KVM_REG_MIPS_CP0_CONFIG4 | 32 | |
2037 | MIPS | KVM_REG_MIPS_CP0_CONFIG5 | 32 | |
2038 | MIPS | KVM_REG_MIPS_CP0_CONFIG7 | 32 | |
2039 | MIPS | KVM_REG_MIPS_CP0_XCONTEXT | 64 | |
2040 | MIPS | KVM_REG_MIPS_CP0_ERROREPC | 64 | |
2041 | MIPS | KVM_REG_MIPS_CP0_KSCRATCH1 | 64 | |
2042 | MIPS | KVM_REG_MIPS_CP0_KSCRATCH2 | 64 | |
2043 | MIPS | KVM_REG_MIPS_CP0_KSCRATCH3 | 64 | |
2044 | MIPS | KVM_REG_MIPS_CP0_KSCRATCH4 | 64 | |
2045 | MIPS | KVM_REG_MIPS_CP0_KSCRATCH5 | 64 | |
2046 | MIPS | KVM_REG_MIPS_CP0_KSCRATCH6 | 64 | |
2047 | MIPS | KVM_REG_MIPS_CP0_MAAR(0..63) | 64 | |
2048 | MIPS | KVM_REG_MIPS_COUNT_CTL | 64 | |
2049 | MIPS | KVM_REG_MIPS_COUNT_RESUME | 64 | |
2050 | MIPS | KVM_REG_MIPS_COUNT_HZ | 64 | |
2051 | MIPS | KVM_REG_MIPS_FPR_32(0..31) | 32 | |
2052 | MIPS | KVM_REG_MIPS_FPR_64(0..31) | 64 | |
2053 | MIPS | KVM_REG_MIPS_VEC_128(0..31) | 128 | |
2054 | MIPS | KVM_REG_MIPS_FCR_IR | 32 | |
2055 | MIPS | KVM_REG_MIPS_FCR_CSR | 32 | |
2056 | MIPS | KVM_REG_MIPS_MSA_IR | 32 | |
2057 | MIPS | KVM_REG_MIPS_MSA_CSR | 32 | |
2058 | ||
2059 | ARM registers are mapped using the lower 32 bits. The upper 16 of that | |
2060 | is the register group type, or coprocessor number: | |
2061 | ||
2062 | ARM core registers have the following id bit patterns: | |
2063 | 0x4020 0000 0010 <index into the kvm_regs struct:16> | |
2064 | ||
2065 | ARM 32-bit CP15 registers have the following id bit patterns: | |
2066 | 0x4020 0000 000F <zero:1> <crn:4> <crm:4> <opc1:4> <opc2:3> | |
2067 | ||
2068 | ARM 64-bit CP15 registers have the following id bit patterns: | |
2069 | 0x4030 0000 000F <zero:1> <zero:4> <crm:4> <opc1:4> <zero:3> | |
2070 | ||
2071 | ARM CCSIDR registers are demultiplexed by CSSELR value: | |
2072 | 0x4020 0000 0011 00 <csselr:8> | |
2073 | ||
2074 | ARM 32-bit VFP control registers have the following id bit patterns: | |
2075 | 0x4020 0000 0012 1 <regno:12> | |
2076 | ||
2077 | ARM 64-bit FP registers have the following id bit patterns: | |
2078 | 0x4030 0000 0012 0 <regno:12> | |
2079 | ||
2080 | ARM firmware pseudo-registers have the following bit pattern: | |
2081 | 0x4030 0000 0014 <regno:16> | |
2082 | ||
2083 | ||
2084 | arm64 registers are mapped using the lower 32 bits. The upper 16 of | |
2085 | that is the register group type, or coprocessor number: | |
2086 | ||
2087 | arm64 core/FP-SIMD registers have the following id bit patterns. Note | |
2088 | that the size of the access is variable, as the kvm_regs structure | |
2089 | contains elements ranging from 32 to 128 bits. The index is a 32bit | |
2090 | value in the kvm_regs structure seen as a 32bit array. | |
2091 | 0x60x0 0000 0010 <index into the kvm_regs struct:16> | |
2092 | ||
2093 | arm64 CCSIDR registers are demultiplexed by CSSELR value: | |
2094 | 0x6020 0000 0011 00 <csselr:8> | |
2095 | ||
2096 | arm64 system registers have the following id bit patterns: | |
2097 | 0x6030 0000 0013 <op0:2> <op1:3> <crn:4> <crm:4> <op2:3> | |
2098 | ||
2099 | arm64 firmware pseudo-registers have the following bit pattern: | |
2100 | 0x6030 0000 0014 <regno:16> | |
2101 | ||
2102 | ||
2103 | MIPS registers are mapped using the lower 32 bits. The upper 16 of that is | |
2104 | the register group type: | |
2105 | ||
2106 | MIPS core registers (see above) have the following id bit patterns: | |
2107 | 0x7030 0000 0000 <reg:16> | |
2108 | ||
2109 | MIPS CP0 registers (see KVM_REG_MIPS_CP0_* above) have the following id bit | |
2110 | patterns depending on whether they're 32-bit or 64-bit registers: | |
2111 | 0x7020 0000 0001 00 <reg:5> <sel:3> (32-bit) | |
2112 | 0x7030 0000 0001 00 <reg:5> <sel:3> (64-bit) | |
2113 | ||
2114 | Note: KVM_REG_MIPS_CP0_ENTRYLO0 and KVM_REG_MIPS_CP0_ENTRYLO1 are the MIPS64 | |
2115 | versions of the EntryLo registers regardless of the word size of the host | |
2116 | hardware, host kernel, guest, and whether XPA is present in the guest, i.e. | |
2117 | with the RI and XI bits (if they exist) in bits 63 and 62 respectively, and | |
2118 | the PFNX field starting at bit 30. | |
2119 | ||
2120 | MIPS MAARs (see KVM_REG_MIPS_CP0_MAAR(*) above) have the following id bit | |
2121 | patterns: | |
2122 | 0x7030 0000 0001 01 <reg:8> | |
2123 | ||
2124 | MIPS KVM control registers (see above) have the following id bit patterns: | |
2125 | 0x7030 0000 0002 <reg:16> | |
2126 | ||
2127 | MIPS FPU registers (see KVM_REG_MIPS_FPR_{32,64}() above) have the following | |
2128 | id bit patterns depending on the size of the register being accessed. They are | |
2129 | always accessed according to the current guest FPU mode (Status.FR and | |
2130 | Config5.FRE), i.e. as the guest would see them, and they become unpredictable | |
2131 | if the guest FPU mode is changed. MIPS SIMD Architecture (MSA) vector | |
2132 | registers (see KVM_REG_MIPS_VEC_128() above) have similar patterns as they | |
2133 | overlap the FPU registers: | |
2134 | 0x7020 0000 0003 00 <0:3> <reg:5> (32-bit FPU registers) | |
2135 | 0x7030 0000 0003 00 <0:3> <reg:5> (64-bit FPU registers) | |
2136 | 0x7040 0000 0003 00 <0:3> <reg:5> (128-bit MSA vector registers) | |
2137 | ||
2138 | MIPS FPU control registers (see KVM_REG_MIPS_FCR_{IR,CSR} above) have the | |
2139 | following id bit patterns: | |
2140 | 0x7020 0000 0003 01 <0:3> <reg:5> | |
2141 | ||
2142 | MIPS MSA control registers (see KVM_REG_MIPS_MSA_{IR,CSR} above) have the | |
2143 | following id bit patterns: | |
2144 | 0x7020 0000 0003 02 <0:3> <reg:5> | |
2145 | ||
2146 | ||
2147 | 4.69 KVM_GET_ONE_REG | |
2148 | ||
2149 | Capability: KVM_CAP_ONE_REG | |
2150 | Architectures: all | |
2151 | Type: vcpu ioctl | |
2152 | Parameters: struct kvm_one_reg (in and out) | |
2153 | Returns: 0 on success, negative value on failure | |
2154 | ||
2155 | This ioctl allows to receive the value of a single register implemented | |
2156 | in a vcpu. The register to read is indicated by the "id" field of the | |
2157 | kvm_one_reg struct passed in. On success, the register value can be found | |
2158 | at the memory location pointed to by "addr". | |
2159 | ||
2160 | The list of registers accessible using this interface is identical to the | |
2161 | list in 4.68. | |
2162 | ||
2163 | ||
2164 | 4.70 KVM_KVMCLOCK_CTRL | |
2165 | ||
2166 | Capability: KVM_CAP_KVMCLOCK_CTRL | |
2167 | Architectures: Any that implement pvclocks (currently x86 only) | |
2168 | Type: vcpu ioctl | |
2169 | Parameters: None | |
2170 | Returns: 0 on success, -1 on error | |
2171 | ||
2172 | This signals to the host kernel that the specified guest is being paused by | |
2173 | userspace. The host will set a flag in the pvclock structure that is checked | |
2174 | from the soft lockup watchdog. The flag is part of the pvclock structure that | |
2175 | is shared between guest and host, specifically the second bit of the flags | |
2176 | field of the pvclock_vcpu_time_info structure. It will be set exclusively by | |
2177 | the host and read/cleared exclusively by the guest. The guest operation of | |
2178 | checking and clearing the flag must an atomic operation so | |
2179 | load-link/store-conditional, or equivalent must be used. There are two cases | |
2180 | where the guest will clear the flag: when the soft lockup watchdog timer resets | |
2181 | itself or when a soft lockup is detected. This ioctl can be called any time | |
2182 | after pausing the vcpu, but before it is resumed. | |
2183 | ||
2184 | ||
2185 | 4.71 KVM_SIGNAL_MSI | |
2186 | ||
2187 | Capability: KVM_CAP_SIGNAL_MSI | |
2188 | Architectures: x86 arm arm64 | |
2189 | Type: vm ioctl | |
2190 | Parameters: struct kvm_msi (in) | |
2191 | Returns: >0 on delivery, 0 if guest blocked the MSI, and -1 on error | |
2192 | ||
2193 | Directly inject a MSI message. Only valid with in-kernel irqchip that handles | |
2194 | MSI messages. | |
2195 | ||
2196 | struct kvm_msi { | |
2197 | __u32 address_lo; | |
2198 | __u32 address_hi; | |
2199 | __u32 data; | |
2200 | __u32 flags; | |
2201 | __u32 devid; | |
2202 | __u8 pad[12]; | |
2203 | }; | |
2204 | ||
2205 | flags: KVM_MSI_VALID_DEVID: devid contains a valid value. The per-VM | |
2206 | KVM_CAP_MSI_DEVID capability advertises the requirement to provide | |
2207 | the device ID. If this capability is not available, userspace | |
2208 | should never set the KVM_MSI_VALID_DEVID flag as the ioctl might fail. | |
2209 | ||
2210 | If KVM_MSI_VALID_DEVID is set, devid contains a unique device identifier | |
2211 | for the device that wrote the MSI message. For PCI, this is usually a | |
2212 | BFD identifier in the lower 16 bits. | |
2213 | ||
2214 | On x86, address_hi is ignored unless the KVM_X2APIC_API_USE_32BIT_IDS | |
2215 | feature of KVM_CAP_X2APIC_API capability is enabled. If it is enabled, | |
2216 | address_hi bits 31-8 provide bits 31-8 of the destination id. Bits 7-0 of | |
2217 | address_hi must be zero. | |
2218 | ||
2219 | ||
2220 | 4.71 KVM_CREATE_PIT2 | |
2221 | ||
2222 | Capability: KVM_CAP_PIT2 | |
2223 | Architectures: x86 | |
2224 | Type: vm ioctl | |
2225 | Parameters: struct kvm_pit_config (in) | |
2226 | Returns: 0 on success, -1 on error | |
2227 | ||
2228 | Creates an in-kernel device model for the i8254 PIT. This call is only valid | |
2229 | after enabling in-kernel irqchip support via KVM_CREATE_IRQCHIP. The following | |
2230 | parameters have to be passed: | |
2231 | ||
2232 | struct kvm_pit_config { | |
2233 | __u32 flags; | |
2234 | __u32 pad[15]; | |
2235 | }; | |
2236 | ||
2237 | Valid flags are: | |
2238 | ||
2239 | #define KVM_PIT_SPEAKER_DUMMY 1 /* emulate speaker port stub */ | |
2240 | ||
2241 | PIT timer interrupts may use a per-VM kernel thread for injection. If it | |
2242 | exists, this thread will have a name of the following pattern: | |
2243 | ||
2244 | kvm-pit/<owner-process-pid> | |
2245 | ||
2246 | When running a guest with elevated priorities, the scheduling parameters of | |
2247 | this thread may have to be adjusted accordingly. | |
2248 | ||
2249 | This IOCTL replaces the obsolete KVM_CREATE_PIT. | |
2250 | ||
2251 | ||
2252 | 4.72 KVM_GET_PIT2 | |
2253 | ||
2254 | Capability: KVM_CAP_PIT_STATE2 | |
2255 | Architectures: x86 | |
2256 | Type: vm ioctl | |
2257 | Parameters: struct kvm_pit_state2 (out) | |
2258 | Returns: 0 on success, -1 on error | |
2259 | ||
2260 | Retrieves the state of the in-kernel PIT model. Only valid after | |
2261 | KVM_CREATE_PIT2. The state is returned in the following structure: | |
2262 | ||
2263 | struct kvm_pit_state2 { | |
2264 | struct kvm_pit_channel_state channels[3]; | |
2265 | __u32 flags; | |
2266 | __u32 reserved[9]; | |
2267 | }; | |
2268 | ||
2269 | Valid flags are: | |
2270 | ||
2271 | /* disable PIT in HPET legacy mode */ | |
2272 | #define KVM_PIT_FLAGS_HPET_LEGACY 0x00000001 | |
2273 | ||
2274 | This IOCTL replaces the obsolete KVM_GET_PIT. | |
2275 | ||
2276 | ||
2277 | 4.73 KVM_SET_PIT2 | |
2278 | ||
2279 | Capability: KVM_CAP_PIT_STATE2 | |
2280 | Architectures: x86 | |
2281 | Type: vm ioctl | |
2282 | Parameters: struct kvm_pit_state2 (in) | |
2283 | Returns: 0 on success, -1 on error | |
2284 | ||
2285 | Sets the state of the in-kernel PIT model. Only valid after KVM_CREATE_PIT2. | |
2286 | See KVM_GET_PIT2 for details on struct kvm_pit_state2. | |
2287 | ||
2288 | This IOCTL replaces the obsolete KVM_SET_PIT. | |
2289 | ||
2290 | ||
2291 | 4.74 KVM_PPC_GET_SMMU_INFO | |
2292 | ||
2293 | Capability: KVM_CAP_PPC_GET_SMMU_INFO | |
2294 | Architectures: powerpc | |
2295 | Type: vm ioctl | |
2296 | Parameters: None | |
2297 | Returns: 0 on success, -1 on error | |
2298 | ||
2299 | This populates and returns a structure describing the features of | |
2300 | the "Server" class MMU emulation supported by KVM. | |
2301 | This can in turn be used by userspace to generate the appropriate | |
2302 | device-tree properties for the guest operating system. | |
2303 | ||
2304 | The structure contains some global information, followed by an | |
2305 | array of supported segment page sizes: | |
2306 | ||
2307 | struct kvm_ppc_smmu_info { | |
2308 | __u64 flags; | |
2309 | __u32 slb_size; | |
2310 | __u32 pad; | |
2311 | struct kvm_ppc_one_seg_page_size sps[KVM_PPC_PAGE_SIZES_MAX_SZ]; | |
2312 | }; | |
2313 | ||
2314 | The supported flags are: | |
2315 | ||
2316 | - KVM_PPC_PAGE_SIZES_REAL: | |
2317 | When that flag is set, guest page sizes must "fit" the backing | |
2318 | store page sizes. When not set, any page size in the list can | |
2319 | be used regardless of how they are backed by userspace. | |
2320 | ||
2321 | - KVM_PPC_1T_SEGMENTS | |
2322 | The emulated MMU supports 1T segments in addition to the | |
2323 | standard 256M ones. | |
2324 | ||
2325 | - KVM_PPC_NO_HASH | |
2326 | This flag indicates that HPT guests are not supported by KVM, | |
2327 | thus all guests must use radix MMU mode. | |
2328 | ||
2329 | The "slb_size" field indicates how many SLB entries are supported | |
2330 | ||
2331 | The "sps" array contains 8 entries indicating the supported base | |
2332 | page sizes for a segment in increasing order. Each entry is defined | |
2333 | as follow: | |
2334 | ||
2335 | struct kvm_ppc_one_seg_page_size { | |
2336 | __u32 page_shift; /* Base page shift of segment (or 0) */ | |
2337 | __u32 slb_enc; /* SLB encoding for BookS */ | |
2338 | struct kvm_ppc_one_page_size enc[KVM_PPC_PAGE_SIZES_MAX_SZ]; | |
2339 | }; | |
2340 | ||
2341 | An entry with a "page_shift" of 0 is unused. Because the array is | |
2342 | organized in increasing order, a lookup can stop when encoutering | |
2343 | such an entry. | |
2344 | ||
2345 | The "slb_enc" field provides the encoding to use in the SLB for the | |
2346 | page size. The bits are in positions such as the value can directly | |
2347 | be OR'ed into the "vsid" argument of the slbmte instruction. | |
2348 | ||
2349 | The "enc" array is a list which for each of those segment base page | |
2350 | size provides the list of supported actual page sizes (which can be | |
2351 | only larger or equal to the base page size), along with the | |
2352 | corresponding encoding in the hash PTE. Similarly, the array is | |
2353 | 8 entries sorted by increasing sizes and an entry with a "0" shift | |
2354 | is an empty entry and a terminator: | |
2355 | ||
2356 | struct kvm_ppc_one_page_size { | |
2357 | __u32 page_shift; /* Page shift (or 0) */ | |
2358 | __u32 pte_enc; /* Encoding in the HPTE (>>12) */ | |
2359 | }; | |
2360 | ||
2361 | The "pte_enc" field provides a value that can OR'ed into the hash | |
2362 | PTE's RPN field (ie, it needs to be shifted left by 12 to OR it | |
2363 | into the hash PTE second double word). | |
2364 | ||
2365 | 4.75 KVM_IRQFD | |
2366 | ||
2367 | Capability: KVM_CAP_IRQFD | |
2368 | Architectures: x86 s390 arm arm64 | |
2369 | Type: vm ioctl | |
2370 | Parameters: struct kvm_irqfd (in) | |
2371 | Returns: 0 on success, -1 on error | |
2372 | ||
2373 | Allows setting an eventfd to directly trigger a guest interrupt. | |
2374 | kvm_irqfd.fd specifies the file descriptor to use as the eventfd and | |
2375 | kvm_irqfd.gsi specifies the irqchip pin toggled by this event. When | |
2376 | an event is triggered on the eventfd, an interrupt is injected into | |
2377 | the guest using the specified gsi pin. The irqfd is removed using | |
2378 | the KVM_IRQFD_FLAG_DEASSIGN flag, specifying both kvm_irqfd.fd | |
2379 | and kvm_irqfd.gsi. | |
2380 | ||
2381 | With KVM_CAP_IRQFD_RESAMPLE, KVM_IRQFD supports a de-assert and notify | |
2382 | mechanism allowing emulation of level-triggered, irqfd-based | |
2383 | interrupts. When KVM_IRQFD_FLAG_RESAMPLE is set the user must pass an | |
2384 | additional eventfd in the kvm_irqfd.resamplefd field. When operating | |
2385 | in resample mode, posting of an interrupt through kvm_irq.fd asserts | |
2386 | the specified gsi in the irqchip. When the irqchip is resampled, such | |
2387 | as from an EOI, the gsi is de-asserted and the user is notified via | |
2388 | kvm_irqfd.resamplefd. It is the user's responsibility to re-queue | |
2389 | the interrupt if the device making use of it still requires service. | |
2390 | Note that closing the resamplefd is not sufficient to disable the | |
2391 | irqfd. The KVM_IRQFD_FLAG_RESAMPLE is only necessary on assignment | |
2392 | and need not be specified with KVM_IRQFD_FLAG_DEASSIGN. | |
2393 | ||
2394 | On arm/arm64, gsi routing being supported, the following can happen: | |
2395 | - in case no routing entry is associated to this gsi, injection fails | |
2396 | - in case the gsi is associated to an irqchip routing entry, | |
2397 | irqchip.pin + 32 corresponds to the injected SPI ID. | |
2398 | - in case the gsi is associated to an MSI routing entry, the MSI | |
2399 | message and device ID are translated into an LPI (support restricted | |
2400 | to GICv3 ITS in-kernel emulation). | |
2401 | ||
2402 | 4.76 KVM_PPC_ALLOCATE_HTAB | |
2403 | ||
2404 | Capability: KVM_CAP_PPC_ALLOC_HTAB | |
2405 | Architectures: powerpc | |
2406 | Type: vm ioctl | |
2407 | Parameters: Pointer to u32 containing hash table order (in/out) | |
2408 | Returns: 0 on success, -1 on error | |
2409 | ||
2410 | This requests the host kernel to allocate an MMU hash table for a | |
2411 | guest using the PAPR paravirtualization interface. This only does | |
2412 | anything if the kernel is configured to use the Book 3S HV style of | |
2413 | virtualization. Otherwise the capability doesn't exist and the ioctl | |
2414 | returns an ENOTTY error. The rest of this description assumes Book 3S | |
2415 | HV. | |
2416 | ||
2417 | There must be no vcpus running when this ioctl is called; if there | |
2418 | are, it will do nothing and return an EBUSY error. | |
2419 | ||
2420 | The parameter is a pointer to a 32-bit unsigned integer variable | |
2421 | containing the order (log base 2) of the desired size of the hash | |
2422 | table, which must be between 18 and 46. On successful return from the | |
2423 | ioctl, the value will not be changed by the kernel. | |
2424 | ||
2425 | If no hash table has been allocated when any vcpu is asked to run | |
2426 | (with the KVM_RUN ioctl), the host kernel will allocate a | |
2427 | default-sized hash table (16 MB). | |
2428 | ||
2429 | If this ioctl is called when a hash table has already been allocated, | |
2430 | with a different order from the existing hash table, the existing hash | |
2431 | table will be freed and a new one allocated. If this is ioctl is | |
2432 | called when a hash table has already been allocated of the same order | |
2433 | as specified, the kernel will clear out the existing hash table (zero | |
2434 | all HPTEs). In either case, if the guest is using the virtualized | |
2435 | real-mode area (VRMA) facility, the kernel will re-create the VMRA | |
2436 | HPTEs on the next KVM_RUN of any vcpu. | |
2437 | ||
2438 | 4.77 KVM_S390_INTERRUPT | |
2439 | ||
2440 | Capability: basic | |
2441 | Architectures: s390 | |
2442 | Type: vm ioctl, vcpu ioctl | |
2443 | Parameters: struct kvm_s390_interrupt (in) | |
2444 | Returns: 0 on success, -1 on error | |
2445 | ||
2446 | Allows to inject an interrupt to the guest. Interrupts can be floating | |
2447 | (vm ioctl) or per cpu (vcpu ioctl), depending on the interrupt type. | |
2448 | ||
2449 | Interrupt parameters are passed via kvm_s390_interrupt: | |
2450 | ||
2451 | struct kvm_s390_interrupt { | |
2452 | __u32 type; | |
2453 | __u32 parm; | |
2454 | __u64 parm64; | |
2455 | }; | |
2456 | ||
2457 | type can be one of the following: | |
2458 | ||
2459 | KVM_S390_SIGP_STOP (vcpu) - sigp stop; optional flags in parm | |
2460 | KVM_S390_PROGRAM_INT (vcpu) - program check; code in parm | |
2461 | KVM_S390_SIGP_SET_PREFIX (vcpu) - sigp set prefix; prefix address in parm | |
2462 | KVM_S390_RESTART (vcpu) - restart | |
2463 | KVM_S390_INT_CLOCK_COMP (vcpu) - clock comparator interrupt | |
2464 | KVM_S390_INT_CPU_TIMER (vcpu) - CPU timer interrupt | |
2465 | KVM_S390_INT_VIRTIO (vm) - virtio external interrupt; external interrupt | |
2466 | parameters in parm and parm64 | |
2467 | KVM_S390_INT_SERVICE (vm) - sclp external interrupt; sclp parameter in parm | |
2468 | KVM_S390_INT_EMERGENCY (vcpu) - sigp emergency; source cpu in parm | |
2469 | KVM_S390_INT_EXTERNAL_CALL (vcpu) - sigp external call; source cpu in parm | |
2470 | KVM_S390_INT_IO(ai,cssid,ssid,schid) (vm) - compound value to indicate an | |
2471 | I/O interrupt (ai - adapter interrupt; cssid,ssid,schid - subchannel); | |
2472 | I/O interruption parameters in parm (subchannel) and parm64 (intparm, | |
2473 | interruption subclass) | |
2474 | KVM_S390_MCHK (vm, vcpu) - machine check interrupt; cr 14 bits in parm, | |
2475 | machine check interrupt code in parm64 (note that | |
2476 | machine checks needing further payload are not | |
2477 | supported by this ioctl) | |
2478 | ||
2479 | Note that the vcpu ioctl is asynchronous to vcpu execution. | |
2480 | ||
2481 | 4.78 KVM_PPC_GET_HTAB_FD | |
2482 | ||
2483 | Capability: KVM_CAP_PPC_HTAB_FD | |
2484 | Architectures: powerpc | |
2485 | Type: vm ioctl | |
2486 | Parameters: Pointer to struct kvm_get_htab_fd (in) | |
2487 | Returns: file descriptor number (>= 0) on success, -1 on error | |
2488 | ||
2489 | This returns a file descriptor that can be used either to read out the | |
2490 | entries in the guest's hashed page table (HPT), or to write entries to | |
2491 | initialize the HPT. The returned fd can only be written to if the | |
2492 | KVM_GET_HTAB_WRITE bit is set in the flags field of the argument, and | |
2493 | can only be read if that bit is clear. The argument struct looks like | |
2494 | this: | |
2495 | ||
2496 | /* For KVM_PPC_GET_HTAB_FD */ | |
2497 | struct kvm_get_htab_fd { | |
2498 | __u64 flags; | |
2499 | __u64 start_index; | |
2500 | __u64 reserved[2]; | |
2501 | }; | |
2502 | ||
2503 | /* Values for kvm_get_htab_fd.flags */ | |
2504 | #define KVM_GET_HTAB_BOLTED_ONLY ((__u64)0x1) | |
2505 | #define KVM_GET_HTAB_WRITE ((__u64)0x2) | |
2506 | ||
2507 | The `start_index' field gives the index in the HPT of the entry at | |
2508 | which to start reading. It is ignored when writing. | |
2509 | ||
2510 | Reads on the fd will initially supply information about all | |
2511 | "interesting" HPT entries. Interesting entries are those with the | |
2512 | bolted bit set, if the KVM_GET_HTAB_BOLTED_ONLY bit is set, otherwise | |
2513 | all entries. When the end of the HPT is reached, the read() will | |
2514 | return. If read() is called again on the fd, it will start again from | |
2515 | the beginning of the HPT, but will only return HPT entries that have | |
2516 | changed since they were last read. | |
2517 | ||
2518 | Data read or written is structured as a header (8 bytes) followed by a | |
2519 | series of valid HPT entries (16 bytes) each. The header indicates how | |
2520 | many valid HPT entries there are and how many invalid entries follow | |
2521 | the valid entries. The invalid entries are not represented explicitly | |
2522 | in the stream. The header format is: | |
2523 | ||
2524 | struct kvm_get_htab_header { | |
2525 | __u32 index; | |
2526 | __u16 n_valid; | |
2527 | __u16 n_invalid; | |
2528 | }; | |
2529 | ||
2530 | Writes to the fd create HPT entries starting at the index given in the | |
2531 | header; first `n_valid' valid entries with contents from the data | |
2532 | written, then `n_invalid' invalid entries, invalidating any previously | |
2533 | valid entries found. | |
2534 | ||
2535 | 4.79 KVM_CREATE_DEVICE | |
2536 | ||
2537 | Capability: KVM_CAP_DEVICE_CTRL | |
2538 | Type: vm ioctl | |
2539 | Parameters: struct kvm_create_device (in/out) | |
2540 | Returns: 0 on success, -1 on error | |
2541 | Errors: | |
2542 | ENODEV: The device type is unknown or unsupported | |
2543 | EEXIST: Device already created, and this type of device may not | |
2544 | be instantiated multiple times | |
2545 | ||
2546 | Other error conditions may be defined by individual device types or | |
2547 | have their standard meanings. | |
2548 | ||
2549 | Creates an emulated device in the kernel. The file descriptor returned | |
2550 | in fd can be used with KVM_SET/GET/HAS_DEVICE_ATTR. | |
2551 | ||
2552 | If the KVM_CREATE_DEVICE_TEST flag is set, only test whether the | |
2553 | device type is supported (not necessarily whether it can be created | |
2554 | in the current vm). | |
2555 | ||
2556 | Individual devices should not define flags. Attributes should be used | |
2557 | for specifying any behavior that is not implied by the device type | |
2558 | number. | |
2559 | ||
2560 | struct kvm_create_device { | |
2561 | __u32 type; /* in: KVM_DEV_TYPE_xxx */ | |
2562 | __u32 fd; /* out: device handle */ | |
2563 | __u32 flags; /* in: KVM_CREATE_DEVICE_xxx */ | |
2564 | }; | |
2565 | ||
2566 | 4.80 KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR | |
2567 | ||
2568 | Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, | |
2569 | KVM_CAP_VCPU_ATTRIBUTES for vcpu device | |
2570 | Type: device ioctl, vm ioctl, vcpu ioctl | |
2571 | Parameters: struct kvm_device_attr | |
2572 | Returns: 0 on success, -1 on error | |
2573 | Errors: | |
2574 | ENXIO: The group or attribute is unknown/unsupported for this device | |
2575 | or hardware support is missing. | |
2576 | EPERM: The attribute cannot (currently) be accessed this way | |
2577 | (e.g. read-only attribute, or attribute that only makes | |
2578 | sense when the device is in a different state) | |
2579 | ||
2580 | Other error conditions may be defined by individual device types. | |
2581 | ||
2582 | Gets/sets a specified piece of device configuration and/or state. The | |
2583 | semantics are device-specific. See individual device documentation in | |
2584 | the "devices" directory. As with ONE_REG, the size of the data | |
2585 | transferred is defined by the particular attribute. | |
2586 | ||
2587 | struct kvm_device_attr { | |
2588 | __u32 flags; /* no flags currently defined */ | |
2589 | __u32 group; /* device-defined */ | |
2590 | __u64 attr; /* group-defined */ | |
2591 | __u64 addr; /* userspace address of attr data */ | |
2592 | }; | |
2593 | ||
2594 | 4.81 KVM_HAS_DEVICE_ATTR | |
2595 | ||
2596 | Capability: KVM_CAP_DEVICE_CTRL, KVM_CAP_VM_ATTRIBUTES for vm device, | |
2597 | KVM_CAP_VCPU_ATTRIBUTES for vcpu device | |
2598 | Type: device ioctl, vm ioctl, vcpu ioctl | |
2599 | Parameters: struct kvm_device_attr | |
2600 | Returns: 0 on success, -1 on error | |
2601 | Errors: | |
2602 | ENXIO: The group or attribute is unknown/unsupported for this device | |
2603 | or hardware support is missing. | |
2604 | ||
2605 | Tests whether a device supports a particular attribute. A successful | |
2606 | return indicates the attribute is implemented. It does not necessarily | |
2607 | indicate that the attribute can be read or written in the device's | |
2608 | current state. "addr" is ignored. | |
2609 | ||
2610 | 4.82 KVM_ARM_VCPU_INIT | |
2611 | ||
2612 | Capability: basic | |
2613 | Architectures: arm, arm64 | |
2614 | Type: vcpu ioctl | |
2615 | Parameters: struct kvm_vcpu_init (in) | |
2616 | Returns: 0 on success; -1 on error | |
2617 | Errors: | |
2618 | Â EINVAL: Â Â Â the target is unknown, or the combination of features is invalid. | |
2619 | Â ENOENT: Â Â Â a features bit specified is unknown. | |
2620 | ||
2621 | This tells KVM what type of CPU to present to the guest, and what | |
2622 | optional features it should have. Â This will cause a reset of the cpu | |
2623 | registers to their initial values. Â If this is not called, KVM_RUN will | |
2624 | return ENOEXEC for that vcpu. | |
2625 | ||
2626 | Note that because some registers reflect machine topology, all vcpus | |
2627 | should be created before this ioctl is invoked. | |
2628 | ||
2629 | Userspace can call this function multiple times for a given vcpu, including | |
2630 | after the vcpu has been run. This will reset the vcpu to its initial | |
2631 | state. All calls to this function after the initial call must use the same | |
2632 | target and same set of feature flags, otherwise EINVAL will be returned. | |
2633 | ||
2634 | Possible features: | |
2635 | - KVM_ARM_VCPU_POWER_OFF: Starts the CPU in a power-off state. | |
2636 | Depends on KVM_CAP_ARM_PSCI. If not set, the CPU will be powered on | |
2637 | and execute guest code when KVM_RUN is called. | |
2638 | - KVM_ARM_VCPU_EL1_32BIT: Starts the CPU in a 32bit mode. | |
2639 | Depends on KVM_CAP_ARM_EL1_32BIT (arm64 only). | |
2640 | - KVM_ARM_VCPU_PSCI_0_2: Emulate PSCI v0.2 (or a future revision | |
2641 | backward compatible with v0.2) for the CPU. | |
2642 | Depends on KVM_CAP_ARM_PSCI_0_2. | |
2643 | - KVM_ARM_VCPU_PMU_V3: Emulate PMUv3 for the CPU. | |
2644 | Depends on KVM_CAP_ARM_PMU_V3. | |
2645 | ||
2646 | ||
2647 | 4.83 KVM_ARM_PREFERRED_TARGET | |
2648 | ||
2649 | Capability: basic | |
2650 | Architectures: arm, arm64 | |
2651 | Type: vm ioctl | |
2652 | Parameters: struct struct kvm_vcpu_init (out) | |
2653 | Returns: 0 on success; -1 on error | |
2654 | Errors: | |
2655 | ENODEV: no preferred target available for the host | |
2656 | ||
2657 | This queries KVM for preferred CPU target type which can be emulated | |
2658 | by KVM on underlying host. | |
2659 | ||
2660 | The ioctl returns struct kvm_vcpu_init instance containing information | |
2661 | about preferred CPU target type and recommended features for it. The | |
2662 | kvm_vcpu_init->features bitmap returned will have feature bits set if | |
2663 | the preferred target recommends setting these features, but this is | |
2664 | not mandatory. | |
2665 | ||
2666 | The information returned by this ioctl can be used to prepare an instance | |
2667 | of struct kvm_vcpu_init for KVM_ARM_VCPU_INIT ioctl which will result in | |
2668 | in VCPU matching underlying host. | |
2669 | ||
2670 | ||
2671 | 4.84 KVM_GET_REG_LIST | |
2672 | ||
2673 | Capability: basic | |
2674 | Architectures: arm, arm64, mips | |
2675 | Type: vcpu ioctl | |
2676 | Parameters: struct kvm_reg_list (in/out) | |
2677 | Returns: 0 on success; -1 on error | |
2678 | Errors: | |
2679 | Â E2BIG: Â Â Â Â the reg index list is too big to fit in the array specified by | |
2680 | Â Â Â Â Â Â Â Â Â Â Â Â the user (the number required will be written into n). | |
2681 | ||
2682 | struct kvm_reg_list { | |
2683 | __u64 n; /* number of registers in reg[] */ | |
2684 | __u64 reg[0]; | |
2685 | }; | |
2686 | ||
2687 | This ioctl returns the guest registers that are supported for the | |
2688 | KVM_GET_ONE_REG/KVM_SET_ONE_REG calls. | |
2689 | ||
2690 | ||
2691 | 4.85 KVM_ARM_SET_DEVICE_ADDR (deprecated) | |
2692 | ||
2693 | Capability: KVM_CAP_ARM_SET_DEVICE_ADDR | |
2694 | Architectures: arm, arm64 | |
2695 | Type: vm ioctl | |
2696 | Parameters: struct kvm_arm_device_address (in) | |
2697 | Returns: 0 on success, -1 on error | |
2698 | Errors: | |
2699 | ENODEV: The device id is unknown | |
2700 | ENXIO: Device not supported on current system | |
2701 | EEXIST: Address already set | |
2702 | E2BIG: Address outside guest physical address space | |
2703 | EBUSY: Address overlaps with other device range | |
2704 | ||
2705 | struct kvm_arm_device_addr { | |
2706 | __u64 id; | |
2707 | __u64 addr; | |
2708 | }; | |
2709 | ||
2710 | Specify a device address in the guest's physical address space where guests | |
2711 | can access emulated or directly exposed devices, which the host kernel needs | |
2712 | to know about. The id field is an architecture specific identifier for a | |
2713 | specific device. | |
2714 | ||
2715 | ARM/arm64 divides the id field into two parts, a device id and an | |
2716 | address type id specific to the individual device. | |
2717 | ||
2718 | Â bits: | 63 ... 32 | 31 ... 16 | 15 ... 0 | | |
2719 | field: | 0x00000000 | device id | addr type id | | |
2720 | ||
2721 | ARM/arm64 currently only require this when using the in-kernel GIC | |
2722 | support for the hardware VGIC features, using KVM_ARM_DEVICE_VGIC_V2 | |
2723 | as the device id. When setting the base address for the guest's | |
2724 | mapping of the VGIC virtual CPU and distributor interface, the ioctl | |
2725 | must be called after calling KVM_CREATE_IRQCHIP, but before calling | |
2726 | KVM_RUN on any of the VCPUs. Calling this ioctl twice for any of the | |
2727 | base addresses will return -EEXIST. | |
2728 | ||
2729 | Note, this IOCTL is deprecated and the more flexible SET/GET_DEVICE_ATTR API | |
2730 | should be used instead. | |
2731 | ||
2732 | ||
2733 | 4.86 KVM_PPC_RTAS_DEFINE_TOKEN | |
2734 | ||
2735 | Capability: KVM_CAP_PPC_RTAS | |
2736 | Architectures: ppc | |
2737 | Type: vm ioctl | |
2738 | Parameters: struct kvm_rtas_token_args | |
2739 | Returns: 0 on success, -1 on error | |
2740 | ||
2741 | Defines a token value for a RTAS (Run Time Abstraction Services) | |
2742 | service in order to allow it to be handled in the kernel. The | |
2743 | argument struct gives the name of the service, which must be the name | |
2744 | of a service that has a kernel-side implementation. If the token | |
2745 | value is non-zero, it will be associated with that service, and | |
2746 | subsequent RTAS calls by the guest specifying that token will be | |
2747 | handled by the kernel. If the token value is 0, then any token | |
2748 | associated with the service will be forgotten, and subsequent RTAS | |
2749 | calls by the guest for that service will be passed to userspace to be | |
2750 | handled. | |
2751 | ||
2752 | 4.87 KVM_SET_GUEST_DEBUG | |
2753 | ||
2754 | Capability: KVM_CAP_SET_GUEST_DEBUG | |
2755 | Architectures: x86, s390, ppc, arm64 | |
2756 | Type: vcpu ioctl | |
2757 | Parameters: struct kvm_guest_debug (in) | |
2758 | Returns: 0 on success; -1 on error | |
2759 | ||
2760 | struct kvm_guest_debug { | |
2761 | __u32 control; | |
2762 | __u32 pad; | |
2763 | struct kvm_guest_debug_arch arch; | |
2764 | }; | |
2765 | ||
2766 | Set up the processor specific debug registers and configure vcpu for | |
2767 | handling guest debug events. There are two parts to the structure, the | |
2768 | first a control bitfield indicates the type of debug events to handle | |
2769 | when running. Common control bits are: | |
2770 | ||
2771 | - KVM_GUESTDBG_ENABLE: guest debugging is enabled | |
2772 | - KVM_GUESTDBG_SINGLESTEP: the next run should single-step | |
2773 | ||
2774 | The top 16 bits of the control field are architecture specific control | |
2775 | flags which can include the following: | |
2776 | ||
2777 | - KVM_GUESTDBG_USE_SW_BP: using software breakpoints [x86, arm64] | |
2778 | - KVM_GUESTDBG_USE_HW_BP: using hardware breakpoints [x86, s390, arm64] | |
2779 | - KVM_GUESTDBG_INJECT_DB: inject DB type exception [x86] | |
2780 | - KVM_GUESTDBG_INJECT_BP: inject BP type exception [x86] | |
2781 | - KVM_GUESTDBG_EXIT_PENDING: trigger an immediate guest exit [s390] | |
2782 | ||
2783 | For example KVM_GUESTDBG_USE_SW_BP indicates that software breakpoints | |
2784 | are enabled in memory so we need to ensure breakpoint exceptions are | |
2785 | correctly trapped and the KVM run loop exits at the breakpoint and not | |
2786 | running off into the normal guest vector. For KVM_GUESTDBG_USE_HW_BP | |
2787 | we need to ensure the guest vCPUs architecture specific registers are | |
2788 | updated to the correct (supplied) values. | |
2789 | ||
2790 | The second part of the structure is architecture specific and | |
2791 | typically contains a set of debug registers. | |
2792 | ||
2793 | For arm64 the number of debug registers is implementation defined and | |
2794 | can be determined by querying the KVM_CAP_GUEST_DEBUG_HW_BPS and | |
2795 | KVM_CAP_GUEST_DEBUG_HW_WPS capabilities which return a positive number | |
2796 | indicating the number of supported registers. | |
2797 | ||
2798 | When debug events exit the main run loop with the reason | |
2799 | KVM_EXIT_DEBUG with the kvm_debug_exit_arch part of the kvm_run | |
2800 | structure containing architecture specific debug information. | |
2801 | ||
2802 | 4.88 KVM_GET_EMULATED_CPUID | |
2803 | ||
2804 | Capability: KVM_CAP_EXT_EMUL_CPUID | |
2805 | Architectures: x86 | |
2806 | Type: system ioctl | |
2807 | Parameters: struct kvm_cpuid2 (in/out) | |
2808 | Returns: 0 on success, -1 on error | |
2809 | ||
2810 | struct kvm_cpuid2 { | |
2811 | __u32 nent; | |
2812 | __u32 flags; | |
2813 | struct kvm_cpuid_entry2 entries[0]; | |
2814 | }; | |
2815 | ||
2816 | The member 'flags' is used for passing flags from userspace. | |
2817 | ||
2818 | #define KVM_CPUID_FLAG_SIGNIFCANT_INDEX BIT(0) | |
2819 | #define KVM_CPUID_FLAG_STATEFUL_FUNC BIT(1) | |
2820 | #define KVM_CPUID_FLAG_STATE_READ_NEXT BIT(2) | |
2821 | ||
2822 | struct kvm_cpuid_entry2 { | |
2823 | __u32 function; | |
2824 | __u32 index; | |
2825 | __u32 flags; | |
2826 | __u32 eax; | |
2827 | __u32 ebx; | |
2828 | __u32 ecx; | |
2829 | __u32 edx; | |
2830 | __u32 padding[3]; | |
2831 | }; | |
2832 | ||
2833 | This ioctl returns x86 cpuid features which are emulated by | |
2834 | kvm.Userspace can use the information returned by this ioctl to query | |
2835 | which features are emulated by kvm instead of being present natively. | |
2836 | ||
2837 | Userspace invokes KVM_GET_EMULATED_CPUID by passing a kvm_cpuid2 | |
2838 | structure with the 'nent' field indicating the number of entries in | |
2839 | the variable-size array 'entries'. If the number of entries is too low | |
2840 | to describe the cpu capabilities, an error (E2BIG) is returned. If the | |
2841 | number is too high, the 'nent' field is adjusted and an error (ENOMEM) | |
2842 | is returned. If the number is just right, the 'nent' field is adjusted | |
2843 | to the number of valid entries in the 'entries' array, which is then | |
2844 | filled. | |
2845 | ||
2846 | The entries returned are the set CPUID bits of the respective features | |
2847 | which kvm emulates, as returned by the CPUID instruction, with unknown | |
2848 | or unsupported feature bits cleared. | |
2849 | ||
2850 | Features like x2apic, for example, may not be present in the host cpu | |
2851 | but are exposed by kvm in KVM_GET_SUPPORTED_CPUID because they can be | |
2852 | emulated efficiently and thus not included here. | |
2853 | ||
2854 | The fields in each entry are defined as follows: | |
2855 | ||
2856 | function: the eax value used to obtain the entry | |
2857 | index: the ecx value used to obtain the entry (for entries that are | |
2858 | affected by ecx) | |
2859 | flags: an OR of zero or more of the following: | |
2860 | KVM_CPUID_FLAG_SIGNIFCANT_INDEX: | |
2861 | if the index field is valid | |
2862 | KVM_CPUID_FLAG_STATEFUL_FUNC: | |
2863 | if cpuid for this function returns different values for successive | |
2864 | invocations; there will be several entries with the same function, | |
2865 | all with this flag set | |
2866 | KVM_CPUID_FLAG_STATE_READ_NEXT: | |
2867 | for KVM_CPUID_FLAG_STATEFUL_FUNC entries, set if this entry is | |
2868 | the first entry to be read by a cpu | |
2869 | eax, ebx, ecx, edx: the values returned by the cpuid instruction for | |
2870 | this function/index combination | |
2871 | ||
2872 | 4.89 KVM_S390_MEM_OP | |
2873 | ||
2874 | Capability: KVM_CAP_S390_MEM_OP | |
2875 | Architectures: s390 | |
2876 | Type: vcpu ioctl | |
2877 | Parameters: struct kvm_s390_mem_op (in) | |
2878 | Returns: = 0 on success, | |
2879 | < 0 on generic error (e.g. -EFAULT or -ENOMEM), | |
2880 | > 0 if an exception occurred while walking the page tables | |
2881 | ||
2882 | Read or write data from/to the logical (virtual) memory of a VCPU. | |
2883 | ||
2884 | Parameters are specified via the following structure: | |
2885 | ||
2886 | struct kvm_s390_mem_op { | |
2887 | __u64 gaddr; /* the guest address */ | |
2888 | __u64 flags; /* flags */ | |
2889 | __u32 size; /* amount of bytes */ | |
2890 | __u32 op; /* type of operation */ | |
2891 | __u64 buf; /* buffer in userspace */ | |
2892 | __u8 ar; /* the access register number */ | |
2893 | __u8 reserved[31]; /* should be set to 0 */ | |
2894 | }; | |
2895 | ||
2896 | The type of operation is specified in the "op" field. It is either | |
2897 | KVM_S390_MEMOP_LOGICAL_READ for reading from logical memory space or | |
2898 | KVM_S390_MEMOP_LOGICAL_WRITE for writing to logical memory space. The | |
2899 | KVM_S390_MEMOP_F_CHECK_ONLY flag can be set in the "flags" field to check | |
2900 | whether the corresponding memory access would create an access exception | |
2901 | (without touching the data in the memory at the destination). In case an | |
2902 | access exception occurred while walking the MMU tables of the guest, the | |
2903 | ioctl returns a positive error number to indicate the type of exception. | |
2904 | This exception is also raised directly at the corresponding VCPU if the | |
2905 | flag KVM_S390_MEMOP_F_INJECT_EXCEPTION is set in the "flags" field. | |
2906 | ||
2907 | The start address of the memory region has to be specified in the "gaddr" | |
2908 | field, and the length of the region in the "size" field. "buf" is the buffer | |
2909 | supplied by the userspace application where the read data should be written | |
2910 | to for KVM_S390_MEMOP_LOGICAL_READ, or where the data that should be written | |
2911 | is stored for a KVM_S390_MEMOP_LOGICAL_WRITE. "buf" is unused and can be NULL | |
2912 | when KVM_S390_MEMOP_F_CHECK_ONLY is specified. "ar" designates the access | |
2913 | register number to be used. | |
2914 | ||
2915 | The "reserved" field is meant for future extensions. It is not used by | |
2916 | KVM with the currently defined set of flags. | |
2917 | ||
2918 | 4.90 KVM_S390_GET_SKEYS | |
2919 | ||
2920 | Capability: KVM_CAP_S390_SKEYS | |
2921 | Architectures: s390 | |
2922 | Type: vm ioctl | |
2923 | Parameters: struct kvm_s390_skeys | |
2924 | Returns: 0 on success, KVM_S390_GET_KEYS_NONE if guest is not using storage | |
2925 | keys, negative value on error | |
2926 | ||
2927 | This ioctl is used to get guest storage key values on the s390 | |
2928 | architecture. The ioctl takes parameters via the kvm_s390_skeys struct. | |
2929 | ||
2930 | struct kvm_s390_skeys { | |
2931 | __u64 start_gfn; | |
2932 | __u64 count; | |
2933 | __u64 skeydata_addr; | |
2934 | __u32 flags; | |
2935 | __u32 reserved[9]; | |
2936 | }; | |
2937 | ||
2938 | The start_gfn field is the number of the first guest frame whose storage keys | |
2939 | you want to get. | |
2940 | ||
2941 | The count field is the number of consecutive frames (starting from start_gfn) | |
2942 | whose storage keys to get. The count field must be at least 1 and the maximum | |
2943 | allowed value is defined as KVM_S390_SKEYS_ALLOC_MAX. Values outside this range | |
2944 | will cause the ioctl to return -EINVAL. | |
2945 | ||
2946 | The skeydata_addr field is the address to a buffer large enough to hold count | |
2947 | bytes. This buffer will be filled with storage key data by the ioctl. | |
2948 | ||
2949 | 4.91 KVM_S390_SET_SKEYS | |
2950 | ||
2951 | Capability: KVM_CAP_S390_SKEYS | |
2952 | Architectures: s390 | |
2953 | Type: vm ioctl | |
2954 | Parameters: struct kvm_s390_skeys | |
2955 | Returns: 0 on success, negative value on error | |
2956 | ||
2957 | This ioctl is used to set guest storage key values on the s390 | |
2958 | architecture. The ioctl takes parameters via the kvm_s390_skeys struct. | |
2959 | See section on KVM_S390_GET_SKEYS for struct definition. | |
2960 | ||
2961 | The start_gfn field is the number of the first guest frame whose storage keys | |
2962 | you want to set. | |
2963 | ||
2964 | The count field is the number of consecutive frames (starting from start_gfn) | |
2965 | whose storage keys to get. The count field must be at least 1 and the maximum | |
2966 | allowed value is defined as KVM_S390_SKEYS_ALLOC_MAX. Values outside this range | |
2967 | will cause the ioctl to return -EINVAL. | |
2968 | ||
2969 | The skeydata_addr field is the address to a buffer containing count bytes of | |
2970 | storage keys. Each byte in the buffer will be set as the storage key for a | |
2971 | single frame starting at start_gfn for count frames. | |
2972 | ||
2973 | Note: If any architecturally invalid key value is found in the given data then | |
2974 | the ioctl will return -EINVAL. | |
2975 | ||
2976 | 4.92 KVM_S390_IRQ | |
2977 | ||
2978 | Capability: KVM_CAP_S390_INJECT_IRQ | |
2979 | Architectures: s390 | |
2980 | Type: vcpu ioctl | |
2981 | Parameters: struct kvm_s390_irq (in) | |
2982 | Returns: 0 on success, -1 on error | |
2983 | Errors: | |
2984 | EINVAL: interrupt type is invalid | |
2985 | type is KVM_S390_SIGP_STOP and flag parameter is invalid value | |
2986 | type is KVM_S390_INT_EXTERNAL_CALL and code is bigger | |
2987 | than the maximum of VCPUs | |
2988 | EBUSY: type is KVM_S390_SIGP_SET_PREFIX and vcpu is not stopped | |
2989 | type is KVM_S390_SIGP_STOP and a stop irq is already pending | |
2990 | type is KVM_S390_INT_EXTERNAL_CALL and an external call interrupt | |
2991 | is already pending | |
2992 | ||
2993 | Allows to inject an interrupt to the guest. | |
2994 | ||
2995 | Using struct kvm_s390_irq as a parameter allows | |
2996 | to inject additional payload which is not | |
2997 | possible via KVM_S390_INTERRUPT. | |
2998 | ||
2999 | Interrupt parameters are passed via kvm_s390_irq: | |
3000 | ||
3001 | struct kvm_s390_irq { | |
3002 | __u64 type; | |
3003 | union { | |
3004 | struct kvm_s390_io_info io; | |
3005 | struct kvm_s390_ext_info ext; | |
3006 | struct kvm_s390_pgm_info pgm; | |
3007 | struct kvm_s390_emerg_info emerg; | |
3008 | struct kvm_s390_extcall_info extcall; | |
3009 | struct kvm_s390_prefix_info prefix; | |
3010 | struct kvm_s390_stop_info stop; | |
3011 | struct kvm_s390_mchk_info mchk; | |
3012 | char reserved[64]; | |
3013 | } u; | |
3014 | }; | |
3015 | ||
3016 | type can be one of the following: | |
3017 | ||
3018 | KVM_S390_SIGP_STOP - sigp stop; parameter in .stop | |
3019 | KVM_S390_PROGRAM_INT - program check; parameters in .pgm | |
3020 | KVM_S390_SIGP_SET_PREFIX - sigp set prefix; parameters in .prefix | |
3021 | KVM_S390_RESTART - restart; no parameters | |
3022 | KVM_S390_INT_CLOCK_COMP - clock comparator interrupt; no parameters | |
3023 | KVM_S390_INT_CPU_TIMER - CPU timer interrupt; no parameters | |
3024 | KVM_S390_INT_EMERGENCY - sigp emergency; parameters in .emerg | |
3025 | KVM_S390_INT_EXTERNAL_CALL - sigp external call; parameters in .extcall | |
3026 | KVM_S390_MCHK - machine check interrupt; parameters in .mchk | |
3027 | ||
3028 | ||
3029 | Note that the vcpu ioctl is asynchronous to vcpu execution. | |
3030 | ||
3031 | 4.94 KVM_S390_GET_IRQ_STATE | |
3032 | ||
3033 | Capability: KVM_CAP_S390_IRQ_STATE | |
3034 | Architectures: s390 | |
3035 | Type: vcpu ioctl | |
3036 | Parameters: struct kvm_s390_irq_state (out) | |
3037 | Returns: >= number of bytes copied into buffer, | |
3038 | -EINVAL if buffer size is 0, | |
3039 | -ENOBUFS if buffer size is too small to fit all pending interrupts, | |
3040 | -EFAULT if the buffer address was invalid | |
3041 | ||
3042 | This ioctl allows userspace to retrieve the complete state of all currently | |
3043 | pending interrupts in a single buffer. Use cases include migration | |
3044 | and introspection. The parameter structure contains the address of a | |
3045 | userspace buffer and its length: | |
3046 | ||
3047 | struct kvm_s390_irq_state { | |
3048 | __u64 buf; | |
3049 | __u32 flags; /* will stay unused for compatibility reasons */ | |
3050 | __u32 len; | |
3051 | __u32 reserved[4]; /* will stay unused for compatibility reasons */ | |
3052 | }; | |
3053 | ||
3054 | Userspace passes in the above struct and for each pending interrupt a | |
3055 | struct kvm_s390_irq is copied to the provided buffer. | |
3056 | ||
3057 | The structure contains a flags and a reserved field for future extensions. As | |
3058 | the kernel never checked for flags == 0 and QEMU never pre-zeroed flags and | |
3059 | reserved, these fields can not be used in the future without breaking | |
3060 | compatibility. | |
3061 | ||
3062 | If -ENOBUFS is returned the buffer provided was too small and userspace | |
3063 | may retry with a bigger buffer. | |
3064 | ||
3065 | 4.95 KVM_S390_SET_IRQ_STATE | |
3066 | ||
3067 | Capability: KVM_CAP_S390_IRQ_STATE | |
3068 | Architectures: s390 | |
3069 | Type: vcpu ioctl | |
3070 | Parameters: struct kvm_s390_irq_state (in) | |
3071 | Returns: 0 on success, | |
3072 | -EFAULT if the buffer address was invalid, | |
3073 | -EINVAL for an invalid buffer length (see below), | |
3074 | -EBUSY if there were already interrupts pending, | |
3075 | errors occurring when actually injecting the | |
3076 | interrupt. See KVM_S390_IRQ. | |
3077 | ||
3078 | This ioctl allows userspace to set the complete state of all cpu-local | |
3079 | interrupts currently pending for the vcpu. It is intended for restoring | |
3080 | interrupt state after a migration. The input parameter is a userspace buffer | |
3081 | containing a struct kvm_s390_irq_state: | |
3082 | ||
3083 | struct kvm_s390_irq_state { | |
3084 | __u64 buf; | |
3085 | __u32 flags; /* will stay unused for compatibility reasons */ | |
3086 | __u32 len; | |
3087 | __u32 reserved[4]; /* will stay unused for compatibility reasons */ | |
3088 | }; | |
3089 | ||
3090 | The restrictions for flags and reserved apply as well. | |
3091 | (see KVM_S390_GET_IRQ_STATE) | |
3092 | ||
3093 | The userspace memory referenced by buf contains a struct kvm_s390_irq | |
3094 | for each interrupt to be injected into the guest. | |
3095 | If one of the interrupts could not be injected for some reason the | |
3096 | ioctl aborts. | |
3097 | ||
3098 | len must be a multiple of sizeof(struct kvm_s390_irq). It must be > 0 | |
3099 | and it must not exceed (max_vcpus + 32) * sizeof(struct kvm_s390_irq), | |
3100 | which is the maximum number of possibly pending cpu-local interrupts. | |
3101 | ||
3102 | 4.96 KVM_SMI | |
3103 | ||
3104 | Capability: KVM_CAP_X86_SMM | |
3105 | Architectures: x86 | |
3106 | Type: vcpu ioctl | |
3107 | Parameters: none | |
3108 | Returns: 0 on success, -1 on error | |
3109 | ||
3110 | Queues an SMI on the thread's vcpu. | |
3111 | ||
3112 | 4.97 KVM_CAP_PPC_MULTITCE | |
3113 | ||
3114 | Capability: KVM_CAP_PPC_MULTITCE | |
3115 | Architectures: ppc | |
3116 | Type: vm | |
3117 | ||
3118 | This capability means the kernel is capable of handling hypercalls | |
3119 | H_PUT_TCE_INDIRECT and H_STUFF_TCE without passing those into the user | |
3120 | space. This significantly accelerates DMA operations for PPC KVM guests. | |
3121 | User space should expect that its handlers for these hypercalls | |
3122 | are not going to be called if user space previously registered LIOBN | |
3123 | in KVM (via KVM_CREATE_SPAPR_TCE or similar calls). | |
3124 | ||
3125 | In order to enable H_PUT_TCE_INDIRECT and H_STUFF_TCE use in the guest, | |
3126 | user space might have to advertise it for the guest. For example, | |
3127 | IBM pSeries (sPAPR) guest starts using them if "hcall-multi-tce" is | |
3128 | present in the "ibm,hypertas-functions" device-tree property. | |
3129 | ||
3130 | The hypercalls mentioned above may or may not be processed successfully | |
3131 | in the kernel based fast path. If they can not be handled by the kernel, | |
3132 | they will get passed on to user space. So user space still has to have | |
3133 | an implementation for these despite the in kernel acceleration. | |
3134 | ||
3135 | This capability is always enabled. | |
3136 | ||
3137 | 4.98 KVM_CREATE_SPAPR_TCE_64 | |
3138 | ||
3139 | Capability: KVM_CAP_SPAPR_TCE_64 | |
3140 | Architectures: powerpc | |
3141 | Type: vm ioctl | |
3142 | Parameters: struct kvm_create_spapr_tce_64 (in) | |
3143 | Returns: file descriptor for manipulating the created TCE table | |
3144 | ||
3145 | This is an extension for KVM_CAP_SPAPR_TCE which only supports 32bit | |
3146 | windows, described in 4.62 KVM_CREATE_SPAPR_TCE | |
3147 | ||
3148 | This capability uses extended struct in ioctl interface: | |
3149 | ||
3150 | /* for KVM_CAP_SPAPR_TCE_64 */ | |
3151 | struct kvm_create_spapr_tce_64 { | |
3152 | __u64 liobn; | |
3153 | __u32 page_shift; | |
3154 | __u32 flags; | |
3155 | __u64 offset; /* in pages */ | |
3156 | __u64 size; /* in pages */ | |
3157 | }; | |
3158 | ||
3159 | The aim of extension is to support an additional bigger DMA window with | |
3160 | a variable page size. | |
3161 | KVM_CREATE_SPAPR_TCE_64 receives a 64bit window size, an IOMMU page shift and | |
3162 | a bus offset of the corresponding DMA window, @size and @offset are numbers | |
3163 | of IOMMU pages. | |
3164 | ||
3165 | @flags are not used at the moment. | |
3166 | ||
3167 | The rest of functionality is identical to KVM_CREATE_SPAPR_TCE. | |
3168 | ||
3169 | 4.99 KVM_REINJECT_CONTROL | |
3170 | ||
3171 | Capability: KVM_CAP_REINJECT_CONTROL | |
3172 | Architectures: x86 | |
3173 | Type: vm ioctl | |
3174 | Parameters: struct kvm_reinject_control (in) | |
3175 | Returns: 0 on success, | |
3176 | -EFAULT if struct kvm_reinject_control cannot be read, | |
3177 | -ENXIO if KVM_CREATE_PIT or KVM_CREATE_PIT2 didn't succeed earlier. | |
3178 | ||
3179 | i8254 (PIT) has two modes, reinject and !reinject. The default is reinject, | |
3180 | where KVM queues elapsed i8254 ticks and monitors completion of interrupt from | |
3181 | vector(s) that i8254 injects. Reinject mode dequeues a tick and injects its | |
3182 | interrupt whenever there isn't a pending interrupt from i8254. | |
3183 | !reinject mode injects an interrupt as soon as a tick arrives. | |
3184 | ||
3185 | struct kvm_reinject_control { | |
3186 | __u8 pit_reinject; | |
3187 | __u8 reserved[31]; | |
3188 | }; | |
3189 | ||
3190 | pit_reinject = 0 (!reinject mode) is recommended, unless running an old | |
3191 | operating system that uses the PIT for timing (e.g. Linux 2.4.x). | |
3192 | ||
3193 | 4.100 KVM_PPC_CONFIGURE_V3_MMU | |
3194 | ||
3195 | Capability: KVM_CAP_PPC_RADIX_MMU or KVM_CAP_PPC_HASH_MMU_V3 | |
3196 | Architectures: ppc | |
3197 | Type: vm ioctl | |
3198 | Parameters: struct kvm_ppc_mmuv3_cfg (in) | |
3199 | Returns: 0 on success, | |
3200 | -EFAULT if struct kvm_ppc_mmuv3_cfg cannot be read, | |
3201 | -EINVAL if the configuration is invalid | |
3202 | ||
3203 | This ioctl controls whether the guest will use radix or HPT (hashed | |
3204 | page table) translation, and sets the pointer to the process table for | |
3205 | the guest. | |
3206 | ||
3207 | struct kvm_ppc_mmuv3_cfg { | |
3208 | __u64 flags; | |
3209 | __u64 process_table; | |
3210 | }; | |
3211 | ||
3212 | There are two bits that can be set in flags; KVM_PPC_MMUV3_RADIX and | |
3213 | KVM_PPC_MMUV3_GTSE. KVM_PPC_MMUV3_RADIX, if set, configures the guest | |
3214 | to use radix tree translation, and if clear, to use HPT translation. | |
3215 | KVM_PPC_MMUV3_GTSE, if set and if KVM permits it, configures the guest | |
3216 | to be able to use the global TLB and SLB invalidation instructions; | |
3217 | if clear, the guest may not use these instructions. | |
3218 | ||
3219 | The process_table field specifies the address and size of the guest | |
3220 | process table, which is in the guest's space. This field is formatted | |
3221 | as the second doubleword of the partition table entry, as defined in | |
3222 | the Power ISA V3.00, Book III section 5.7.6.1. | |
3223 | ||
3224 | 4.101 KVM_PPC_GET_RMMU_INFO | |
3225 | ||
3226 | Capability: KVM_CAP_PPC_RADIX_MMU | |
3227 | Architectures: ppc | |
3228 | Type: vm ioctl | |
3229 | Parameters: struct kvm_ppc_rmmu_info (out) | |
3230 | Returns: 0 on success, | |
3231 | -EFAULT if struct kvm_ppc_rmmu_info cannot be written, | |
3232 | -EINVAL if no useful information can be returned | |
3233 | ||
3234 | This ioctl returns a structure containing two things: (a) a list | |
3235 | containing supported radix tree geometries, and (b) a list that maps | |
3236 | page sizes to put in the "AP" (actual page size) field for the tlbie | |
3237 | (TLB invalidate entry) instruction. | |
3238 | ||
3239 | struct kvm_ppc_rmmu_info { | |
3240 | struct kvm_ppc_radix_geom { | |
3241 | __u8 page_shift; | |
3242 | __u8 level_bits[4]; | |
3243 | __u8 pad[3]; | |
3244 | } geometries[8]; | |
3245 | __u32 ap_encodings[8]; | |
3246 | }; | |
3247 | ||
3248 | The geometries[] field gives up to 8 supported geometries for the | |
3249 | radix page table, in terms of the log base 2 of the smallest page | |
3250 | size, and the number of bits indexed at each level of the tree, from | |
3251 | the PTE level up to the PGD level in that order. Any unused entries | |
3252 | will have 0 in the page_shift field. | |
3253 | ||
3254 | The ap_encodings gives the supported page sizes and their AP field | |
3255 | encodings, encoded with the AP value in the top 3 bits and the log | |
3256 | base 2 of the page size in the bottom 6 bits. | |
3257 | ||
3258 | 4.102 KVM_PPC_RESIZE_HPT_PREPARE | |
3259 | ||
3260 | Capability: KVM_CAP_SPAPR_RESIZE_HPT | |
3261 | Architectures: powerpc | |
3262 | Type: vm ioctl | |
3263 | Parameters: struct kvm_ppc_resize_hpt (in) | |
3264 | Returns: 0 on successful completion, | |
3265 | >0 if a new HPT is being prepared, the value is an estimated | |
3266 | number of milliseconds until preparation is complete | |
3267 | -EFAULT if struct kvm_reinject_control cannot be read, | |
3268 | -EINVAL if the supplied shift or flags are invalid | |
3269 | -ENOMEM if unable to allocate the new HPT | |
3270 | -ENOSPC if there was a hash collision when moving existing | |
3271 | HPT entries to the new HPT | |
3272 | -EIO on other error conditions | |
3273 | ||
3274 | Used to implement the PAPR extension for runtime resizing of a guest's | |
3275 | Hashed Page Table (HPT). Specifically this starts, stops or monitors | |
3276 | the preparation of a new potential HPT for the guest, essentially | |
3277 | implementing the H_RESIZE_HPT_PREPARE hypercall. | |
3278 | ||
3279 | If called with shift > 0 when there is no pending HPT for the guest, | |
3280 | this begins preparation of a new pending HPT of size 2^(shift) bytes. | |
3281 | It then returns a positive integer with the estimated number of | |
3282 | milliseconds until preparation is complete. | |
3283 | ||
3284 | If called when there is a pending HPT whose size does not match that | |
3285 | requested in the parameters, discards the existing pending HPT and | |
3286 | creates a new one as above. | |
3287 | ||
3288 | If called when there is a pending HPT of the size requested, will: | |
3289 | * If preparation of the pending HPT is already complete, return 0 | |
3290 | * If preparation of the pending HPT has failed, return an error | |
3291 | code, then discard the pending HPT. | |
3292 | * If preparation of the pending HPT is still in progress, return an | |
3293 | estimated number of milliseconds until preparation is complete. | |
3294 | ||
3295 | If called with shift == 0, discards any currently pending HPT and | |
3296 | returns 0 (i.e. cancels any in-progress preparation). | |
3297 | ||
3298 | flags is reserved for future expansion, currently setting any bits in | |
3299 | flags will result in an -EINVAL. | |
3300 | ||
3301 | Normally this will be called repeatedly with the same parameters until | |
3302 | it returns <= 0. The first call will initiate preparation, subsequent | |
3303 | ones will monitor preparation until it completes or fails. | |
3304 | ||
3305 | struct kvm_ppc_resize_hpt { | |
3306 | __u64 flags; | |
3307 | __u32 shift; | |
3308 | __u32 pad; | |
3309 | }; | |
3310 | ||
3311 | 4.103 KVM_PPC_RESIZE_HPT_COMMIT | |
3312 | ||
3313 | Capability: KVM_CAP_SPAPR_RESIZE_HPT | |
3314 | Architectures: powerpc | |
3315 | Type: vm ioctl | |
3316 | Parameters: struct kvm_ppc_resize_hpt (in) | |
3317 | Returns: 0 on successful completion, | |
3318 | -EFAULT if struct kvm_reinject_control cannot be read, | |
3319 | -EINVAL if the supplied shift or flags are invalid | |
3320 | -ENXIO is there is no pending HPT, or the pending HPT doesn't | |
3321 | have the requested size | |
3322 | -EBUSY if the pending HPT is not fully prepared | |
3323 | -ENOSPC if there was a hash collision when moving existing | |
3324 | HPT entries to the new HPT | |
3325 | -EIO on other error conditions | |
3326 | ||
3327 | Used to implement the PAPR extension for runtime resizing of a guest's | |
3328 | Hashed Page Table (HPT). Specifically this requests that the guest be | |
3329 | transferred to working with the new HPT, essentially implementing the | |
3330 | H_RESIZE_HPT_COMMIT hypercall. | |
3331 | ||
3332 | This should only be called after KVM_PPC_RESIZE_HPT_PREPARE has | |
3333 | returned 0 with the same parameters. In other cases | |
3334 | KVM_PPC_RESIZE_HPT_COMMIT will return an error (usually -ENXIO or | |
3335 | -EBUSY, though others may be possible if the preparation was started, | |
3336 | but failed). | |
3337 | ||
3338 | This will have undefined effects on the guest if it has not already | |
3339 | placed itself in a quiescent state where no vcpu will make MMU enabled | |
3340 | memory accesses. | |
3341 | ||
3342 | On succsful completion, the pending HPT will become the guest's active | |
3343 | HPT and the previous HPT will be discarded. | |
3344 | ||
3345 | On failure, the guest will still be operating on its previous HPT. | |
3346 | ||
3347 | struct kvm_ppc_resize_hpt { | |
3348 | __u64 flags; | |
3349 | __u32 shift; | |
3350 | __u32 pad; | |
3351 | }; | |
3352 | ||
3353 | 4.104 KVM_X86_GET_MCE_CAP_SUPPORTED | |
3354 | ||
3355 | Capability: KVM_CAP_MCE | |
3356 | Architectures: x86 | |
3357 | Type: system ioctl | |
3358 | Parameters: u64 mce_cap (out) | |
3359 | Returns: 0 on success, -1 on error | |
3360 | ||
3361 | Returns supported MCE capabilities. The u64 mce_cap parameter | |
3362 | has the same format as the MSR_IA32_MCG_CAP register. Supported | |
3363 | capabilities will have the corresponding bits set. | |
3364 | ||
3365 | 4.105 KVM_X86_SETUP_MCE | |
3366 | ||
3367 | Capability: KVM_CAP_MCE | |
3368 | Architectures: x86 | |
3369 | Type: vcpu ioctl | |
3370 | Parameters: u64 mcg_cap (in) | |
3371 | Returns: 0 on success, | |
3372 | -EFAULT if u64 mcg_cap cannot be read, | |
3373 | -EINVAL if the requested number of banks is invalid, | |
3374 | -EINVAL if requested MCE capability is not supported. | |
3375 | ||
3376 | Initializes MCE support for use. The u64 mcg_cap parameter | |
3377 | has the same format as the MSR_IA32_MCG_CAP register and | |
3378 | specifies which capabilities should be enabled. The maximum | |
3379 | supported number of error-reporting banks can be retrieved when | |
3380 | checking for KVM_CAP_MCE. The supported capabilities can be | |
3381 | retrieved with KVM_X86_GET_MCE_CAP_SUPPORTED. | |
3382 | ||
3383 | 4.106 KVM_X86_SET_MCE | |
3384 | ||
3385 | Capability: KVM_CAP_MCE | |
3386 | Architectures: x86 | |
3387 | Type: vcpu ioctl | |
3388 | Parameters: struct kvm_x86_mce (in) | |
3389 | Returns: 0 on success, | |
3390 | -EFAULT if struct kvm_x86_mce cannot be read, | |
3391 | -EINVAL if the bank number is invalid, | |
3392 | -EINVAL if VAL bit is not set in status field. | |
3393 | ||
3394 | Inject a machine check error (MCE) into the guest. The input | |
3395 | parameter is: | |
3396 | ||
3397 | struct kvm_x86_mce { | |
3398 | __u64 status; | |
3399 | __u64 addr; | |
3400 | __u64 misc; | |
3401 | __u64 mcg_status; | |
3402 | __u8 bank; | |
3403 | __u8 pad1[7]; | |
3404 | __u64 pad2[3]; | |
3405 | }; | |
3406 | ||
3407 | If the MCE being reported is an uncorrected error, KVM will | |
3408 | inject it as an MCE exception into the guest. If the guest | |
3409 | MCG_STATUS register reports that an MCE is in progress, KVM | |
3410 | causes an KVM_EXIT_SHUTDOWN vmexit. | |
3411 | ||
3412 | Otherwise, if the MCE is a corrected error, KVM will just | |
3413 | store it in the corresponding bank (provided this bank is | |
3414 | not holding a previously reported uncorrected error). | |
3415 | ||
3416 | 4.107 KVM_S390_GET_CMMA_BITS | |
3417 | ||
3418 | Capability: KVM_CAP_S390_CMMA_MIGRATION | |
3419 | Architectures: s390 | |
3420 | Type: vm ioctl | |
3421 | Parameters: struct kvm_s390_cmma_log (in, out) | |
3422 | Returns: 0 on success, a negative value on error | |
3423 | ||
3424 | This ioctl is used to get the values of the CMMA bits on the s390 | |
3425 | architecture. It is meant to be used in two scenarios: | |
3426 | - During live migration to save the CMMA values. Live migration needs | |
3427 | to be enabled via the KVM_REQ_START_MIGRATION VM property. | |
3428 | - To non-destructively peek at the CMMA values, with the flag | |
3429 | KVM_S390_CMMA_PEEK set. | |
3430 | ||
3431 | The ioctl takes parameters via the kvm_s390_cmma_log struct. The desired | |
3432 | values are written to a buffer whose location is indicated via the "values" | |
3433 | member in the kvm_s390_cmma_log struct. The values in the input struct are | |
3434 | also updated as needed. | |
3435 | Each CMMA value takes up one byte. | |
3436 | ||
3437 | struct kvm_s390_cmma_log { | |
3438 | __u64 start_gfn; | |
3439 | __u32 count; | |
3440 | __u32 flags; | |
3441 | union { | |
3442 | __u64 remaining; | |
3443 | __u64 mask; | |
3444 | }; | |
3445 | __u64 values; | |
3446 | }; | |
3447 | ||
3448 | start_gfn is the number of the first guest frame whose CMMA values are | |
3449 | to be retrieved, | |
3450 | ||
3451 | count is the length of the buffer in bytes, | |
3452 | ||
3453 | values points to the buffer where the result will be written to. | |
3454 | ||
3455 | If count is greater than KVM_S390_SKEYS_MAX, then it is considered to be | |
3456 | KVM_S390_SKEYS_MAX. KVM_S390_SKEYS_MAX is re-used for consistency with | |
3457 | other ioctls. | |
3458 | ||
3459 | The result is written in the buffer pointed to by the field values, and | |
3460 | the values of the input parameter are updated as follows. | |
3461 | ||
3462 | Depending on the flags, different actions are performed. The only | |
3463 | supported flag so far is KVM_S390_CMMA_PEEK. | |
3464 | ||
3465 | The default behaviour if KVM_S390_CMMA_PEEK is not set is: | |
3466 | start_gfn will indicate the first page frame whose CMMA bits were dirty. | |
3467 | It is not necessarily the same as the one passed as input, as clean pages | |
3468 | are skipped. | |
3469 | ||
3470 | count will indicate the number of bytes actually written in the buffer. | |
3471 | It can (and very often will) be smaller than the input value, since the | |
3472 | buffer is only filled until 16 bytes of clean values are found (which | |
3473 | are then not copied in the buffer). Since a CMMA migration block needs | |
3474 | the base address and the length, for a total of 16 bytes, we will send | |
3475 | back some clean data if there is some dirty data afterwards, as long as | |
3476 | the size of the clean data does not exceed the size of the header. This | |
3477 | allows to minimize the amount of data to be saved or transferred over | |
3478 | the network at the expense of more roundtrips to userspace. The next | |
3479 | invocation of the ioctl will skip over all the clean values, saving | |
3480 | potentially more than just the 16 bytes we found. | |
3481 | ||
3482 | If KVM_S390_CMMA_PEEK is set: | |
3483 | the existing storage attributes are read even when not in migration | |
3484 | mode, and no other action is performed; | |
3485 | ||
3486 | the output start_gfn will be equal to the input start_gfn, | |
3487 | ||
3488 | the output count will be equal to the input count, except if the end of | |
3489 | memory has been reached. | |
3490 | ||
3491 | In both cases: | |
3492 | the field "remaining" will indicate the total number of dirty CMMA values | |
3493 | still remaining, or 0 if KVM_S390_CMMA_PEEK is set and migration mode is | |
3494 | not enabled. | |
3495 | ||
3496 | mask is unused. | |
3497 | ||
3498 | values points to the userspace buffer where the result will be stored. | |
3499 | ||
3500 | This ioctl can fail with -ENOMEM if not enough memory can be allocated to | |
3501 | complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if | |
3502 | KVM_S390_CMMA_PEEK is not set but migration mode was not enabled, with | |
3503 | -EFAULT if the userspace address is invalid or if no page table is | |
3504 | present for the addresses (e.g. when using hugepages). | |
3505 | ||
3506 | 4.108 KVM_S390_SET_CMMA_BITS | |
3507 | ||
3508 | Capability: KVM_CAP_S390_CMMA_MIGRATION | |
3509 | Architectures: s390 | |
3510 | Type: vm ioctl | |
3511 | Parameters: struct kvm_s390_cmma_log (in) | |
3512 | Returns: 0 on success, a negative value on error | |
3513 | ||
3514 | This ioctl is used to set the values of the CMMA bits on the s390 | |
3515 | architecture. It is meant to be used during live migration to restore | |
3516 | the CMMA values, but there are no restrictions on its use. | |
3517 | The ioctl takes parameters via the kvm_s390_cmma_values struct. | |
3518 | Each CMMA value takes up one byte. | |
3519 | ||
3520 | struct kvm_s390_cmma_log { | |
3521 | __u64 start_gfn; | |
3522 | __u32 count; | |
3523 | __u32 flags; | |
3524 | union { | |
3525 | __u64 remaining; | |
3526 | __u64 mask; | |
3527 | }; | |
3528 | __u64 values; | |
3529 | }; | |
3530 | ||
3531 | start_gfn indicates the starting guest frame number, | |
3532 | ||
3533 | count indicates how many values are to be considered in the buffer, | |
3534 | ||
3535 | flags is not used and must be 0. | |
3536 | ||
3537 | mask indicates which PGSTE bits are to be considered. | |
3538 | ||
3539 | remaining is not used. | |
3540 | ||
3541 | values points to the buffer in userspace where to store the values. | |
3542 | ||
3543 | This ioctl can fail with -ENOMEM if not enough memory can be allocated to | |
3544 | complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if | |
3545 | the count field is too large (e.g. more than KVM_S390_CMMA_SIZE_MAX) or | |
3546 | if the flags field was not 0, with -EFAULT if the userspace address is | |
3547 | invalid, if invalid pages are written to (e.g. after the end of memory) | |
3548 | or if no page table is present for the addresses (e.g. when using | |
3549 | hugepages). | |
3550 | ||
3551 | 4.109 KVM_PPC_GET_CPU_CHAR | |
3552 | ||
3553 | Capability: KVM_CAP_PPC_GET_CPU_CHAR | |
3554 | Architectures: powerpc | |
3555 | Type: vm ioctl | |
3556 | Parameters: struct kvm_ppc_cpu_char (out) | |
3557 | Returns: 0 on successful completion | |
3558 | -EFAULT if struct kvm_ppc_cpu_char cannot be written | |
3559 | ||
3560 | This ioctl gives userspace information about certain characteristics | |
3561 | of the CPU relating to speculative execution of instructions and | |
3562 | possible information leakage resulting from speculative execution (see | |
3563 | CVE-2017-5715, CVE-2017-5753 and CVE-2017-5754). The information is | |
3564 | returned in struct kvm_ppc_cpu_char, which looks like this: | |
3565 | ||
3566 | struct kvm_ppc_cpu_char { | |
3567 | __u64 character; /* characteristics of the CPU */ | |
3568 | __u64 behaviour; /* recommended software behaviour */ | |
3569 | __u64 character_mask; /* valid bits in character */ | |
3570 | __u64 behaviour_mask; /* valid bits in behaviour */ | |
3571 | }; | |
3572 | ||
3573 | For extensibility, the character_mask and behaviour_mask fields | |
3574 | indicate which bits of character and behaviour have been filled in by | |
3575 | the kernel. If the set of defined bits is extended in future then | |
3576 | userspace will be able to tell whether it is running on a kernel that | |
3577 | knows about the new bits. | |
3578 | ||
3579 | The character field describes attributes of the CPU which can help | |
3580 | with preventing inadvertent information disclosure - specifically, | |
3581 | whether there is an instruction to flash-invalidate the L1 data cache | |
3582 | (ori 30,30,0 or mtspr SPRN_TRIG2,rN), whether the L1 data cache is set | |
3583 | to a mode where entries can only be used by the thread that created | |
3584 | them, whether the bcctr[l] instruction prevents speculation, and | |
3585 | whether a speculation barrier instruction (ori 31,31,0) is provided. | |
3586 | ||
3587 | The behaviour field describes actions that software should take to | |
3588 | prevent inadvertent information disclosure, and thus describes which | |
3589 | vulnerabilities the hardware is subject to; specifically whether the | |
3590 | L1 data cache should be flushed when returning to user mode from the | |
3591 | kernel, and whether a speculation barrier should be placed between an | |
3592 | array bounds check and the array access. | |
3593 | ||
3594 | These fields use the same bit definitions as the new | |
3595 | H_GET_CPU_CHARACTERISTICS hypercall. | |
3596 | ||
3597 | 4.110 KVM_MEMORY_ENCRYPT_OP | |
3598 | ||
3599 | Capability: basic | |
3600 | Architectures: x86 | |
3601 | Type: system | |
3602 | Parameters: an opaque platform specific structure (in/out) | |
3603 | Returns: 0 on success; -1 on error | |
3604 | ||
3605 | If the platform supports creating encrypted VMs then this ioctl can be used | |
3606 | for issuing platform-specific memory encryption commands to manage those | |
3607 | encrypted VMs. | |
3608 | ||
3609 | Currently, this ioctl is used for issuing Secure Encrypted Virtualization | |
3610 | (SEV) commands on AMD Processors. The SEV commands are defined in | |
3611 | Documentation/virtual/kvm/amd-memory-encryption.rst. | |
3612 | ||
3613 | 4.111 KVM_MEMORY_ENCRYPT_REG_REGION | |
3614 | ||
3615 | Capability: basic | |
3616 | Architectures: x86 | |
3617 | Type: system | |
3618 | Parameters: struct kvm_enc_region (in) | |
3619 | Returns: 0 on success; -1 on error | |
3620 | ||
3621 | This ioctl can be used to register a guest memory region which may | |
3622 | contain encrypted data (e.g. guest RAM, SMRAM etc). | |
3623 | ||
3624 | It is used in the SEV-enabled guest. When encryption is enabled, a guest | |
3625 | memory region may contain encrypted data. The SEV memory encryption | |
3626 | engine uses a tweak such that two identical plaintext pages, each at | |
3627 | different locations will have differing ciphertexts. So swapping or | |
3628 | moving ciphertext of those pages will not result in plaintext being | |
3629 | swapped. So relocating (or migrating) physical backing pages for the SEV | |
3630 | guest will require some additional steps. | |
3631 | ||
3632 | Note: The current SEV key management spec does not provide commands to | |
3633 | swap or migrate (move) ciphertext pages. Hence, for now we pin the guest | |
3634 | memory region registered with the ioctl. | |
3635 | ||
3636 | 4.112 KVM_MEMORY_ENCRYPT_UNREG_REGION | |
3637 | ||
3638 | Capability: basic | |
3639 | Architectures: x86 | |
3640 | Type: system | |
3641 | Parameters: struct kvm_enc_region (in) | |
3642 | Returns: 0 on success; -1 on error | |
3643 | ||
3644 | This ioctl can be used to unregister the guest memory region registered | |
3645 | with KVM_MEMORY_ENCRYPT_REG_REGION ioctl above. | |
3646 | ||
3647 | 4.113 KVM_HYPERV_EVENTFD | |
3648 | ||
3649 | Capability: KVM_CAP_HYPERV_EVENTFD | |
3650 | Architectures: x86 | |
3651 | Type: vm ioctl | |
3652 | Parameters: struct kvm_hyperv_eventfd (in) | |
3653 | ||
3654 | This ioctl (un)registers an eventfd to receive notifications from the guest on | |
3655 | the specified Hyper-V connection id through the SIGNAL_EVENT hypercall, without | |
3656 | causing a user exit. SIGNAL_EVENT hypercall with non-zero event flag number | |
3657 | (bits 24-31) still triggers a KVM_EXIT_HYPERV_HCALL user exit. | |
3658 | ||
3659 | struct kvm_hyperv_eventfd { | |
3660 | __u32 conn_id; | |
3661 | __s32 fd; | |
3662 | __u32 flags; | |
3663 | __u32 padding[3]; | |
3664 | }; | |
3665 | ||
3666 | The conn_id field should fit within 24 bits: | |
3667 | ||
3668 | #define KVM_HYPERV_CONN_ID_MASK 0x00ffffff | |
3669 | ||
3670 | The acceptable values for the flags field are: | |
3671 | ||
3672 | #define KVM_HYPERV_EVENTFD_DEASSIGN (1 << 0) | |
3673 | ||
3674 | Returns: 0 on success, | |
3675 | -EINVAL if conn_id or flags is outside the allowed range | |
3676 | -ENOENT on deassign if the conn_id isn't registered | |
3677 | -EEXIST on assign if the conn_id is already registered | |
3678 | ||
3679 | 4.114 KVM_GET_NESTED_STATE | |
3680 | ||
3681 | Capability: KVM_CAP_NESTED_STATE | |
3682 | Architectures: x86 | |
3683 | Type: vcpu ioctl | |
3684 | Parameters: struct kvm_nested_state (in/out) | |
3685 | Returns: 0 on success, -1 on error | |
3686 | Errors: | |
3687 | E2BIG: the total state size (including the fixed-size part of struct | |
3688 | kvm_nested_state) exceeds the value of 'size' specified by | |
3689 | the user; the size required will be written into size. | |
3690 | ||
3691 | struct kvm_nested_state { | |
3692 | __u16 flags; | |
3693 | __u16 format; | |
3694 | __u32 size; | |
3695 | union { | |
3696 | struct kvm_vmx_nested_state vmx; | |
3697 | struct kvm_svm_nested_state svm; | |
3698 | __u8 pad[120]; | |
3699 | }; | |
3700 | __u8 data[0]; | |
3701 | }; | |
3702 | ||
3703 | #define KVM_STATE_NESTED_GUEST_MODE 0x00000001 | |
3704 | #define KVM_STATE_NESTED_RUN_PENDING 0x00000002 | |
3705 | ||
3706 | #define KVM_STATE_NESTED_SMM_GUEST_MODE 0x00000001 | |
3707 | #define KVM_STATE_NESTED_SMM_VMXON 0x00000002 | |
3708 | ||
3709 | struct kvm_vmx_nested_state { | |
3710 | __u64 vmxon_pa; | |
3711 | __u64 vmcs_pa; | |
3712 | ||
3713 | struct { | |
3714 | __u16 flags; | |
3715 | } smm; | |
3716 | }; | |
3717 | ||
3718 | This ioctl copies the vcpu's nested virtualization state from the kernel to | |
3719 | userspace. | |
3720 | ||
3721 | The maximum size of the state, including the fixed-size part of struct | |
3722 | kvm_nested_state, can be retrieved by passing KVM_CAP_NESTED_STATE to | |
3723 | the KVM_CHECK_EXTENSION ioctl(). | |
3724 | ||
3725 | 4.115 KVM_SET_NESTED_STATE | |
3726 | ||
3727 | Capability: KVM_CAP_NESTED_STATE | |
3728 | Architectures: x86 | |
3729 | Type: vcpu ioctl | |
3730 | Parameters: struct kvm_nested_state (in) | |
3731 | Returns: 0 on success, -1 on error | |
3732 | ||
3733 | This copies the vcpu's kvm_nested_state struct from userspace to the kernel. For | |
3734 | the definition of struct kvm_nested_state, see KVM_GET_NESTED_STATE. | |
3735 | ||
3736 | 4.116 KVM_(UN)REGISTER_COALESCED_MMIO | |
3737 | ||
3738 | Capability: KVM_CAP_COALESCED_MMIO (for coalesced mmio) | |
3739 | KVM_CAP_COALESCED_PIO (for coalesced pio) | |
3740 | Architectures: all | |
3741 | Type: vm ioctl | |
3742 | Parameters: struct kvm_coalesced_mmio_zone | |
3743 | Returns: 0 on success, < 0 on error | |
3744 | ||
3745 | Coalesced I/O is a performance optimization that defers hardware | |
3746 | register write emulation so that userspace exits are avoided. It is | |
3747 | typically used to reduce the overhead of emulating frequently accessed | |
3748 | hardware registers. | |
3749 | ||
3750 | When a hardware register is configured for coalesced I/O, write accesses | |
3751 | do not exit to userspace and their value is recorded in a ring buffer | |
3752 | that is shared between kernel and userspace. | |
3753 | ||
3754 | Coalesced I/O is used if one or more write accesses to a hardware | |
3755 | register can be deferred until a read or a write to another hardware | |
3756 | register on the same device. This last access will cause a vmexit and | |
3757 | userspace will process accesses from the ring buffer before emulating | |
3758 | it. That will avoid exiting to userspace on repeated writes. | |
3759 | ||
3760 | Coalesced pio is based on coalesced mmio. There is little difference | |
3761 | between coalesced mmio and pio except that coalesced pio records accesses | |
3762 | to I/O ports. | |
3763 | ||
3764 | 4.117 KVM_CLEAR_DIRTY_LOG (vm ioctl) | |
3765 | ||
3766 | Capability: KVM_CAP_MANUAL_DIRTY_LOG_PROTECT | |
3767 | Architectures: x86 | |
3768 | Type: vm ioctl | |
3769 | Parameters: struct kvm_dirty_log (in) | |
3770 | Returns: 0 on success, -1 on error | |
3771 | ||
3772 | /* for KVM_CLEAR_DIRTY_LOG */ | |
3773 | struct kvm_clear_dirty_log { | |
3774 | __u32 slot; | |
3775 | __u32 num_pages; | |
3776 | __u64 first_page; | |
3777 | union { | |
3778 | void __user *dirty_bitmap; /* one bit per page */ | |
3779 | __u64 padding; | |
3780 | }; | |
3781 | }; | |
3782 | ||
3783 | The ioctl clears the dirty status of pages in a memory slot, according to | |
3784 | the bitmap that is passed in struct kvm_clear_dirty_log's dirty_bitmap | |
3785 | field. Bit 0 of the bitmap corresponds to page "first_page" in the | |
3786 | memory slot, and num_pages is the size in bits of the input bitmap. | |
3787 | Both first_page and num_pages must be a multiple of 64. For each bit | |
3788 | that is set in the input bitmap, the corresponding page is marked "clean" | |
3789 | in KVM's dirty bitmap, and dirty tracking is re-enabled for that page | |
3790 | (for example via write-protection, or by clearing the dirty bit in | |
3791 | a page table entry). | |
3792 | ||
3793 | If KVM_CAP_MULTI_ADDRESS_SPACE is available, bits 16-31 specifies | |
3794 | the address space for which you want to return the dirty bitmap. | |
3795 | They must be less than the value that KVM_CHECK_EXTENSION returns for | |
3796 | the KVM_CAP_MULTI_ADDRESS_SPACE capability. | |
3797 | ||
3798 | This ioctl is mostly useful when KVM_CAP_MANUAL_DIRTY_LOG_PROTECT | |
3799 | is enabled; for more information, see the description of the capability. | |
3800 | However, it can always be used as long as KVM_CHECK_EXTENSION confirms | |
3801 | that KVM_CAP_MANUAL_DIRTY_LOG_PROTECT is present. | |
3802 | ||
3803 | ||
3804 | 5. The kvm_run structure | |
3805 | ------------------------ | |
3806 | ||
3807 | Application code obtains a pointer to the kvm_run structure by | |
3808 | mmap()ing a vcpu fd. From that point, application code can control | |
3809 | execution by changing fields in kvm_run prior to calling the KVM_RUN | |
3810 | ioctl, and obtain information about the reason KVM_RUN returned by | |
3811 | looking up structure members. | |
3812 | ||
3813 | struct kvm_run { | |
3814 | /* in */ | |
3815 | __u8 request_interrupt_window; | |
3816 | ||
3817 | Request that KVM_RUN return when it becomes possible to inject external | |
3818 | interrupts into the guest. Useful in conjunction with KVM_INTERRUPT. | |
3819 | ||
3820 | __u8 immediate_exit; | |
3821 | ||
3822 | This field is polled once when KVM_RUN starts; if non-zero, KVM_RUN | |
3823 | exits immediately, returning -EINTR. In the common scenario where a | |
3824 | signal is used to "kick" a VCPU out of KVM_RUN, this field can be used | |
3825 | to avoid usage of KVM_SET_SIGNAL_MASK, which has worse scalability. | |
3826 | Rather than blocking the signal outside KVM_RUN, userspace can set up | |
3827 | a signal handler that sets run->immediate_exit to a non-zero value. | |
3828 | ||
3829 | This field is ignored if KVM_CAP_IMMEDIATE_EXIT is not available. | |
3830 | ||
3831 | __u8 padding1[6]; | |
3832 | ||
3833 | /* out */ | |
3834 | __u32 exit_reason; | |
3835 | ||
3836 | When KVM_RUN has returned successfully (return value 0), this informs | |
3837 | application code why KVM_RUN has returned. Allowable values for this | |
3838 | field are detailed below. | |
3839 | ||
3840 | __u8 ready_for_interrupt_injection; | |
3841 | ||
3842 | If request_interrupt_window has been specified, this field indicates | |
3843 | an interrupt can be injected now with KVM_INTERRUPT. | |
3844 | ||
3845 | __u8 if_flag; | |
3846 | ||
3847 | The value of the current interrupt flag. Only valid if in-kernel | |
3848 | local APIC is not used. | |
3849 | ||
3850 | __u16 flags; | |
3851 | ||
3852 | More architecture-specific flags detailing state of the VCPU that may | |
3853 | affect the device's behavior. The only currently defined flag is | |
3854 | KVM_RUN_X86_SMM, which is valid on x86 machines and is set if the | |
3855 | VCPU is in system management mode. | |
3856 | ||
3857 | /* in (pre_kvm_run), out (post_kvm_run) */ | |
3858 | __u64 cr8; | |
3859 | ||
3860 | The value of the cr8 register. Only valid if in-kernel local APIC is | |
3861 | not used. Both input and output. | |
3862 | ||
3863 | __u64 apic_base; | |
3864 | ||
3865 | The value of the APIC BASE msr. Only valid if in-kernel local | |
3866 | APIC is not used. Both input and output. | |
3867 | ||
3868 | union { | |
3869 | /* KVM_EXIT_UNKNOWN */ | |
3870 | struct { | |
3871 | __u64 hardware_exit_reason; | |
3872 | } hw; | |
3873 | ||
3874 | If exit_reason is KVM_EXIT_UNKNOWN, the vcpu has exited due to unknown | |
3875 | reasons. Further architecture-specific information is available in | |
3876 | hardware_exit_reason. | |
3877 | ||
3878 | /* KVM_EXIT_FAIL_ENTRY */ | |
3879 | struct { | |
3880 | __u64 hardware_entry_failure_reason; | |
3881 | } fail_entry; | |
3882 | ||
3883 | If exit_reason is KVM_EXIT_FAIL_ENTRY, the vcpu could not be run due | |
3884 | to unknown reasons. Further architecture-specific information is | |
3885 | available in hardware_entry_failure_reason. | |
3886 | ||
3887 | /* KVM_EXIT_EXCEPTION */ | |
3888 | struct { | |
3889 | __u32 exception; | |
3890 | __u32 error_code; | |
3891 | } ex; | |
3892 | ||
3893 | Unused. | |
3894 | ||
3895 | /* KVM_EXIT_IO */ | |
3896 | struct { | |
3897 | #define KVM_EXIT_IO_IN 0 | |
3898 | #define KVM_EXIT_IO_OUT 1 | |
3899 | __u8 direction; | |
3900 | __u8 size; /* bytes */ | |
3901 | __u16 port; | |
3902 | __u32 count; | |
3903 | __u64 data_offset; /* relative to kvm_run start */ | |
3904 | } io; | |
3905 | ||
3906 | If exit_reason is KVM_EXIT_IO, then the vcpu has | |
3907 | executed a port I/O instruction which could not be satisfied by kvm. | |
3908 | data_offset describes where the data is located (KVM_EXIT_IO_OUT) or | |
3909 | where kvm expects application code to place the data for the next | |
3910 | KVM_RUN invocation (KVM_EXIT_IO_IN). Data format is a packed array. | |
3911 | ||
3912 | /* KVM_EXIT_DEBUG */ | |
3913 | struct { | |
3914 | struct kvm_debug_exit_arch arch; | |
3915 | } debug; | |
3916 | ||
3917 | If the exit_reason is KVM_EXIT_DEBUG, then a vcpu is processing a debug event | |
3918 | for which architecture specific information is returned. | |
3919 | ||
3920 | /* KVM_EXIT_MMIO */ | |
3921 | struct { | |
3922 | __u64 phys_addr; | |
3923 | __u8 data[8]; | |
3924 | __u32 len; | |
3925 | __u8 is_write; | |
3926 | } mmio; | |
3927 | ||
3928 | If exit_reason is KVM_EXIT_MMIO, then the vcpu has | |
3929 | executed a memory-mapped I/O instruction which could not be satisfied | |
3930 | by kvm. The 'data' member contains the written data if 'is_write' is | |
3931 | true, and should be filled by application code otherwise. | |
3932 | ||
3933 | The 'data' member contains, in its first 'len' bytes, the value as it would | |
3934 | appear if the VCPU performed a load or store of the appropriate width directly | |
3935 | to the byte array. | |
3936 | ||
3937 | NOTE: For KVM_EXIT_IO, KVM_EXIT_MMIO, KVM_EXIT_OSI, KVM_EXIT_PAPR and | |
3938 | KVM_EXIT_EPR the corresponding | |
3939 | operations are complete (and guest state is consistent) only after userspace | |
3940 | has re-entered the kernel with KVM_RUN. The kernel side will first finish | |
3941 | incomplete operations and then check for pending signals. Userspace | |
3942 | can re-enter the guest with an unmasked signal pending to complete | |
3943 | pending operations. | |
3944 | ||
3945 | /* KVM_EXIT_HYPERCALL */ | |
3946 | struct { | |
3947 | __u64 nr; | |
3948 | __u64 args[6]; | |
3949 | __u64 ret; | |
3950 | __u32 longmode; | |
3951 | __u32 pad; | |
3952 | } hypercall; | |
3953 | ||
3954 | Unused. This was once used for 'hypercall to userspace'. To implement | |
3955 | such functionality, use KVM_EXIT_IO (x86) or KVM_EXIT_MMIO (all except s390). | |
3956 | Note KVM_EXIT_IO is significantly faster than KVM_EXIT_MMIO. | |
3957 | ||
3958 | /* KVM_EXIT_TPR_ACCESS */ | |
3959 | struct { | |
3960 | __u64 rip; | |
3961 | __u32 is_write; | |
3962 | __u32 pad; | |
3963 | } tpr_access; | |
3964 | ||
3965 | To be documented (KVM_TPR_ACCESS_REPORTING). | |
3966 | ||
3967 | /* KVM_EXIT_S390_SIEIC */ | |
3968 | struct { | |
3969 | __u8 icptcode; | |
3970 | __u64 mask; /* psw upper half */ | |
3971 | __u64 addr; /* psw lower half */ | |
3972 | __u16 ipa; | |
3973 | __u32 ipb; | |
3974 | } s390_sieic; | |
3975 | ||
3976 | s390 specific. | |
3977 | ||
3978 | /* KVM_EXIT_S390_RESET */ | |
3979 | #define KVM_S390_RESET_POR 1 | |
3980 | #define KVM_S390_RESET_CLEAR 2 | |
3981 | #define KVM_S390_RESET_SUBSYSTEM 4 | |
3982 | #define KVM_S390_RESET_CPU_INIT 8 | |
3983 | #define KVM_S390_RESET_IPL 16 | |
3984 | __u64 s390_reset_flags; | |
3985 | ||
3986 | s390 specific. | |
3987 | ||
3988 | /* KVM_EXIT_S390_UCONTROL */ | |
3989 | struct { | |
3990 | __u64 trans_exc_code; | |
3991 | __u32 pgm_code; | |
3992 | } s390_ucontrol; | |
3993 | ||
3994 | s390 specific. A page fault has occurred for a user controlled virtual | |
3995 | machine (KVM_VM_S390_UNCONTROL) on it's host page table that cannot be | |
3996 | resolved by the kernel. | |
3997 | The program code and the translation exception code that were placed | |
3998 | in the cpu's lowcore are presented here as defined by the z Architecture | |
3999 | Principles of Operation Book in the Chapter for Dynamic Address Translation | |
4000 | (DAT) | |
4001 | ||
4002 | /* KVM_EXIT_DCR */ | |
4003 | struct { | |
4004 | __u32 dcrn; | |
4005 | __u32 data; | |
4006 | __u8 is_write; | |
4007 | } dcr; | |
4008 | ||
4009 | Deprecated - was used for 440 KVM. | |
4010 | ||
4011 | /* KVM_EXIT_OSI */ | |
4012 | struct { | |
4013 | __u64 gprs[32]; | |
4014 | } osi; | |
4015 | ||
4016 | MOL uses a special hypercall interface it calls 'OSI'. To enable it, we catch | |
4017 | hypercalls and exit with this exit struct that contains all the guest gprs. | |
4018 | ||
4019 | If exit_reason is KVM_EXIT_OSI, then the vcpu has triggered such a hypercall. | |
4020 | Userspace can now handle the hypercall and when it's done modify the gprs as | |
4021 | necessary. Upon guest entry all guest GPRs will then be replaced by the values | |
4022 | in this struct. | |
4023 | ||
4024 | /* KVM_EXIT_PAPR_HCALL */ | |
4025 | struct { | |
4026 | __u64 nr; | |
4027 | __u64 ret; | |
4028 | __u64 args[9]; | |
4029 | } papr_hcall; | |
4030 | ||
4031 | This is used on 64-bit PowerPC when emulating a pSeries partition, | |
4032 | e.g. with the 'pseries' machine type in qemu. It occurs when the | |
4033 | guest does a hypercall using the 'sc 1' instruction. The 'nr' field | |
4034 | contains the hypercall number (from the guest R3), and 'args' contains | |
4035 | the arguments (from the guest R4 - R12). Userspace should put the | |
4036 | return code in 'ret' and any extra returned values in args[]. | |
4037 | The possible hypercalls are defined in the Power Architecture Platform | |
4038 | Requirements (PAPR) document available from www.power.org (free | |
4039 | developer registration required to access it). | |
4040 | ||
4041 | /* KVM_EXIT_S390_TSCH */ | |
4042 | struct { | |
4043 | __u16 subchannel_id; | |
4044 | __u16 subchannel_nr; | |
4045 | __u32 io_int_parm; | |
4046 | __u32 io_int_word; | |
4047 | __u32 ipb; | |
4048 | __u8 dequeued; | |
4049 | } s390_tsch; | |
4050 | ||
4051 | s390 specific. This exit occurs when KVM_CAP_S390_CSS_SUPPORT has been enabled | |
4052 | and TEST SUBCHANNEL was intercepted. If dequeued is set, a pending I/O | |
4053 | interrupt for the target subchannel has been dequeued and subchannel_id, | |
4054 | subchannel_nr, io_int_parm and io_int_word contain the parameters for that | |
4055 | interrupt. ipb is needed for instruction parameter decoding. | |
4056 | ||
4057 | /* KVM_EXIT_EPR */ | |
4058 | struct { | |
4059 | __u32 epr; | |
4060 | } epr; | |
4061 | ||
4062 | On FSL BookE PowerPC chips, the interrupt controller has a fast patch | |
4063 | interrupt acknowledge path to the core. When the core successfully | |
4064 | delivers an interrupt, it automatically populates the EPR register with | |
4065 | the interrupt vector number and acknowledges the interrupt inside | |
4066 | the interrupt controller. | |
4067 | ||
4068 | In case the interrupt controller lives in user space, we need to do | |
4069 | the interrupt acknowledge cycle through it to fetch the next to be | |
4070 | delivered interrupt vector using this exit. | |
4071 | ||
4072 | It gets triggered whenever both KVM_CAP_PPC_EPR are enabled and an | |
4073 | external interrupt has just been delivered into the guest. User space | |
4074 | should put the acknowledged interrupt vector into the 'epr' field. | |
4075 | ||
4076 | /* KVM_EXIT_SYSTEM_EVENT */ | |
4077 | struct { | |
4078 | #define KVM_SYSTEM_EVENT_SHUTDOWN 1 | |
4079 | #define KVM_SYSTEM_EVENT_RESET 2 | |
4080 | #define KVM_SYSTEM_EVENT_CRASH 3 | |
4081 | __u32 type; | |
4082 | __u64 flags; | |
4083 | } system_event; | |
4084 | ||
4085 | If exit_reason is KVM_EXIT_SYSTEM_EVENT then the vcpu has triggered | |
4086 | a system-level event using some architecture specific mechanism (hypercall | |
4087 | or some special instruction). In case of ARM/ARM64, this is triggered using | |
4088 | HVC instruction based PSCI call from the vcpu. The 'type' field describes | |
4089 | the system-level event type. The 'flags' field describes architecture | |
4090 | specific flags for the system-level event. | |
4091 | ||
4092 | Valid values for 'type' are: | |
4093 | KVM_SYSTEM_EVENT_SHUTDOWN -- the guest has requested a shutdown of the | |
4094 | VM. Userspace is not obliged to honour this, and if it does honour | |
4095 | this does not need to destroy the VM synchronously (ie it may call | |
4096 | KVM_RUN again before shutdown finally occurs). | |
4097 | KVM_SYSTEM_EVENT_RESET -- the guest has requested a reset of the VM. | |
4098 | As with SHUTDOWN, userspace can choose to ignore the request, or | |
4099 | to schedule the reset to occur in the future and may call KVM_RUN again. | |
4100 | KVM_SYSTEM_EVENT_CRASH -- the guest crash occurred and the guest | |
4101 | has requested a crash condition maintenance. Userspace can choose | |
4102 | to ignore the request, or to gather VM memory core dump and/or | |
4103 | reset/shutdown of the VM. | |
4104 | ||
4105 | /* KVM_EXIT_IOAPIC_EOI */ | |
4106 | struct { | |
4107 | __u8 vector; | |
4108 | } eoi; | |
4109 | ||
4110 | Indicates that the VCPU's in-kernel local APIC received an EOI for a | |
4111 | level-triggered IOAPIC interrupt. This exit only triggers when the | |
4112 | IOAPIC is implemented in userspace (i.e. KVM_CAP_SPLIT_IRQCHIP is enabled); | |
4113 | the userspace IOAPIC should process the EOI and retrigger the interrupt if | |
4114 | it is still asserted. Vector is the LAPIC interrupt vector for which the | |
4115 | EOI was received. | |
4116 | ||
4117 | struct kvm_hyperv_exit { | |
4118 | #define KVM_EXIT_HYPERV_SYNIC 1 | |
4119 | #define KVM_EXIT_HYPERV_HCALL 2 | |
4120 | __u32 type; | |
4121 | union { | |
4122 | struct { | |
4123 | __u32 msr; | |
4124 | __u64 control; | |
4125 | __u64 evt_page; | |
4126 | __u64 msg_page; | |
4127 | } synic; | |
4128 | struct { | |
4129 | __u64 input; | |
4130 | __u64 result; | |
4131 | __u64 params[2]; | |
4132 | } hcall; | |
4133 | } u; | |
4134 | }; | |
4135 | /* KVM_EXIT_HYPERV */ | |
4136 | struct kvm_hyperv_exit hyperv; | |
4137 | Indicates that the VCPU exits into userspace to process some tasks | |
4138 | related to Hyper-V emulation. | |
4139 | Valid values for 'type' are: | |
4140 | KVM_EXIT_HYPERV_SYNIC -- synchronously notify user-space about | |
4141 | Hyper-V SynIC state change. Notification is used to remap SynIC | |
4142 | event/message pages and to enable/disable SynIC messages/events processing | |
4143 | in userspace. | |
4144 | ||
4145 | /* Fix the size of the union. */ | |
4146 | char padding[256]; | |
4147 | }; | |
4148 | ||
4149 | /* | |
4150 | * shared registers between kvm and userspace. | |
4151 | * kvm_valid_regs specifies the register classes set by the host | |
4152 | * kvm_dirty_regs specified the register classes dirtied by userspace | |
4153 | * struct kvm_sync_regs is architecture specific, as well as the | |
4154 | * bits for kvm_valid_regs and kvm_dirty_regs | |
4155 | */ | |
4156 | __u64 kvm_valid_regs; | |
4157 | __u64 kvm_dirty_regs; | |
4158 | union { | |
4159 | struct kvm_sync_regs regs; | |
4160 | char padding[SYNC_REGS_SIZE_BYTES]; | |
4161 | } s; | |
4162 | ||
4163 | If KVM_CAP_SYNC_REGS is defined, these fields allow userspace to access | |
4164 | certain guest registers without having to call SET/GET_*REGS. Thus we can | |
4165 | avoid some system call overhead if userspace has to handle the exit. | |
4166 | Userspace can query the validity of the structure by checking | |
4167 | kvm_valid_regs for specific bits. These bits are architecture specific | |
4168 | and usually define the validity of a groups of registers. (e.g. one bit | |
4169 | for general purpose registers) | |
4170 | ||
4171 | Please note that the kernel is allowed to use the kvm_run structure as the | |
4172 | primary storage for certain register types. Therefore, the kernel may use the | |
4173 | values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set. | |
4174 | ||
4175 | }; | |
4176 | ||
4177 | ||
4178 | ||
4179 | 6. Capabilities that can be enabled on vCPUs | |
4180 | -------------------------------------------- | |
4181 | ||
4182 | There are certain capabilities that change the behavior of the virtual CPU or | |
4183 | the virtual machine when enabled. To enable them, please see section 4.37. | |
4184 | Below you can find a list of capabilities and what their effect on the vCPU or | |
4185 | the virtual machine is when enabling them. | |
4186 | ||
4187 | The following information is provided along with the description: | |
4188 | ||
4189 | Architectures: which instruction set architectures provide this ioctl. | |
4190 | x86 includes both i386 and x86_64. | |
4191 | ||
4192 | Target: whether this is a per-vcpu or per-vm capability. | |
4193 | ||
4194 | Parameters: what parameters are accepted by the capability. | |
4195 | ||
4196 | Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL) | |
4197 | are not detailed, but errors with specific meanings are. | |
4198 | ||
4199 | ||
4200 | 6.1 KVM_CAP_PPC_OSI | |
4201 | ||
4202 | Architectures: ppc | |
4203 | Target: vcpu | |
4204 | Parameters: none | |
4205 | Returns: 0 on success; -1 on error | |
4206 | ||
4207 | This capability enables interception of OSI hypercalls that otherwise would | |
4208 | be treated as normal system calls to be injected into the guest. OSI hypercalls | |
4209 | were invented by Mac-on-Linux to have a standardized communication mechanism | |
4210 | between the guest and the host. | |
4211 | ||
4212 | When this capability is enabled, KVM_EXIT_OSI can occur. | |
4213 | ||
4214 | ||
4215 | 6.2 KVM_CAP_PPC_PAPR | |
4216 | ||
4217 | Architectures: ppc | |
4218 | Target: vcpu | |
4219 | Parameters: none | |
4220 | Returns: 0 on success; -1 on error | |
4221 | ||
4222 | This capability enables interception of PAPR hypercalls. PAPR hypercalls are | |
4223 | done using the hypercall instruction "sc 1". | |
4224 | ||
4225 | It also sets the guest privilege level to "supervisor" mode. Usually the guest | |
4226 | runs in "hypervisor" privilege mode with a few missing features. | |
4227 | ||
4228 | In addition to the above, it changes the semantics of SDR1. In this mode, the | |
4229 | HTAB address part of SDR1 contains an HVA instead of a GPA, as PAPR keeps the | |
4230 | HTAB invisible to the guest. | |
4231 | ||
4232 | When this capability is enabled, KVM_EXIT_PAPR_HCALL can occur. | |
4233 | ||
4234 | ||
4235 | 6.3 KVM_CAP_SW_TLB | |
4236 | ||
4237 | Architectures: ppc | |
4238 | Target: vcpu | |
4239 | Parameters: args[0] is the address of a struct kvm_config_tlb | |
4240 | Returns: 0 on success; -1 on error | |
4241 | ||
4242 | struct kvm_config_tlb { | |
4243 | __u64 params; | |
4244 | __u64 array; | |
4245 | __u32 mmu_type; | |
4246 | __u32 array_len; | |
4247 | }; | |
4248 | ||
4249 | Configures the virtual CPU's TLB array, establishing a shared memory area | |
4250 | between userspace and KVM. The "params" and "array" fields are userspace | |
4251 | addresses of mmu-type-specific data structures. The "array_len" field is an | |
4252 | safety mechanism, and should be set to the size in bytes of the memory that | |
4253 | userspace has reserved for the array. It must be at least the size dictated | |
4254 | by "mmu_type" and "params". | |
4255 | ||
4256 | While KVM_RUN is active, the shared region is under control of KVM. Its | |
4257 | contents are undefined, and any modification by userspace results in | |
4258 | boundedly undefined behavior. | |
4259 | ||
4260 | On return from KVM_RUN, the shared region will reflect the current state of | |
4261 | the guest's TLB. If userspace makes any changes, it must call KVM_DIRTY_TLB | |
4262 | to tell KVM which entries have been changed, prior to calling KVM_RUN again | |
4263 | on this vcpu. | |
4264 | ||
4265 | For mmu types KVM_MMU_FSL_BOOKE_NOHV and KVM_MMU_FSL_BOOKE_HV: | |
4266 | - The "params" field is of type "struct kvm_book3e_206_tlb_params". | |
4267 | - The "array" field points to an array of type "struct | |
4268 | kvm_book3e_206_tlb_entry". | |
4269 | - The array consists of all entries in the first TLB, followed by all | |
4270 | entries in the second TLB. | |
4271 | - Within a TLB, entries are ordered first by increasing set number. Within a | |
4272 | set, entries are ordered by way (increasing ESEL). | |
4273 | - The hash for determining set number in TLB0 is: (MAS2 >> 12) & (num_sets - 1) | |
4274 | where "num_sets" is the tlb_sizes[] value divided by the tlb_ways[] value. | |
4275 | - The tsize field of mas1 shall be set to 4K on TLB0, even though the | |
4276 | hardware ignores this value for TLB0. | |
4277 | ||
4278 | 6.4 KVM_CAP_S390_CSS_SUPPORT | |
4279 | ||
4280 | Architectures: s390 | |
4281 | Target: vcpu | |
4282 | Parameters: none | |
4283 | Returns: 0 on success; -1 on error | |
4284 | ||
4285 | This capability enables support for handling of channel I/O instructions. | |
4286 | ||
4287 | TEST PENDING INTERRUPTION and the interrupt portion of TEST SUBCHANNEL are | |
4288 | handled in-kernel, while the other I/O instructions are passed to userspace. | |
4289 | ||
4290 | When this capability is enabled, KVM_EXIT_S390_TSCH will occur on TEST | |
4291 | SUBCHANNEL intercepts. | |
4292 | ||
4293 | Note that even though this capability is enabled per-vcpu, the complete | |
4294 | virtual machine is affected. | |
4295 | ||
4296 | 6.5 KVM_CAP_PPC_EPR | |
4297 | ||
4298 | Architectures: ppc | |
4299 | Target: vcpu | |
4300 | Parameters: args[0] defines whether the proxy facility is active | |
4301 | Returns: 0 on success; -1 on error | |
4302 | ||
4303 | This capability enables or disables the delivery of interrupts through the | |
4304 | external proxy facility. | |
4305 | ||
4306 | When enabled (args[0] != 0), every time the guest gets an external interrupt | |
4307 | delivered, it automatically exits into user space with a KVM_EXIT_EPR exit | |
4308 | to receive the topmost interrupt vector. | |
4309 | ||
4310 | When disabled (args[0] == 0), behavior is as if this facility is unsupported. | |
4311 | ||
4312 | When this capability is enabled, KVM_EXIT_EPR can occur. | |
4313 | ||
4314 | 6.6 KVM_CAP_IRQ_MPIC | |
4315 | ||
4316 | Architectures: ppc | |
4317 | Parameters: args[0] is the MPIC device fd | |
4318 | args[1] is the MPIC CPU number for this vcpu | |
4319 | ||
4320 | This capability connects the vcpu to an in-kernel MPIC device. | |
4321 | ||
4322 | 6.7 KVM_CAP_IRQ_XICS | |
4323 | ||
4324 | Architectures: ppc | |
4325 | Target: vcpu | |
4326 | Parameters: args[0] is the XICS device fd | |
4327 | args[1] is the XICS CPU number (server ID) for this vcpu | |
4328 | ||
4329 | This capability connects the vcpu to an in-kernel XICS device. | |
4330 | ||
4331 | 6.8 KVM_CAP_S390_IRQCHIP | |
4332 | ||
4333 | Architectures: s390 | |
4334 | Target: vm | |
4335 | Parameters: none | |
4336 | ||
4337 | This capability enables the in-kernel irqchip for s390. Please refer to | |
4338 | "4.24 KVM_CREATE_IRQCHIP" for details. | |
4339 | ||
4340 | 6.9 KVM_CAP_MIPS_FPU | |
4341 | ||
4342 | Architectures: mips | |
4343 | Target: vcpu | |
4344 | Parameters: args[0] is reserved for future use (should be 0). | |
4345 | ||
4346 | This capability allows the use of the host Floating Point Unit by the guest. It | |
4347 | allows the Config1.FP bit to be set to enable the FPU in the guest. Once this is | |
4348 | done the KVM_REG_MIPS_FPR_* and KVM_REG_MIPS_FCR_* registers can be accessed | |
4349 | (depending on the current guest FPU register mode), and the Status.FR, | |
4350 | Config5.FRE bits are accessible via the KVM API and also from the guest, | |
4351 | depending on them being supported by the FPU. | |
4352 | ||
4353 | 6.10 KVM_CAP_MIPS_MSA | |
4354 | ||
4355 | Architectures: mips | |
4356 | Target: vcpu | |
4357 | Parameters: args[0] is reserved for future use (should be 0). | |
4358 | ||
4359 | This capability allows the use of the MIPS SIMD Architecture (MSA) by the guest. | |
4360 | It allows the Config3.MSAP bit to be set to enable the use of MSA by the guest. | |
4361 | Once this is done the KVM_REG_MIPS_VEC_* and KVM_REG_MIPS_MSA_* registers can be | |
4362 | accessed, and the Config5.MSAEn bit is accessible via the KVM API and also from | |
4363 | the guest. | |
4364 | ||
4365 | 6.74 KVM_CAP_SYNC_REGS | |
4366 | Architectures: s390, x86 | |
4367 | Target: s390: always enabled, x86: vcpu | |
4368 | Parameters: none | |
4369 | Returns: x86: KVM_CHECK_EXTENSION returns a bit-array indicating which register | |
4370 | sets are supported (bitfields defined in arch/x86/include/uapi/asm/kvm.h). | |
4371 | ||
4372 | As described above in the kvm_sync_regs struct info in section 5 (kvm_run): | |
4373 | KVM_CAP_SYNC_REGS "allow[s] userspace to access certain guest registers | |
4374 | without having to call SET/GET_*REGS". This reduces overhead by eliminating | |
4375 | repeated ioctl calls for setting and/or getting register values. This is | |
4376 | particularly important when userspace is making synchronous guest state | |
4377 | modifications, e.g. when emulating and/or intercepting instructions in | |
4378 | userspace. | |
4379 | ||
4380 | For s390 specifics, please refer to the source code. | |
4381 | ||
4382 | For x86: | |
4383 | - the register sets to be copied out to kvm_run are selectable | |
4384 | by userspace (rather that all sets being copied out for every exit). | |
4385 | - vcpu_events are available in addition to regs and sregs. | |
4386 | ||
4387 | For x86, the 'kvm_valid_regs' field of struct kvm_run is overloaded to | |
4388 | function as an input bit-array field set by userspace to indicate the | |
4389 | specific register sets to be copied out on the next exit. | |
4390 | ||
4391 | To indicate when userspace has modified values that should be copied into | |
4392 | the vCPU, the all architecture bitarray field, 'kvm_dirty_regs' must be set. | |
4393 | This is done using the same bitflags as for the 'kvm_valid_regs' field. | |
4394 | If the dirty bit is not set, then the register set values will not be copied | |
4395 | into the vCPU even if they've been modified. | |
4396 | ||
4397 | Unused bitfields in the bitarrays must be set to zero. | |
4398 | ||
4399 | struct kvm_sync_regs { | |
4400 | struct kvm_regs regs; | |
4401 | struct kvm_sregs sregs; | |
4402 | struct kvm_vcpu_events events; | |
4403 | }; | |
4404 | ||
4405 | 7. Capabilities that can be enabled on VMs | |
4406 | ------------------------------------------ | |
4407 | ||
4408 | There are certain capabilities that change the behavior of the virtual | |
4409 | machine when enabled. To enable them, please see section 4.37. Below | |
4410 | you can find a list of capabilities and what their effect on the VM | |
4411 | is when enabling them. | |
4412 | ||
4413 | The following information is provided along with the description: | |
4414 | ||
4415 | Architectures: which instruction set architectures provide this ioctl. | |
4416 | x86 includes both i386 and x86_64. | |
4417 | ||
4418 | Parameters: what parameters are accepted by the capability. | |
4419 | ||
4420 | Returns: the return value. General error numbers (EBADF, ENOMEM, EINVAL) | |
4421 | are not detailed, but errors with specific meanings are. | |
4422 | ||
4423 | ||
4424 | 7.1 KVM_CAP_PPC_ENABLE_HCALL | |
4425 | ||
4426 | Architectures: ppc | |
4427 | Parameters: args[0] is the sPAPR hcall number | |
4428 | args[1] is 0 to disable, 1 to enable in-kernel handling | |
4429 | ||
4430 | This capability controls whether individual sPAPR hypercalls (hcalls) | |
4431 | get handled by the kernel or not. Enabling or disabling in-kernel | |
4432 | handling of an hcall is effective across the VM. On creation, an | |
4433 | initial set of hcalls are enabled for in-kernel handling, which | |
4434 | consists of those hcalls for which in-kernel handlers were implemented | |
4435 | before this capability was implemented. If disabled, the kernel will | |
4436 | not to attempt to handle the hcall, but will always exit to userspace | |
4437 | to handle it. Note that it may not make sense to enable some and | |
4438 | disable others of a group of related hcalls, but KVM does not prevent | |
4439 | userspace from doing that. | |
4440 | ||
4441 | If the hcall number specified is not one that has an in-kernel | |
4442 | implementation, the KVM_ENABLE_CAP ioctl will fail with an EINVAL | |
4443 | error. | |
4444 | ||
4445 | 7.2 KVM_CAP_S390_USER_SIGP | |
4446 | ||
4447 | Architectures: s390 | |
4448 | Parameters: none | |
4449 | ||
4450 | This capability controls which SIGP orders will be handled completely in user | |
4451 | space. With this capability enabled, all fast orders will be handled completely | |
4452 | in the kernel: | |
4453 | - SENSE | |
4454 | - SENSE RUNNING | |
4455 | - EXTERNAL CALL | |
4456 | - EMERGENCY SIGNAL | |
4457 | - CONDITIONAL EMERGENCY SIGNAL | |
4458 | ||
4459 | All other orders will be handled completely in user space. | |
4460 | ||
4461 | Only privileged operation exceptions will be checked for in the kernel (or even | |
4462 | in the hardware prior to interception). If this capability is not enabled, the | |
4463 | old way of handling SIGP orders is used (partially in kernel and user space). | |
4464 | ||
4465 | 7.3 KVM_CAP_S390_VECTOR_REGISTERS | |
4466 | ||
4467 | Architectures: s390 | |
4468 | Parameters: none | |
4469 | Returns: 0 on success, negative value on error | |
4470 | ||
4471 | Allows use of the vector registers introduced with z13 processor, and | |
4472 | provides for the synchronization between host and user space. Will | |
4473 | return -EINVAL if the machine does not support vectors. | |
4474 | ||
4475 | 7.4 KVM_CAP_S390_USER_STSI | |
4476 | ||
4477 | Architectures: s390 | |
4478 | Parameters: none | |
4479 | ||
4480 | This capability allows post-handlers for the STSI instruction. After | |
4481 | initial handling in the kernel, KVM exits to user space with | |
4482 | KVM_EXIT_S390_STSI to allow user space to insert further data. | |
4483 | ||
4484 | Before exiting to userspace, kvm handlers should fill in s390_stsi field of | |
4485 | vcpu->run: | |
4486 | struct { | |
4487 | __u64 addr; | |
4488 | __u8 ar; | |
4489 | __u8 reserved; | |
4490 | __u8 fc; | |
4491 | __u8 sel1; | |
4492 | __u16 sel2; | |
4493 | } s390_stsi; | |
4494 | ||
4495 | @addr - guest address of STSI SYSIB | |
4496 | @fc - function code | |
4497 | @sel1 - selector 1 | |
4498 | @sel2 - selector 2 | |
4499 | @ar - access register number | |
4500 | ||
4501 | KVM handlers should exit to userspace with rc = -EREMOTE. | |
4502 | ||
4503 | 7.5 KVM_CAP_SPLIT_IRQCHIP | |
4504 | ||
4505 | Architectures: x86 | |
4506 | Parameters: args[0] - number of routes reserved for userspace IOAPICs | |
4507 | Returns: 0 on success, -1 on error | |
4508 | ||
4509 | Create a local apic for each processor in the kernel. This can be used | |
4510 | instead of KVM_CREATE_IRQCHIP if the userspace VMM wishes to emulate the | |
4511 | IOAPIC and PIC (and also the PIT, even though this has to be enabled | |
4512 | separately). | |
4513 | ||
4514 | This capability also enables in kernel routing of interrupt requests; | |
4515 | when KVM_CAP_SPLIT_IRQCHIP only routes of KVM_IRQ_ROUTING_MSI type are | |
4516 | used in the IRQ routing table. The first args[0] MSI routes are reserved | |
4517 | for the IOAPIC pins. Whenever the LAPIC receives an EOI for these routes, | |
4518 | a KVM_EXIT_IOAPIC_EOI vmexit will be reported to userspace. | |
4519 | ||
4520 | Fails if VCPU has already been created, or if the irqchip is already in the | |
4521 | kernel (i.e. KVM_CREATE_IRQCHIP has already been called). | |
4522 | ||
4523 | 7.6 KVM_CAP_S390_RI | |
4524 | ||
4525 | Architectures: s390 | |
4526 | Parameters: none | |
4527 | ||
4528 | Allows use of runtime-instrumentation introduced with zEC12 processor. | |
4529 | Will return -EINVAL if the machine does not support runtime-instrumentation. | |
4530 | Will return -EBUSY if a VCPU has already been created. | |
4531 | ||
4532 | 7.7 KVM_CAP_X2APIC_API | |
4533 | ||
4534 | Architectures: x86 | |
4535 | Parameters: args[0] - features that should be enabled | |
4536 | Returns: 0 on success, -EINVAL when args[0] contains invalid features | |
4537 | ||
4538 | Valid feature flags in args[0] are | |
4539 | ||
4540 | #define KVM_X2APIC_API_USE_32BIT_IDS (1ULL << 0) | |
4541 | #define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK (1ULL << 1) | |
4542 | ||
4543 | Enabling KVM_X2APIC_API_USE_32BIT_IDS changes the behavior of | |
4544 | KVM_SET_GSI_ROUTING, KVM_SIGNAL_MSI, KVM_SET_LAPIC, and KVM_GET_LAPIC, | |
4545 | allowing the use of 32-bit APIC IDs. See KVM_CAP_X2APIC_API in their | |
4546 | respective sections. | |
4547 | ||
4548 | KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK must be enabled for x2APIC to work | |
4549 | in logical mode or with more than 255 VCPUs. Otherwise, KVM treats 0xff | |
4550 | as a broadcast even in x2APIC mode in order to support physical x2APIC | |
4551 | without interrupt remapping. This is undesirable in logical mode, | |
4552 | where 0xff represents CPUs 0-7 in cluster 0. | |
4553 | ||
4554 | 7.8 KVM_CAP_S390_USER_INSTR0 | |
4555 | ||
4556 | Architectures: s390 | |
4557 | Parameters: none | |
4558 | ||
4559 | With this capability enabled, all illegal instructions 0x0000 (2 bytes) will | |
4560 | be intercepted and forwarded to user space. User space can use this | |
4561 | mechanism e.g. to realize 2-byte software breakpoints. The kernel will | |
4562 | not inject an operating exception for these instructions, user space has | |
4563 | to take care of that. | |
4564 | ||
4565 | This capability can be enabled dynamically even if VCPUs were already | |
4566 | created and are running. | |
4567 | ||
4568 | 7.9 KVM_CAP_S390_GS | |
4569 | ||
4570 | Architectures: s390 | |
4571 | Parameters: none | |
4572 | Returns: 0 on success; -EINVAL if the machine does not support | |
4573 | guarded storage; -EBUSY if a VCPU has already been created. | |
4574 | ||
4575 | Allows use of guarded storage for the KVM guest. | |
4576 | ||
4577 | 7.10 KVM_CAP_S390_AIS | |
4578 | ||
4579 | Architectures: s390 | |
4580 | Parameters: none | |
4581 | ||
4582 | Allow use of adapter-interruption suppression. | |
4583 | Returns: 0 on success; -EBUSY if a VCPU has already been created. | |
4584 | ||
4585 | 7.11 KVM_CAP_PPC_SMT | |
4586 | ||
4587 | Architectures: ppc | |
4588 | Parameters: vsmt_mode, flags | |
4589 | ||
4590 | Enabling this capability on a VM provides userspace with a way to set | |
4591 | the desired virtual SMT mode (i.e. the number of virtual CPUs per | |
4592 | virtual core). The virtual SMT mode, vsmt_mode, must be a power of 2 | |
4593 | between 1 and 8. On POWER8, vsmt_mode must also be no greater than | |
4594 | the number of threads per subcore for the host. Currently flags must | |
4595 | be 0. A successful call to enable this capability will result in | |
4596 | vsmt_mode being returned when the KVM_CAP_PPC_SMT capability is | |
4597 | subsequently queried for the VM. This capability is only supported by | |
4598 | HV KVM, and can only be set before any VCPUs have been created. | |
4599 | The KVM_CAP_PPC_SMT_POSSIBLE capability indicates which virtual SMT | |
4600 | modes are available. | |
4601 | ||
4602 | 7.12 KVM_CAP_PPC_FWNMI | |
4603 | ||
4604 | Architectures: ppc | |
4605 | Parameters: none | |
4606 | ||
4607 | With this capability a machine check exception in the guest address | |
4608 | space will cause KVM to exit the guest with NMI exit reason. This | |
4609 | enables QEMU to build error log and branch to guest kernel registered | |
4610 | machine check handling routine. Without this capability KVM will | |
4611 | branch to guests' 0x200 interrupt vector. | |
4612 | ||
4613 | 7.13 KVM_CAP_X86_DISABLE_EXITS | |
4614 | ||
4615 | Architectures: x86 | |
4616 | Parameters: args[0] defines which exits are disabled | |
4617 | Returns: 0 on success, -EINVAL when args[0] contains invalid exits | |
4618 | ||
4619 | Valid bits in args[0] are | |
4620 | ||
4621 | #define KVM_X86_DISABLE_EXITS_MWAIT (1 << 0) | |
4622 | #define KVM_X86_DISABLE_EXITS_HLT (1 << 1) | |
4623 | ||
4624 | Enabling this capability on a VM provides userspace with a way to no | |
4625 | longer intercept some instructions for improved latency in some | |
4626 | workloads, and is suggested when vCPUs are associated to dedicated | |
4627 | physical CPUs. More bits can be added in the future; userspace can | |
4628 | just pass the KVM_CHECK_EXTENSION result to KVM_ENABLE_CAP to disable | |
4629 | all such vmexits. | |
4630 | ||
4631 | Do not enable KVM_FEATURE_PV_UNHALT if you disable HLT exits. | |
4632 | ||
4633 | 7.14 KVM_CAP_S390_HPAGE_1M | |
4634 | ||
4635 | Architectures: s390 | |
4636 | Parameters: none | |
4637 | Returns: 0 on success, -EINVAL if hpage module parameter was not set | |
4638 | or cmma is enabled, or the VM has the KVM_VM_S390_UCONTROL | |
4639 | flag set | |
4640 | ||
4641 | With this capability the KVM support for memory backing with 1m pages | |
4642 | through hugetlbfs can be enabled for a VM. After the capability is | |
4643 | enabled, cmma can't be enabled anymore and pfmfi and the storage key | |
4644 | interpretation are disabled. If cmma has already been enabled or the | |
4645 | hpage module parameter is not set to 1, -EINVAL is returned. | |
4646 | ||
4647 | While it is generally possible to create a huge page backed VM without | |
4648 | this capability, the VM will not be able to run. | |
4649 | ||
4650 | 7.15 KVM_CAP_MSR_PLATFORM_INFO | |
4651 | ||
4652 | Architectures: x86 | |
4653 | Parameters: args[0] whether feature should be enabled or not | |
4654 | ||
4655 | With this capability, a guest may read the MSR_PLATFORM_INFO MSR. Otherwise, | |
4656 | a #GP would be raised when the guest tries to access. Currently, this | |
4657 | capability does not enable write permissions of this MSR for the guest. | |
4658 | ||
4659 | 7.16 KVM_CAP_PPC_NESTED_HV | |
4660 | ||
4661 | Architectures: ppc | |
4662 | Parameters: none | |
4663 | Returns: 0 on success, -EINVAL when the implementation doesn't support | |
4664 | nested-HV virtualization. | |
4665 | ||
4666 | HV-KVM on POWER9 and later systems allows for "nested-HV" | |
4667 | virtualization, which provides a way for a guest VM to run guests that | |
4668 | can run using the CPU's supervisor mode (privileged non-hypervisor | |
4669 | state). Enabling this capability on a VM depends on the CPU having | |
4670 | the necessary functionality and on the facility being enabled with a | |
4671 | kvm-hv module parameter. | |
4672 | ||
4673 | 7.17 KVM_CAP_EXCEPTION_PAYLOAD | |
4674 | ||
4675 | Architectures: x86 | |
4676 | Parameters: args[0] whether feature should be enabled or not | |
4677 | ||
4678 | With this capability enabled, CR2 will not be modified prior to the | |
4679 | emulated VM-exit when L1 intercepts a #PF exception that occurs in | |
4680 | L2. Similarly, for kvm-intel only, DR6 will not be modified prior to | |
4681 | the emulated VM-exit when L1 intercepts a #DB exception that occurs in | |
4682 | L2. As a result, when KVM_GET_VCPU_EVENTS reports a pending #PF (or | |
4683 | #DB) exception for L2, exception.has_payload will be set and the | |
4684 | faulting address (or the new DR6 bits*) will be reported in the | |
4685 | exception_payload field. Similarly, when userspace injects a #PF (or | |
4686 | #DB) into L2 using KVM_SET_VCPU_EVENTS, it is expected to set | |
4687 | exception.has_payload and to put the faulting address (or the new DR6 | |
4688 | bits*) in the exception_payload field. | |
4689 | ||
4690 | This capability also enables exception.pending in struct | |
4691 | kvm_vcpu_events, which allows userspace to distinguish between pending | |
4692 | and injected exceptions. | |
4693 | ||
4694 | ||
4695 | * For the new DR6 bits, note that bit 16 is set iff the #DB exception | |
4696 | will clear DR6.RTM. | |
4697 | ||
4698 | 7.18 KVM_CAP_MANUAL_DIRTY_LOG_PROTECT | |
4699 | ||
4700 | Architectures: all | |
4701 | Parameters: args[0] whether feature should be enabled or not | |
4702 | ||
4703 | With this capability enabled, KVM_GET_DIRTY_LOG will not automatically | |
4704 | clear and write-protect all pages that are returned as dirty. | |
4705 | Rather, userspace will have to do this operation separately using | |
4706 | KVM_CLEAR_DIRTY_LOG. | |
4707 | ||
4708 | At the cost of a slightly more complicated operation, this provides better | |
4709 | scalability and responsiveness for two reasons. First, | |
4710 | KVM_CLEAR_DIRTY_LOG ioctl can operate on a 64-page granularity rather | |
4711 | than requiring to sync a full memslot; this ensures that KVM does not | |
4712 | take spinlocks for an extended period of time. Second, in some cases a | |
4713 | large amount of time can pass between a call to KVM_GET_DIRTY_LOG and | |
4714 | userspace actually using the data in the page. Pages can be modified | |
4715 | during this time, which is inefficint for both the guest and userspace: | |
4716 | the guest will incur a higher penalty due to write protection faults, | |
4717 | while userspace can see false reports of dirty pages. Manual reprotection | |
4718 | helps reducing this time, improving guest performance and reducing the | |
4719 | number of dirty log false positives. | |
4720 | ||
4721 | ||
4722 | 8. Other capabilities. | |
4723 | ---------------------- | |
4724 | ||
4725 | This section lists capabilities that give information about other | |
4726 | features of the KVM implementation. | |
4727 | ||
4728 | 8.1 KVM_CAP_PPC_HWRNG | |
4729 | ||
4730 | Architectures: ppc | |
4731 | ||
4732 | This capability, if KVM_CHECK_EXTENSION indicates that it is | |
4733 | available, means that that the kernel has an implementation of the | |
4734 | H_RANDOM hypercall backed by a hardware random-number generator. | |
4735 | If present, the kernel H_RANDOM handler can be enabled for guest use | |
4736 | with the KVM_CAP_PPC_ENABLE_HCALL capability. | |
4737 | ||
4738 | 8.2 KVM_CAP_HYPERV_SYNIC | |
4739 | ||
4740 | Architectures: x86 | |
4741 | This capability, if KVM_CHECK_EXTENSION indicates that it is | |
4742 | available, means that that the kernel has an implementation of the | |
4743 | Hyper-V Synthetic interrupt controller(SynIC). Hyper-V SynIC is | |
4744 | used to support Windows Hyper-V based guest paravirt drivers(VMBus). | |
4745 | ||
4746 | In order to use SynIC, it has to be activated by setting this | |
4747 | capability via KVM_ENABLE_CAP ioctl on the vcpu fd. Note that this | |
4748 | will disable the use of APIC hardware virtualization even if supported | |
4749 | by the CPU, as it's incompatible with SynIC auto-EOI behavior. | |
4750 | ||
4751 | 8.3 KVM_CAP_PPC_RADIX_MMU | |
4752 | ||
4753 | Architectures: ppc | |
4754 | ||
4755 | This capability, if KVM_CHECK_EXTENSION indicates that it is | |
4756 | available, means that that the kernel can support guests using the | |
4757 | radix MMU defined in Power ISA V3.00 (as implemented in the POWER9 | |
4758 | processor). | |
4759 | ||
4760 | 8.4 KVM_CAP_PPC_HASH_MMU_V3 | |
4761 | ||
4762 | Architectures: ppc | |
4763 | ||
4764 | This capability, if KVM_CHECK_EXTENSION indicates that it is | |
4765 | available, means that that the kernel can support guests using the | |
4766 | hashed page table MMU defined in Power ISA V3.00 (as implemented in | |
4767 | the POWER9 processor), including in-memory segment tables. | |
4768 | ||
4769 | 8.5 KVM_CAP_MIPS_VZ | |
4770 | ||
4771 | Architectures: mips | |
4772 | ||
4773 | This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that | |
4774 | it is available, means that full hardware assisted virtualization capabilities | |
4775 | of the hardware are available for use through KVM. An appropriate | |
4776 | KVM_VM_MIPS_* type must be passed to KVM_CREATE_VM to create a VM which | |
4777 | utilises it. | |
4778 | ||
4779 | If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is | |
4780 | available, it means that the VM is using full hardware assisted virtualization | |
4781 | capabilities of the hardware. This is useful to check after creating a VM with | |
4782 | KVM_VM_MIPS_DEFAULT. | |
4783 | ||
4784 | The value returned by KVM_CHECK_EXTENSION should be compared against known | |
4785 | values (see below). All other values are reserved. This is to allow for the | |
4786 | possibility of other hardware assisted virtualization implementations which | |
4787 | may be incompatible with the MIPS VZ ASE. | |
4788 | ||
4789 | 0: The trap & emulate implementation is in use to run guest code in user | |
4790 | mode. Guest virtual memory segments are rearranged to fit the guest in the | |
4791 | user mode address space. | |
4792 | ||
4793 | 1: The MIPS VZ ASE is in use, providing full hardware assisted | |
4794 | virtualization, including standard guest virtual memory segments. | |
4795 | ||
4796 | 8.6 KVM_CAP_MIPS_TE | |
4797 | ||
4798 | Architectures: mips | |
4799 | ||
4800 | This capability, if KVM_CHECK_EXTENSION on the main kvm handle indicates that | |
4801 | it is available, means that the trap & emulate implementation is available to | |
4802 | run guest code in user mode, even if KVM_CAP_MIPS_VZ indicates that hardware | |
4803 | assisted virtualisation is also available. KVM_VM_MIPS_TE (0) must be passed | |
4804 | to KVM_CREATE_VM to create a VM which utilises it. | |
4805 | ||
4806 | If KVM_CHECK_EXTENSION on a kvm VM handle indicates that this capability is | |
4807 | available, it means that the VM is using trap & emulate. | |
4808 | ||
4809 | 8.7 KVM_CAP_MIPS_64BIT | |
4810 | ||
4811 | Architectures: mips | |
4812 | ||
4813 | This capability indicates the supported architecture type of the guest, i.e. the | |
4814 | supported register and address width. | |
4815 | ||
4816 | The values returned when this capability is checked by KVM_CHECK_EXTENSION on a | |
4817 | kvm VM handle correspond roughly to the CP0_Config.AT register field, and should | |
4818 | be checked specifically against known values (see below). All other values are | |
4819 | reserved. | |
4820 | ||
4821 | 0: MIPS32 or microMIPS32. | |
4822 | Both registers and addresses are 32-bits wide. | |
4823 | It will only be possible to run 32-bit guest code. | |
4824 | ||
4825 | 1: MIPS64 or microMIPS64 with access only to 32-bit compatibility segments. | |
4826 | Registers are 64-bits wide, but addresses are 32-bits wide. | |
4827 | 64-bit guest code may run but cannot access MIPS64 memory segments. | |
4828 | It will also be possible to run 32-bit guest code. | |
4829 | ||
4830 | 2: MIPS64 or microMIPS64 with access to all address segments. | |
4831 | Both registers and addresses are 64-bits wide. | |
4832 | It will be possible to run 64-bit or 32-bit guest code. | |
4833 | ||
4834 | 8.9 KVM_CAP_ARM_USER_IRQ | |
4835 | ||
4836 | Architectures: arm, arm64 | |
4837 | This capability, if KVM_CHECK_EXTENSION indicates that it is available, means | |
4838 | that if userspace creates a VM without an in-kernel interrupt controller, it | |
4839 | will be notified of changes to the output level of in-kernel emulated devices, | |
4840 | which can generate virtual interrupts, presented to the VM. | |
4841 | For such VMs, on every return to userspace, the kernel | |
4842 | updates the vcpu's run->s.regs.device_irq_level field to represent the actual | |
4843 | output level of the device. | |
4844 | ||
4845 | Whenever kvm detects a change in the device output level, kvm guarantees at | |
4846 | least one return to userspace before running the VM. This exit could either | |
4847 | be a KVM_EXIT_INTR or any other exit event, like KVM_EXIT_MMIO. This way, | |
4848 | userspace can always sample the device output level and re-compute the state of | |
4849 | the userspace interrupt controller. Userspace should always check the state | |
4850 | of run->s.regs.device_irq_level on every kvm exit. | |
4851 | The value in run->s.regs.device_irq_level can represent both level and edge | |
4852 | triggered interrupt signals, depending on the device. Edge triggered interrupt | |
4853 | signals will exit to userspace with the bit in run->s.regs.device_irq_level | |
4854 | set exactly once per edge signal. | |
4855 | ||
4856 | The field run->s.regs.device_irq_level is available independent of | |
4857 | run->kvm_valid_regs or run->kvm_dirty_regs bits. | |
4858 | ||
4859 | If KVM_CAP_ARM_USER_IRQ is supported, the KVM_CHECK_EXTENSION ioctl returns a | |
4860 | number larger than 0 indicating the version of this capability is implemented | |
4861 | and thereby which bits in in run->s.regs.device_irq_level can signal values. | |
4862 | ||
4863 | Currently the following bits are defined for the device_irq_level bitmap: | |
4864 | ||
4865 | KVM_CAP_ARM_USER_IRQ >= 1: | |
4866 | ||
4867 | KVM_ARM_DEV_EL1_VTIMER - EL1 virtual timer | |
4868 | KVM_ARM_DEV_EL1_PTIMER - EL1 physical timer | |
4869 | KVM_ARM_DEV_PMU - ARM PMU overflow interrupt signal | |
4870 | ||
4871 | Future versions of kvm may implement additional events. These will get | |
4872 | indicated by returning a higher number from KVM_CHECK_EXTENSION and will be | |
4873 | listed above. | |
4874 | ||
4875 | 8.10 KVM_CAP_PPC_SMT_POSSIBLE | |
4876 | ||
4877 | Architectures: ppc | |
4878 | ||
4879 | Querying this capability returns a bitmap indicating the possible | |
4880 | virtual SMT modes that can be set using KVM_CAP_PPC_SMT. If bit N | |
4881 | (counting from the right) is set, then a virtual SMT mode of 2^N is | |
4882 | available. | |
4883 | ||
4884 | 8.11 KVM_CAP_HYPERV_SYNIC2 | |
4885 | ||
4886 | Architectures: x86 | |
4887 | ||
4888 | This capability enables a newer version of Hyper-V Synthetic interrupt | |
4889 | controller (SynIC). The only difference with KVM_CAP_HYPERV_SYNIC is that KVM | |
4890 | doesn't clear SynIC message and event flags pages when they are enabled by | |
4891 | writing to the respective MSRs. | |
4892 | ||
4893 | 8.12 KVM_CAP_HYPERV_VP_INDEX | |
4894 | ||
4895 | Architectures: x86 | |
4896 | ||
4897 | This capability indicates that userspace can load HV_X64_MSR_VP_INDEX msr. Its | |
4898 | value is used to denote the target vcpu for a SynIC interrupt. For | |
4899 | compatibilty, KVM initializes this msr to KVM's internal vcpu index. When this | |
4900 | capability is absent, userspace can still query this msr's value. | |
4901 | ||
4902 | 8.13 KVM_CAP_S390_AIS_MIGRATION | |
4903 | ||
4904 | Architectures: s390 | |
4905 | Parameters: none | |
4906 | ||
4907 | This capability indicates if the flic device will be able to get/set the | |
4908 | AIS states for migration via the KVM_DEV_FLIC_AISM_ALL attribute and allows | |
4909 | to discover this without having to create a flic device. | |
4910 | ||
4911 | 8.14 KVM_CAP_S390_PSW | |
4912 | ||
4913 | Architectures: s390 | |
4914 | ||
4915 | This capability indicates that the PSW is exposed via the kvm_run structure. | |
4916 | ||
4917 | 8.15 KVM_CAP_S390_GMAP | |
4918 | ||
4919 | Architectures: s390 | |
4920 | ||
4921 | This capability indicates that the user space memory used as guest mapping can | |
4922 | be anywhere in the user memory address space, as long as the memory slots are | |
4923 | aligned and sized to a segment (1MB) boundary. | |
4924 | ||
4925 | 8.16 KVM_CAP_S390_COW | |
4926 | ||
4927 | Architectures: s390 | |
4928 | ||
4929 | This capability indicates that the user space memory used as guest mapping can | |
4930 | use copy-on-write semantics as well as dirty pages tracking via read-only page | |
4931 | tables. | |
4932 | ||
4933 | 8.17 KVM_CAP_S390_BPB | |
4934 | ||
4935 | Architectures: s390 | |
4936 | ||
4937 | This capability indicates that kvm will implement the interfaces to handle | |
4938 | reset, migration and nested KVM for branch prediction blocking. The stfle | |
4939 | facility 82 should not be provided to the guest without this capability. | |
4940 | ||
4941 | 8.18 KVM_CAP_HYPERV_TLBFLUSH | |
4942 | ||
4943 | Architectures: x86 | |
4944 | ||
4945 | This capability indicates that KVM supports paravirtualized Hyper-V TLB Flush | |
4946 | hypercalls: | |
4947 | HvFlushVirtualAddressSpace, HvFlushVirtualAddressSpaceEx, | |
4948 | HvFlushVirtualAddressList, HvFlushVirtualAddressListEx. | |
4949 | ||
4950 | 8.19 KVM_CAP_ARM_INJECT_SERROR_ESR | |
4951 | ||
4952 | Architectures: arm, arm64 | |
4953 | ||
4954 | This capability indicates that userspace can specify (via the | |
4955 | KVM_SET_VCPU_EVENTS ioctl) the syndrome value reported to the guest when it | |
4956 | takes a virtual SError interrupt exception. | |
4957 | If KVM advertises this capability, userspace can only specify the ISS field for | |
4958 | the ESR syndrome. Other parts of the ESR, such as the EC are generated by the | |
4959 | CPU when the exception is taken. If this virtual SError is taken to EL1 using | |
4960 | AArch64, this value will be reported in the ISS field of ESR_ELx. | |
4961 | ||
4962 | See KVM_CAP_VCPU_EVENTS for more details. | |
4963 | 8.20 KVM_CAP_HYPERV_SEND_IPI | |
4964 | ||
4965 | Architectures: x86 | |
4966 | ||
4967 | This capability indicates that KVM supports paravirtualized Hyper-V IPI send | |
4968 | hypercalls: | |
4969 | HvCallSendSyntheticClusterIpi, HvCallSendSyntheticClusterIpiEx. |