4 Copyright (c) 2014 Virtual Open Systems Sarl.
6 This work is licensed under the terms of the GNU GPL, version 2 or later.
7 See the COPYING file in the top-level directory.
10 This protocol is aiming to complement the ioctl interface used to control the
11 vhost implementation in the Linux kernel. It implements the control plane needed
12 to establish virtqueue sharing with a user space process on the same host. It
13 uses communication over a Unix domain socket to share file descriptors in the
14 ancillary data of the message.
16 The protocol defines 2 sides of the communication, master and slave. Master is
17 the application that shares its virtqueues, in our case QEMU. Slave is the
18 consumer of the virtqueues.
20 In the current implementation QEMU is the Master, and the Slave is intended to
21 be a software Ethernet switch running in user space, such as Snabbswitch.
23 Master and slave can be either a client (i.e. connecting) or server (listening)
24 in the socket communication.
29 Note that all numbers are in the machine native byte order. A vhost-user message
30 consists of 3 header fields and a payload:
32 ------------------------------------
33 | request | flags | size | payload |
34 ------------------------------------
36 * Request: 32-bit type of the request
37 * Flags: 32-bit bit field:
38 - Lower 2 bits are the version (currently 0x01)
39 - Bit 2 is the reply flag - needs to be sent on each reply from the slave
40 - Bit 3 is the need_reply flag - see VHOST_USER_PROTOCOL_F_REPLY_ACK for
42 * Size - 32-bit size of the payload
45 Depending on the request type, payload can be:
47 * A single 64-bit integer
52 u64: a 64-bit unsigned integer
54 * A vring state description
62 * A vring address description
63 --------------------------------------------------------------
64 | index | flags | size | descriptor | used | available | log |
65 --------------------------------------------------------------
67 Index: a 32-bit vring index
68 Flags: a 32-bit vring flags
69 Descriptor: a 64-bit ring address of the vring descriptor table
70 Used: a 64-bit ring address of the vring used ring
71 Available: a 64-bit ring address of the vring available ring
72 Log: a 64-bit guest address for logging
74 Note that a ring address is an IOVA if VIRTIO_F_IOMMU_PLATFORM has been
75 negotiated. Otherwise it is a user address.
77 * Memory regions description
78 ---------------------------------------------------
79 | num regions | padding | region0 | ... | region7 |
80 ---------------------------------------------------
82 Num regions: a 32-bit number of regions
86 -----------------------------------------------------
87 | guest address | size | user address | mmap offset |
88 -----------------------------------------------------
90 Guest address: a 64-bit guest address of the region
92 User address: a 64-bit user address
93 mmap offset: 64-bit offset where region starts in the mapped memory
96 ---------------------------
97 | log size | log offset |
98 ---------------------------
99 log size: size of area used for logging
100 log offset: offset from start of supplied file descriptor
101 where logging starts (i.e. where guest address 0 would be logged)
104 ---------------------------------------------------------
105 | iova | size | user address | permissions flags | type |
106 ---------------------------------------------------------
108 IOVA: a 64-bit I/O virtual address programmed by the guest
110 User address: a 64-bit user address
111 Permissions: a 8-bit value:
115 - 3: Read/Write access
116 Type: a 8-bit IOTLB message type:
119 - 3: IOTLB invalidate
120 - 4: IOTLB access fail
122 * Virtio device config space
123 -----------------------------------
124 | offset | size | flags | payload |
125 -----------------------------------
127 Offset: a 32-bit offset of virtio device's configuration space
128 Size: a 32-bit configuration space access size in bytes
129 Flags: a 32-bit value:
130 - 0: Vhost master messages used for writeable fields
131 - 1: Vhost master messages used for live migration
132 Payload: Size bytes array holding the contents of the virtio
133 device's configuration space
135 In QEMU the vhost-user message is implemented with the following struct:
137 typedef struct VhostUserMsg {
138 VhostUserRequest request;
143 struct vhost_vring_state state;
144 struct vhost_vring_addr addr;
145 VhostUserMemory memory;
147 struct vhost_iotlb_msg iotlb;
148 VhostUserConfig config;
150 } QEMU_PACKED VhostUserMsg;
155 The protocol for vhost-user is based on the existing implementation of vhost
156 for the Linux Kernel. Most messages that can be sent via the Unix domain socket
157 implementing vhost-user have an equivalent ioctl to the kernel implementation.
159 The communication consists of master sending message requests and slave sending
160 message replies. Most of the requests don't require replies. Here is a list of
163 * VHOST_USER_GET_FEATURES
164 * VHOST_USER_GET_PROTOCOL_FEATURES
165 * VHOST_USER_GET_VRING_BASE
166 * VHOST_USER_SET_LOG_BASE (if VHOST_USER_PROTOCOL_F_LOG_SHMFD)
168 [ Also see the section on REPLY_ACK protocol extension. ]
170 There are several messages that the master sends with file descriptors passed
171 in the ancillary data:
173 * VHOST_USER_SET_MEM_TABLE
174 * VHOST_USER_SET_LOG_BASE (if VHOST_USER_PROTOCOL_F_LOG_SHMFD)
175 * VHOST_USER_SET_LOG_FD
176 * VHOST_USER_SET_VRING_KICK
177 * VHOST_USER_SET_VRING_CALL
178 * VHOST_USER_SET_VRING_ERR
179 * VHOST_USER_SET_SLAVE_REQ_FD
181 If Master is unable to send the full message or receives a wrong reply it will
182 close the connection. An optional reconnection mechanism can be implemented.
184 Any protocol extensions are gated by protocol feature bits,
185 which allows full backwards compatibility on both master
187 As older slaves don't support negotiating protocol features,
188 a feature bit was dedicated for this purpose:
189 #define VHOST_USER_F_PROTOCOL_FEATURES 30
191 Starting and stopping rings
192 ----------------------
193 Client must only process each ring when it is started.
195 Client must only pass data between the ring and the
196 backend, when the ring is enabled.
198 If ring is started but disabled, client must process the
199 ring without talking to the backend.
201 For example, for a networking device, in the disabled state
202 client must not supply any new RX packets, but must process
203 and discard any TX packets.
205 If VHOST_USER_F_PROTOCOL_FEATURES has not been negotiated, the ring is initialized
208 If VHOST_USER_F_PROTOCOL_FEATURES has been negotiated, the ring is initialized
209 in a disabled state. Client must not pass data to/from the backend until ring is enabled by
210 VHOST_USER_SET_VRING_ENABLE with parameter 1, or after it has been disabled by
211 VHOST_USER_SET_VRING_ENABLE with parameter 0.
213 Each ring is initialized in a stopped state, client must not process it until
214 ring is started, or after it has been stopped.
216 Client must start ring upon receiving a kick (that is, detecting that file
217 descriptor is readable) on the descriptor specified by
218 VHOST_USER_SET_VRING_KICK, and stop ring upon receiving
219 VHOST_USER_GET_VRING_BASE.
221 While processing the rings (whether they are enabled or not), client must
222 support changing some configuration aspects on the fly.
224 Multiple queue support
225 ----------------------
227 Multiple queue is treated as a protocol extension, hence the slave has to
228 implement protocol features first. The multiple queues feature is supported
229 only when the protocol feature VHOST_USER_PROTOCOL_F_MQ (bit 0) is set.
231 The max number of queue pairs the slave supports can be queried with message
232 VHOST_USER_GET_QUEUE_NUM. Master should stop when the number of
233 requested queues is bigger than that.
235 As all queues share one connection, the master uses a unique index for each
236 queue in the sent message to identify a specified queue. One queue pair
237 is enabled initially. More queues are enabled dynamically, by sending
238 message VHOST_USER_SET_VRING_ENABLE.
243 During live migration, the master may need to track the modifications
244 the slave makes to the memory mapped regions. The client should mark
245 the dirty pages in a log. Once it complies to this logging, it may
246 declare the VHOST_F_LOG_ALL vhost feature.
248 To start/stop logging of data/used ring writes, server may send messages
249 VHOST_USER_SET_FEATURES with VHOST_F_LOG_ALL and VHOST_USER_SET_VRING_ADDR with
250 VHOST_VRING_F_LOG in ring's flags set to 1/0, respectively.
252 All the modifications to memory pointed by vring "descriptor" should
253 be marked. Modifications to "used" vring should be marked if
254 VHOST_VRING_F_LOG is part of ring's flags.
256 Dirty pages are of size:
257 #define VHOST_LOG_PAGE 0x1000
259 The log memory fd is provided in the ancillary data of
260 VHOST_USER_SET_LOG_BASE message when the slave has
261 VHOST_USER_PROTOCOL_F_LOG_SHMFD protocol feature.
263 The size of the log is supplied as part of VhostUserMsg
264 which should be large enough to cover all known guest
265 addresses. Log starts at the supplied offset in the
266 supplied file descriptor.
267 The log covers from address 0 to the maximum of guest
268 regions. In pseudo-code, to mark page at "addr" as dirty:
270 page = addr / VHOST_LOG_PAGE
271 log[page / 8] |= 1 << page % 8
273 Where addr is the guest physical address.
275 Use atomic operations, as the log may be concurrently manipulated.
277 Note that when logging modifications to the used ring (when VHOST_VRING_F_LOG
278 is set for this ring), log_guest_addr should be used to calculate the log
279 offset: the write to first byte of the used ring is logged at this offset from
280 log start. Also note that this value might be outside the legal guest physical
281 address range (i.e. does not have to be covered by the VhostUserMemory table),
282 but the bit offset of the last byte of the ring must fall within
283 the size supplied by VhostUserLog.
285 VHOST_USER_SET_LOG_FD is an optional message with an eventfd in
286 ancillary data, it may be used to inform the master that the log has
289 Once the source has finished migration, rings will be stopped by
290 the source. No further update must be done before rings are
293 In postcopy migration the slave is started before all the memory has been
294 received from the source host, and care must be taken to avoid accessing pages
295 that have yet to be received. The slave opens a 'userfault'-fd and registers
296 the memory with it; this fd is then passed back over to the master.
297 The master services requests on the userfaultfd for pages that are accessed
298 and when the page is available it performs WAKE ioctl's on the userfaultfd
299 to wake the stalled slave. The client indicates support for this via the
300 VHOST_USER_PROTOCOL_F_PAGEFAULT feature.
305 The master sends a list of vhost memory regions to the slave using the
306 VHOST_USER_SET_MEM_TABLE message. Each region has two base addresses: a guest
307 address and a user address.
309 Messages contain guest addresses and/or user addresses to reference locations
310 within the shared memory. The mapping of these addresses works as follows.
312 User addresses map to the vhost memory region containing that user address.
314 When the VIRTIO_F_IOMMU_PLATFORM feature has not been negotiated:
316 * Guest addresses map to the vhost memory region containing that guest
319 When the VIRTIO_F_IOMMU_PLATFORM feature has been negotiated:
321 * Guest addresses are also called I/O virtual addresses (IOVAs). They are
322 translated to user addresses via the IOTLB.
324 * The vhost memory region guest address is not used.
329 When the VIRTIO_F_IOMMU_PLATFORM feature has been negotiated, the master
330 sends IOTLB entries update & invalidation by sending VHOST_USER_IOTLB_MSG
331 requests to the slave with a struct vhost_iotlb_msg as payload. For update
332 events, the iotlb payload has to be filled with the update message type (2),
333 the I/O virtual address, the size, the user virtual address, and the
334 permissions flags. Addresses and size must be within vhost memory regions set
335 via the VHOST_USER_SET_MEM_TABLE request. For invalidation events, the iotlb
336 payload has to be filled with the invalidation message type (3), the I/O virtual
337 address and the size. On success, the slave is expected to reply with a zero
338 payload, non-zero otherwise.
340 The slave relies on the slave communcation channel (see "Slave communication"
341 section below) to send IOTLB miss and access failure events, by sending
342 VHOST_USER_SLAVE_IOTLB_MSG requests to the master with a struct vhost_iotlb_msg
343 as payload. For miss events, the iotlb payload has to be filled with the miss
344 message type (1), the I/O virtual address and the permissions flags. For access
345 failure event, the iotlb payload has to be filled with the access failure
346 message type (4), the I/O virtual address and the permissions flags.
347 For synchronization purpose, the slave may rely on the reply-ack feature,
348 so the master may send a reply when operation is completed if the reply-ack
349 feature is negotiated and slaves requests a reply. For miss events, completed
350 operation means either master sent an update message containing the IOTLB entry
351 containing requested address and permission, or master sent nothing if the IOTLB
352 miss message is invalid (invalid IOVA or permission).
354 The master isn't expected to take the initiative to send IOTLB update messages,
355 as the slave sends IOTLB miss messages for the guest virtual memory areas it
361 An optional communication channel is provided if the slave declares
362 VHOST_USER_PROTOCOL_F_SLAVE_REQ protocol feature, to allow the slave to make
363 requests to the master.
365 The fd is provided via VHOST_USER_SET_SLAVE_REQ_FD ancillary data.
367 A slave may then send VHOST_USER_SLAVE_* messages to the master
368 using this fd communication channel.
373 #define VHOST_USER_PROTOCOL_F_MQ 0
374 #define VHOST_USER_PROTOCOL_F_LOG_SHMFD 1
375 #define VHOST_USER_PROTOCOL_F_RARP 2
376 #define VHOST_USER_PROTOCOL_F_REPLY_ACK 3
377 #define VHOST_USER_PROTOCOL_F_MTU 4
378 #define VHOST_USER_PROTOCOL_F_SLAVE_REQ 5
379 #define VHOST_USER_PROTOCOL_F_CROSS_ENDIAN 6
380 #define VHOST_USER_PROTOCOL_F_CRYPTO_SESSION 7
381 #define VHOST_USER_PROTOCOL_F_PAGEFAULT 8
382 #define VHOST_USER_PROTOCOL_F_CONFIG 9
387 * VHOST_USER_GET_FEATURES
390 Equivalent ioctl: VHOST_GET_FEATURES
394 Get from the underlying vhost implementation the features bitmask.
395 Feature bit VHOST_USER_F_PROTOCOL_FEATURES signals slave support for
396 VHOST_USER_GET_PROTOCOL_FEATURES and VHOST_USER_SET_PROTOCOL_FEATURES.
398 * VHOST_USER_SET_FEATURES
401 Ioctl: VHOST_SET_FEATURES
404 Enable features in the underlying vhost implementation using a bitmask.
405 Feature bit VHOST_USER_F_PROTOCOL_FEATURES signals slave support for
406 VHOST_USER_GET_PROTOCOL_FEATURES and VHOST_USER_SET_PROTOCOL_FEATURES.
408 * VHOST_USER_GET_PROTOCOL_FEATURES
411 Equivalent ioctl: VHOST_GET_FEATURES
415 Get the protocol feature bitmask from the underlying vhost implementation.
416 Only legal if feature bit VHOST_USER_F_PROTOCOL_FEATURES is present in
417 VHOST_USER_GET_FEATURES.
418 Note: slave that reported VHOST_USER_F_PROTOCOL_FEATURES must support
419 this message even before VHOST_USER_SET_FEATURES was called.
421 * VHOST_USER_SET_PROTOCOL_FEATURES
424 Ioctl: VHOST_SET_FEATURES
427 Enable protocol features in the underlying vhost implementation.
428 Only legal if feature bit VHOST_USER_F_PROTOCOL_FEATURES is present in
429 VHOST_USER_GET_FEATURES.
430 Note: slave that reported VHOST_USER_F_PROTOCOL_FEATURES must support
431 this message even before VHOST_USER_SET_FEATURES was called.
433 * VHOST_USER_SET_OWNER
436 Equivalent ioctl: VHOST_SET_OWNER
439 Issued when a new connection is established. It sets the current Master
440 as an owner of the session. This can be used on the Slave as a
441 "session start" flag.
443 * VHOST_USER_RESET_OWNER
448 This is no longer used. Used to be sent to request disabling
449 all rings, but some clients interpreted it to also discard
450 connection state (this interpretation would lead to bugs).
451 It is recommended that clients either ignore this message,
452 or use it to disable all rings.
454 * VHOST_USER_SET_MEM_TABLE
457 Equivalent ioctl: VHOST_SET_MEM_TABLE
458 Master payload: memory regions description
459 Slave payload: (postcopy only) memory regions description
461 Sets the memory map regions on the slave so it can translate the vring
462 addresses. In the ancillary data there is an array of file descriptors
463 for each memory mapped region. The size and ordering of the fds matches
464 the number and ordering of memory regions.
466 When VHOST_USER_POSTCOPY_LISTEN has been received, SET_MEM_TABLE replies with
467 the bases of the memory mapped regions to the master. The slave must
468 have mmap'd the regions but not yet accessed them and should not yet generate
469 a userfault event. Note NEED_REPLY_MASK is not set in this case.
470 QEMU will then reply back to the list of mappings with an empty
471 VHOST_USER_SET_MEM_TABLE as an acknowledgment; only upon reception of this
472 message may the guest start accessing the memory and generating faults.
474 * VHOST_USER_SET_LOG_BASE
477 Equivalent ioctl: VHOST_SET_LOG_BASE
481 Sets logging shared memory space.
482 When slave has VHOST_USER_PROTOCOL_F_LOG_SHMFD protocol
483 feature, the log memory fd is provided in the ancillary data of
484 VHOST_USER_SET_LOG_BASE message, the size and offset of shared
485 memory area provided in the message.
488 * VHOST_USER_SET_LOG_FD
491 Equivalent ioctl: VHOST_SET_LOG_FD
494 Sets the logging file descriptor, which is passed as ancillary data.
496 * VHOST_USER_SET_VRING_NUM
499 Equivalent ioctl: VHOST_SET_VRING_NUM
500 Master payload: vring state description
502 Set the size of the queue.
504 * VHOST_USER_SET_VRING_ADDR
507 Equivalent ioctl: VHOST_SET_VRING_ADDR
508 Master payload: vring address description
511 Sets the addresses of the different aspects of the vring.
513 * VHOST_USER_SET_VRING_BASE
516 Equivalent ioctl: VHOST_SET_VRING_BASE
517 Master payload: vring state description
519 Sets the base offset in the available vring.
521 * VHOST_USER_GET_VRING_BASE
524 Equivalent ioctl: VHOST_USER_GET_VRING_BASE
525 Master payload: vring state description
526 Slave payload: vring state description
528 Get the available vring base offset.
530 * VHOST_USER_SET_VRING_KICK
533 Equivalent ioctl: VHOST_SET_VRING_KICK
536 Set the event file descriptor for adding buffers to the vring. It
537 is passed in the ancillary data.
538 Bits (0-7) of the payload contain the vring index. Bit 8 is the
539 invalid FD flag. This flag is set when there is no file descriptor
540 in the ancillary data. This signals that polling should be used
541 instead of waiting for a kick.
543 * VHOST_USER_SET_VRING_CALL
546 Equivalent ioctl: VHOST_SET_VRING_CALL
549 Set the event file descriptor to signal when buffers are used. It
550 is passed in the ancillary data.
551 Bits (0-7) of the payload contain the vring index. Bit 8 is the
552 invalid FD flag. This flag is set when there is no file descriptor
553 in the ancillary data. This signals that polling will be used
554 instead of waiting for the call.
556 * VHOST_USER_SET_VRING_ERR
559 Equivalent ioctl: VHOST_SET_VRING_ERR
562 Set the event file descriptor to signal when error occurs. It
563 is passed in the ancillary data.
564 Bits (0-7) of the payload contain the vring index. Bit 8 is the
565 invalid FD flag. This flag is set when there is no file descriptor
566 in the ancillary data.
568 * VHOST_USER_GET_QUEUE_NUM
571 Equivalent ioctl: N/A
575 Query how many queues the backend supports. This request should be
576 sent only when VHOST_USER_PROTOCOL_F_MQ is set in queried protocol
577 features by VHOST_USER_GET_PROTOCOL_FEATURES.
579 * VHOST_USER_SET_VRING_ENABLE
582 Equivalent ioctl: N/A
583 Master payload: vring state description
585 Signal slave to enable or disable corresponding vring.
586 This request should be sent only when VHOST_USER_F_PROTOCOL_FEATURES
589 * VHOST_USER_SEND_RARP
592 Equivalent ioctl: N/A
595 Ask vhost user backend to broadcast a fake RARP to notify the migration
596 is terminated for guest that does not support GUEST_ANNOUNCE.
597 Only legal if feature bit VHOST_USER_F_PROTOCOL_FEATURES is present in
598 VHOST_USER_GET_FEATURES and protocol feature bit VHOST_USER_PROTOCOL_F_RARP
599 is present in VHOST_USER_GET_PROTOCOL_FEATURES.
600 The first 6 bytes of the payload contain the mac address of the guest to
601 allow the vhost user backend to construct and broadcast the fake RARP.
603 * VHOST_USER_NET_SET_MTU
606 Equivalent ioctl: N/A
609 Set host MTU value exposed to the guest.
610 This request should be sent only when VIRTIO_NET_F_MTU feature has been
611 successfully negotiated, VHOST_USER_F_PROTOCOL_FEATURES is present in
612 VHOST_USER_GET_FEATURES and protocol feature bit
613 VHOST_USER_PROTOCOL_F_NET_MTU is present in
614 VHOST_USER_GET_PROTOCOL_FEATURES.
615 If VHOST_USER_PROTOCOL_F_REPLY_ACK is negotiated, slave must respond
616 with zero in case the specified MTU is valid, or non-zero otherwise.
618 * VHOST_USER_SET_SLAVE_REQ_FD
621 Equivalent ioctl: N/A
624 Set the socket file descriptor for slave initiated requests. It is passed
625 in the ancillary data.
626 This request should be sent only when VHOST_USER_F_PROTOCOL_FEATURES
627 has been negotiated, and protocol feature bit VHOST_USER_PROTOCOL_F_SLAVE_REQ
628 bit is present in VHOST_USER_GET_PROTOCOL_FEATURES.
629 If VHOST_USER_PROTOCOL_F_REPLY_ACK is negotiated, slave must respond
630 with zero for success, non-zero otherwise.
632 * VHOST_USER_IOTLB_MSG
635 Equivalent ioctl: N/A (equivalent to VHOST_IOTLB_MSG message type)
636 Master payload: struct vhost_iotlb_msg
639 Send IOTLB messages with struct vhost_iotlb_msg as payload.
640 Master sends such requests to update and invalidate entries in the device
641 IOTLB. The slave has to acknowledge the request with sending zero as u64
642 payload for success, non-zero otherwise.
643 This request should be send only when VIRTIO_F_IOMMU_PLATFORM feature
644 has been successfully negotiated.
646 * VHOST_USER_SET_VRING_ENDIAN
649 Equivalent ioctl: VHOST_SET_VRING_ENDIAN
650 Master payload: vring state description
652 Set the endianess of a VQ for legacy devices. Little-endian is indicated
653 with state.num set to 0 and big-endian is indicated with state.num set
654 to 1. Other values are invalid.
655 This request should be sent only when VHOST_USER_PROTOCOL_F_CROSS_ENDIAN
657 Backends that negotiated this feature should handle both endianesses
658 and expect this message once (per VQ) during device configuration
659 (ie. before the master starts the VQ).
661 * VHOST_USER_GET_CONFIG
664 Equivalent ioctl: N/A
665 Master payload: virtio device config space
666 Slave payload: virtio device config space
668 When VHOST_USER_PROTOCOL_F_CONFIG is negotiated, this message is
669 submitted by the vhost-user master to fetch the contents of the virtio
670 device configuration space, vhost-user slave's payload size MUST match
671 master's request, vhost-user slave uses zero length of payload to
672 indicate an error to vhost-user master. The vhost-user master may
673 cache the contents to avoid repeated VHOST_USER_GET_CONFIG calls.
675 * VHOST_USER_SET_CONFIG
678 Equivalent ioctl: N/A
679 Master payload: virtio device config space
682 When VHOST_USER_PROTOCOL_F_CONFIG is negotiated, this message is
683 submitted by the vhost-user master when the Guest changes the virtio
684 device configuration space and also can be used for live migration
685 on the destination host. The vhost-user slave must check the flags
686 field, and slaves MUST NOT accept SET_CONFIG for read-only
687 configuration space fields unless the live migration bit is set.
689 * VHOST_USER_CREATE_CRYPTO_SESSION
692 Equivalent ioctl: N/A
693 Master payload: crypto session description
694 Slave payload: crypto session description
696 Create a session for crypto operation. The server side must return the
697 session id, 0 or positive for success, negative for failure.
698 This request should be sent only when VHOST_USER_PROTOCOL_F_CRYPTO_SESSION
699 feature has been successfully negotiated.
700 It's a required feature for crypto devices.
702 * VHOST_USER_CLOSE_CRYPTO_SESSION
705 Equivalent ioctl: N/A
708 Close a session for crypto operation which was previously
709 created by VHOST_USER_CREATE_CRYPTO_SESSION.
710 This request should be sent only when VHOST_USER_PROTOCOL_F_CRYPTO_SESSION
711 feature has been successfully negotiated.
712 It's a required feature for crypto devices.
714 * VHOST_USER_POSTCOPY_ADVISE
717 Slave payload: userfault fd
719 When VHOST_USER_PROTOCOL_F_PAGEFAULT is supported, the
720 master advises slave that a migration with postcopy enabled is underway,
721 the slave must open a userfaultfd for later use.
722 Note that at this stage the migration is still in precopy mode.
724 * VHOST_USER_POSTCOPY_LISTEN
728 Master advises slave that a transition to postcopy mode has happened.
729 The slave must ensure that shared memory is registered with userfaultfd
730 to cause faulting of non-present pages.
732 This is always sent sometime after a VHOST_USER_POSTCOPY_ADVISE, and
733 thus only when VHOST_USER_PROTOCOL_F_PAGEFAULT is supported.
735 * VHOST_USER_POSTCOPY_END
739 Master advises that postcopy migration has now completed. The
740 slave must disable the userfaultfd. The response is an acknowledgement
742 When VHOST_USER_PROTOCOL_F_PAGEFAULT is supported, this message
743 is sent at the end of the migration, after VHOST_USER_POSTCOPY_LISTEN
745 The value returned is an error indication; 0 is success.
750 * VHOST_USER_SLAVE_IOTLB_MSG
753 Equivalent ioctl: N/A (equivalent to VHOST_IOTLB_MSG message type)
754 Slave payload: struct vhost_iotlb_msg
757 Send IOTLB messages with struct vhost_iotlb_msg as payload.
758 Slave sends such requests to notify of an IOTLB miss, or an IOTLB
759 access failure. If VHOST_USER_PROTOCOL_F_REPLY_ACK is negotiated,
760 and slave set the VHOST_USER_NEED_REPLY flag, master must respond with
761 zero when operation is successfully completed, or non-zero otherwise.
762 This request should be send only when VIRTIO_F_IOMMU_PLATFORM feature
763 has been successfully negotiated.
765 * VHOST_USER_SLAVE_CONFIG_CHANGE_MSG
768 Equivalent ioctl: N/A
772 When VHOST_USER_PROTOCOL_F_CONFIG is negotiated, vhost-user slave sends
773 such messages to notify that the virtio device's configuration space has
774 changed, for those host devices which can support such feature, host
775 driver can send VHOST_USER_GET_CONFIG message to slave to get the latest
776 content. If VHOST_USER_PROTOCOL_F_REPLY_ACK is negotiated, and slave set
777 the VHOST_USER_NEED_REPLY flag, master must respond with zero when
778 operation is successfully completed, or non-zero otherwise.
780 VHOST_USER_PROTOCOL_F_REPLY_ACK:
781 -------------------------------
782 The original vhost-user specification only demands replies for certain
783 commands. This differs from the vhost protocol implementation where commands
784 are sent over an ioctl() call and block until the client has completed.
786 With this protocol extension negotiated, the sender (QEMU) can set the
787 "need_reply" [Bit 3] flag to any command. This indicates that
788 the client MUST respond with a Payload VhostUserMsg indicating success or
789 failure. The payload should be set to zero on success or non-zero on failure,
790 unless the message already has an explicit reply body.
792 The response payload gives QEMU a deterministic indication of the result
793 of the command. Today, QEMU is expected to terminate the main vhost-user
794 loop upon receiving such errors. In future, qemu could be taught to be more
795 resilient for selective requests.
797 For the message types that already solicit a reply from the client, the
798 presence of VHOST_USER_PROTOCOL_F_REPLY_ACK or need_reply bit being set brings
799 no behavioural change. (See the 'Communication' section for details.)