| // Copyright (C) 2019 Alibaba Cloud. All rights reserved. |
| // SPDX-License-Identifier: Apache-2.0 or BSD-3-Clause |
| |
| //! Virtio Vhost Backend Drivers |
| //! |
| //! Virtio devices use virtqueues to transport data efficiently. The first generation of virtqueue |
| //! is a set of three different single-producer, single-consumer ring structures designed to store |
| //! generic scatter-gather I/O. The virtio specification 1.1 introduces an alternative compact |
| //! virtqueue layout named "Packed Virtqueue", which is more friendly to memory cache system and |
| //! hardware implemented virtio devices. The packed virtqueue uses read-write memory, that means |
| //! the memory will be both read and written by both host and guest. The new Packed Virtqueue is |
| //! preferred for performance. |
| //! |
| //! Vhost is a mechanism to improve performance of Virtio devices by delegate data plane operations |
| //! to dedicated IO service processes. Only the configuration, I/O submission notification, and I/O |
| //! completion interruption are piped through the hypervisor. |
| //! It uses the same virtqueue layout as Virtio to allow Vhost devices to be mapped directly to |
| //! Virtio devices. This allows a Vhost device to be accessed directly by a guest OS inside a |
| //! hypervisor process with an existing Virtio (PCI) driver. |
| //! |
| //! The initial vhost implementation is a part of the Linux kernel and uses ioctl interface to |
| //! communicate with userspace applications. Dedicated kernel worker threads are created to handle |
| //! IO requests from the guest. |
| //! |
| //! Later Vhost-user protocol is introduced to complement the ioctl interface used to control the |
| //! vhost implementation in the Linux kernel. It implements the control plane needed to establish |
| //! virtqueues sharing with a user space process on the same host. It uses communication over a |
| //! Unix domain socket to share file descriptors in the ancillary data of the message. |
| //! The protocol defines 2 sides of the communication, master and slave. Master is the application |
| //! that shares its virtqueues. Slave is the consumer of the virtqueues. Master and slave can be |
| //! either a client (i.e. connecting) or server (listening) in the socket communication. |
| |
| #![deny(missing_docs)] |
| |
| #[cfg(any(feature = "vmm", feature = "device"))] |
| use std::fs::File; |
| use std::io::Error as IOError; |
| |
| use remain::sorted; |
| use thiserror::Error as ThisError; |
| |
| mod backend; |
| pub use backend::*; |
| |
| pub mod message; |
| |
| pub mod connection; |
| |
| #[cfg(feature = "vmm")] |
| mod master; |
| #[cfg(feature = "vmm")] |
| pub use self::master::{Master, VhostUserMaster}; |
| #[cfg(feature = "vmm")] |
| mod master_req_handler; |
| #[cfg(feature = "vmm")] |
| pub use self::master_req_handler::{ |
| MasterReqHandler, VhostUserMasterReqHandler, VhostUserMasterReqHandlerMut, |
| }; |
| |
| #[cfg(feature = "device")] |
| mod slave; |
| #[cfg(feature = "device")] |
| pub use self::slave::SlaveListener; |
| #[cfg(feature = "device")] |
| mod slave_req_handler; |
| #[cfg(feature = "device")] |
| pub use self::slave_req_handler::{ |
| Protocol, SlaveReqHandler, SlaveReqHelper, VhostUserSlaveReqHandler, |
| VhostUserSlaveReqHandlerMut, |
| }; |
| #[cfg(feature = "device")] |
| mod slave_fs_cache; |
| #[cfg(feature = "device")] |
| pub use self::slave_fs_cache::SlaveFsCacheReq; |
| |
| /// Errors for vhost-user operations |
| #[sorted] |
| #[derive(Debug, ThisError)] |
| pub enum Error { |
| /// Virtio/protocol features mismatch. |
| #[error("virtio features mismatch")] |
| FeatureMismatch, |
| /// Fd array in question is too big or too small |
| #[error("wrong number of attached fds")] |
| IncorrectFds, |
| /// Invalid message format, flag or content. |
| #[error("invalid message")] |
| InvalidMessage, |
| /// Unsupported operations due to that the protocol feature hasn't been negotiated. |
| #[error("invalid operation")] |
| InvalidOperation, |
| /// Invalid parameters. |
| #[error("invalid parameters")] |
| InvalidParam, |
| /// Failure from the master side. |
| #[error("master Internal error")] |
| MasterInternalError, |
| /// Message is too large |
| #[error("oversized message")] |
| OversizedMsg, |
| /// Only part of a message have been sent or received successfully |
| #[error("partial message")] |
| PartialMessage, |
| /// Error from request handler |
| #[error("handler failed to handle request: {0}")] |
| ReqHandlerError(IOError), |
| /// Failure from the slave side. |
| #[error("slave internal error")] |
| SlaveInternalError, |
| /// The socket is broken or has been closed. |
| #[error("socket is broken: {0}")] |
| SocketBroken(std::io::Error), |
| /// Can't connect to peer. |
| #[error("can't connect to peer: {0}")] |
| SocketConnect(std::io::Error), |
| /// Generic socket errors. |
| #[error("socket error: {0}")] |
| SocketError(std::io::Error), |
| /// Should retry the socket operation again. |
| #[error("temporary socket error: {0}")] |
| SocketRetry(std::io::Error), |
| /// Error from VFIO device. |
| #[error("error occurred in VFIO device: {0}")] |
| VfioDeviceError(anyhow::Error), |
| } |
| |
| impl Error { |
| /// Determine whether to rebuild the underline communication channel. |
| pub fn should_reconnect(&self) -> bool { |
| match *self { |
| // Should reconnect because it may be caused by temporary network errors. |
| Error::PartialMessage => true, |
| // Should reconnect because the underline socket is broken. |
| Error::SocketBroken(_) => true, |
| // Slave internal error, hope it recovers on reconnect. |
| Error::SlaveInternalError => true, |
| // Master internal error, hope it recovers on reconnect. |
| Error::MasterInternalError => true, |
| // Should just retry the IO operation instead of rebuilding the underline connection. |
| Error::SocketRetry(_) => false, |
| Error::InvalidParam | Error::InvalidOperation => false, |
| Error::InvalidMessage | Error::IncorrectFds | Error::OversizedMsg => false, |
| Error::SocketError(_) | Error::SocketConnect(_) => false, |
| Error::FeatureMismatch => false, |
| Error::ReqHandlerError(_) => false, |
| Error::VfioDeviceError(_) => false, |
| } |
| } |
| } |
| |
| impl std::convert::From<sys_util::Error> for Error { |
| /// Convert raw socket errors into meaningful vhost-user errors. |
| /// |
| /// The sys_util::Error is a simple wrapper over the raw errno, which doesn't means |
| /// much to the vhost-user connection manager. So convert it into meaningful errors to simplify |
| /// the connection manager logic. |
| /// |
| /// # Return: |
| /// * - Error::SocketRetry: temporary error caused by signals or short of resources. |
| /// * - Error::SocketBroken: the underline socket is broken. |
| /// * - Error::SocketError: other socket related errors. |
| #[allow(unreachable_patterns)] // EWOULDBLOCK equals to EGAIN on linux |
| fn from(err: sys_util::Error) -> Self { |
| match err.errno() { |
| // Retry: |
| // * EAGAIN, EWOULDBLOCK: The socket is marked nonblocking and the requested operation |
| // would block. |
| // * EINTR: A signal occurred before any data was transmitted |
| // * ENOBUFS: The output queue for a network interface was full. This generally |
| // indicates that the interface has stopped sending, but may be caused by transient |
| // congestion. |
| // * ENOMEM: No memory available. |
| libc::EAGAIN | libc::EWOULDBLOCK | libc::EINTR | libc::ENOBUFS | libc::ENOMEM => { |
| Error::SocketRetry(err.into()) |
| } |
| // Broken: |
| // * ECONNRESET: Connection reset by peer. |
| // * EPIPE: The local end has been shut down on a connection oriented socket. In this |
| // case the process will also receive a SIGPIPE unless MSG_NOSIGNAL is set. |
| libc::ECONNRESET | libc::EPIPE => Error::SocketBroken(err.into()), |
| // Write permission is denied on the destination socket file, or search permission is |
| // denied for one of the directories the path prefix. |
| libc::EACCES => Error::SocketConnect(IOError::from_raw_os_error(libc::EACCES)), |
| // Catch all other errors |
| e => Error::SocketError(IOError::from_raw_os_error(e)), |
| } |
| } |
| } |
| |
| /// Result of vhost-user operations |
| pub type Result<T> = std::result::Result<T, Error>; |
| |
| /// Result of request handler. |
| pub type HandlerResult<T> = std::result::Result<T, IOError>; |
| |
| /// Utility function to take the first element from option of a vector of files. |
| /// Returns `None` if the vector contains no file or more than one file. |
| #[cfg(any(feature = "vmm", feature = "device"))] |
| pub(crate) fn take_single_file(files: Option<Vec<File>>) -> Option<File> { |
| let mut files = files?; |
| if files.len() != 1 { |
| return None; |
| } |
| Some(files.swap_remove(0)) |
| } |
| |
| #[cfg(all(test, feature = "device"))] |
| mod dummy_slave; |
| |
| #[cfg(all(test, feature = "vmm", feature = "device"))] |
| mod tests { |
| use std::os::unix::io::AsRawFd; |
| use std::path::Path; |
| use std::sync::{Arc, Barrier, Mutex}; |
| use std::thread; |
| |
| use super::connection::socket::{Endpoint, Listener}; |
| use super::dummy_slave::{DummySlaveReqHandler, VIRTIO_FEATURES}; |
| use super::message::*; |
| use super::*; |
| use crate::backend::VhostBackend; |
| use crate::{VhostUserMemoryRegionInfo, VringConfigData}; |
| use base::AsRawDescriptor; |
| use tempfile::{tempfile, Builder, TempDir}; |
| |
| fn temp_dir() -> TempDir { |
| Builder::new().prefix("/tmp/vhost_test").tempdir().unwrap() |
| } |
| |
| fn create_slave<P, S>( |
| path: P, |
| backend: Arc<S>, |
| ) -> ( |
| Master<Endpoint<MasterReq>>, |
| SlaveReqHandler<S, Endpoint<MasterReq>>, |
| ) |
| where |
| P: AsRef<Path>, |
| S: VhostUserSlaveReqHandler, |
| { |
| let listener = Listener::new(&path, true).unwrap(); |
| let mut slave_listener = SlaveListener::new(listener, backend).unwrap(); |
| let master = Master::connect(&path, 1).unwrap(); |
| (master, slave_listener.accept().unwrap().unwrap()) |
| } |
| |
| #[test] |
| fn create_dummy_slave() { |
| let slave = Arc::new(Mutex::new(DummySlaveReqHandler::new())); |
| |
| slave.set_owner().unwrap(); |
| assert!(slave.set_owner().is_err()); |
| } |
| |
| #[test] |
| fn test_set_owner() { |
| let slave_be = Arc::new(Mutex::new(DummySlaveReqHandler::new())); |
| let dir = temp_dir(); |
| let mut path = dir.path().to_owned(); |
| path.push("sock"); |
| let (master, mut slave) = create_slave(&path, slave_be.clone()); |
| |
| assert!(!slave_be.lock().unwrap().owned); |
| master.set_owner().unwrap(); |
| slave.handle_request().unwrap(); |
| assert!(slave_be.lock().unwrap().owned); |
| master.set_owner().unwrap(); |
| assert!(slave.handle_request().is_err()); |
| assert!(slave_be.lock().unwrap().owned); |
| } |
| |
| #[test] |
| fn test_set_features() { |
| let mbar = Arc::new(Barrier::new(2)); |
| let sbar = mbar.clone(); |
| let dir = temp_dir(); |
| let mut path = dir.path().to_owned(); |
| path.push("sock"); |
| let slave_be = Arc::new(Mutex::new(DummySlaveReqHandler::new())); |
| let (mut master, mut slave) = create_slave(&path, slave_be.clone()); |
| |
| thread::spawn(move || { |
| slave.handle_request().unwrap(); |
| assert!(slave_be.lock().unwrap().owned); |
| |
| slave.handle_request().unwrap(); |
| slave.handle_request().unwrap(); |
| assert_eq!( |
| slave_be.lock().unwrap().acked_features, |
| VIRTIO_FEATURES & !0x1 |
| ); |
| |
| slave.handle_request().unwrap(); |
| slave.handle_request().unwrap(); |
| assert_eq!( |
| slave_be.lock().unwrap().acked_protocol_features, |
| VhostUserProtocolFeatures::all().bits() |
| ); |
| |
| sbar.wait(); |
| }); |
| |
| master.set_owner().unwrap(); |
| |
| // set virtio features |
| let features = master.get_features().unwrap(); |
| assert_eq!(features, VIRTIO_FEATURES); |
| master.set_features(VIRTIO_FEATURES & !0x1).unwrap(); |
| |
| // set vhost protocol features |
| let features = master.get_protocol_features().unwrap(); |
| assert_eq!(features.bits(), VhostUserProtocolFeatures::all().bits()); |
| master.set_protocol_features(features).unwrap(); |
| |
| mbar.wait(); |
| } |
| |
| #[test] |
| fn test_master_slave_process() { |
| let mbar = Arc::new(Barrier::new(2)); |
| let sbar = mbar.clone(); |
| let dir = temp_dir(); |
| let mut path = dir.path().to_owned(); |
| path.push("sock"); |
| let slave_be = Arc::new(Mutex::new(DummySlaveReqHandler::new())); |
| let (mut master, mut slave) = create_slave(&path, slave_be.clone()); |
| |
| thread::spawn(move || { |
| // set_own() |
| slave.handle_request().unwrap(); |
| assert!(slave_be.lock().unwrap().owned); |
| |
| // get/set_features() |
| slave.handle_request().unwrap(); |
| slave.handle_request().unwrap(); |
| assert_eq!( |
| slave_be.lock().unwrap().acked_features, |
| VIRTIO_FEATURES & !0x1 |
| ); |
| |
| slave.handle_request().unwrap(); |
| slave.handle_request().unwrap(); |
| assert_eq!( |
| slave_be.lock().unwrap().acked_protocol_features, |
| VhostUserProtocolFeatures::all().bits() |
| ); |
| |
| // get_inflight_fd() |
| slave.handle_request().unwrap(); |
| // set_inflight_fd() |
| slave.handle_request().unwrap(); |
| |
| // get_queue_num() |
| slave.handle_request().unwrap(); |
| |
| // set_mem_table() |
| slave.handle_request().unwrap(); |
| |
| // get/set_config() |
| slave.handle_request().unwrap(); |
| slave.handle_request().unwrap(); |
| |
| // set_slave_request_fd |
| slave.handle_request().unwrap(); |
| |
| // set_vring_enable |
| slave.handle_request().unwrap(); |
| |
| // set_log_base,set_log_fd() |
| slave.handle_request().unwrap_err(); |
| slave.handle_request().unwrap_err(); |
| |
| // set_vring_xxx |
| slave.handle_request().unwrap(); |
| slave.handle_request().unwrap(); |
| slave.handle_request().unwrap(); |
| slave.handle_request().unwrap(); |
| slave.handle_request().unwrap(); |
| slave.handle_request().unwrap(); |
| |
| // get_max_mem_slots() |
| slave.handle_request().unwrap(); |
| |
| // add_mem_region() |
| slave.handle_request().unwrap(); |
| |
| // remove_mem_region() |
| slave.handle_request().unwrap(); |
| |
| sbar.wait(); |
| }); |
| |
| master.set_owner().unwrap(); |
| |
| // set virtio features |
| let features = master.get_features().unwrap(); |
| assert_eq!(features, VIRTIO_FEATURES); |
| master.set_features(VIRTIO_FEATURES & !0x1).unwrap(); |
| |
| // set vhost protocol features |
| let features = master.get_protocol_features().unwrap(); |
| assert_eq!(features.bits(), VhostUserProtocolFeatures::all().bits()); |
| master.set_protocol_features(features).unwrap(); |
| |
| // Retrieve inflight I/O tracking information |
| let (inflight_info, inflight_file) = master |
| .get_inflight_fd(&VhostUserInflight { |
| num_queues: 2, |
| queue_size: 256, |
| ..Default::default() |
| }) |
| .unwrap(); |
| // Set the buffer back to the backend |
| master |
| .set_inflight_fd(&inflight_info, inflight_file.as_raw_fd()) |
| .unwrap(); |
| |
| let num = master.get_queue_num().unwrap(); |
| assert_eq!(num, 2); |
| |
| let eventfd = base::Event::new().unwrap(); |
| let mem = [VhostUserMemoryRegionInfo { |
| guest_phys_addr: 0, |
| memory_size: 0x10_0000, |
| userspace_addr: 0, |
| mmap_offset: 0, |
| mmap_handle: eventfd.as_raw_descriptor(), |
| }]; |
| master.set_mem_table(&mem).unwrap(); |
| |
| master |
| .set_config(0x100, VhostUserConfigFlags::WRITABLE, &[0xa5u8]) |
| .unwrap(); |
| let buf = [0x0u8; 4]; |
| let (reply_body, reply_payload) = master |
| .get_config(0x100, 4, VhostUserConfigFlags::empty(), &buf) |
| .unwrap(); |
| let offset = reply_body.offset; |
| assert_eq!(offset, 0x100); |
| assert_eq!(reply_payload[0], 0xa5); |
| |
| master.set_slave_request_fd(&eventfd).unwrap(); |
| master.set_vring_enable(0, true).unwrap(); |
| |
| // unimplemented yet |
| master |
| .set_log_base(0, Some(eventfd.as_raw_descriptor())) |
| .unwrap(); |
| master.set_log_fd(eventfd.as_raw_descriptor()).unwrap(); |
| |
| master.set_vring_num(0, 256).unwrap(); |
| master.set_vring_base(0, 0).unwrap(); |
| let config = VringConfigData { |
| queue_max_size: 256, |
| queue_size: 128, |
| flags: VhostUserVringAddrFlags::VHOST_VRING_F_LOG.bits(), |
| desc_table_addr: 0x1000, |
| used_ring_addr: 0x2000, |
| avail_ring_addr: 0x3000, |
| log_addr: Some(0x4000), |
| }; |
| master.set_vring_addr(0, &config).unwrap(); |
| master.set_vring_call(0, &eventfd).unwrap(); |
| master.set_vring_kick(0, &eventfd).unwrap(); |
| master.set_vring_err(0, &eventfd).unwrap(); |
| |
| let max_mem_slots = master.get_max_mem_slots().unwrap(); |
| assert_eq!(max_mem_slots, 32); |
| |
| let region_file = tempfile().unwrap(); |
| let region = VhostUserMemoryRegionInfo { |
| guest_phys_addr: 0x10_0000, |
| memory_size: 0x10_0000, |
| userspace_addr: 0, |
| mmap_offset: 0, |
| mmap_handle: region_file.as_raw_fd(), |
| }; |
| master.add_mem_region(®ion).unwrap(); |
| |
| master.remove_mem_region(®ion).unwrap(); |
| |
| mbar.wait(); |
| } |
| |
| #[test] |
| fn test_error_display() { |
| assert_eq!(format!("{}", Error::InvalidParam), "invalid parameters"); |
| assert_eq!(format!("{}", Error::InvalidOperation), "invalid operation"); |
| } |
| |
| #[test] |
| fn test_error_from_sys_util_error() { |
| let e: Error = sys_util::Error::new(libc::EAGAIN).into(); |
| if let Error::SocketRetry(e1) = e { |
| assert_eq!(e1.raw_os_error().unwrap(), libc::EAGAIN); |
| } else { |
| panic!("invalid error code conversion!"); |
| } |
| } |
| } |
