cloud-hypervisor/virtio-devices/src/vhost_user/fs.rs

// Copyright 2019 Intel Corporation. All Rights Reserved.
// SPDX-License-Identifier: Apache-2.0

use super::vu_common_ctrl::VhostUserHandle;
use super::{Error, Result, DEFAULT_VIRTIO_FEATURES};
use crate::seccomp_filters::Thread;
use crate::thread_helper::spawn_virtio_thread;
use crate::vhost_user::VhostUserCommon;
use crate::{
    ActivateError, ActivateResult, UserspaceMapping, VirtioCommon, VirtioDevice, VirtioDeviceType,
    VirtioInterrupt, VirtioSharedMemoryList, VIRTIO_F_IOMMU_PLATFORM,
};
use crate::{GuestMemoryMmap, GuestRegionMmap, MmapRegion};
use libc::{self, c_void, off64_t, pread64, pwrite64};
use seccompiler::SeccompAction;
use std::io;
use std::os::unix::io::AsRawFd;
use std::result;
use std::sync::{Arc, Barrier, Mutex};
use std::thread;
use versionize::{VersionMap, Versionize, VersionizeResult};
use versionize_derive::Versionize;
use vhost::vhost_user::message::{
    VhostUserFSSlaveMsg, VhostUserFSSlaveMsgFlags, VhostUserProtocolFeatures,
    VhostUserVirtioFeatures, VHOST_USER_FS_SLAVE_ENTRIES,
};
use vhost::vhost_user::{
    HandlerResult, MasterReqHandler, VhostUserMaster, VhostUserMasterReqHandler,
};
use virtio_queue::Queue;
use vm_memory::{
    Address, ByteValued, GuestAddress, GuestAddressSpace, GuestMemory, GuestMemoryAtomic,
};
use vm_migration::{
    protocol::MemoryRangeTable, Migratable, MigratableError, Pausable, Snapshot, Snapshottable,
    Transportable, VersionMapped,
};
use vmm_sys_util::eventfd::EventFd;

const NUM_QUEUE_OFFSET: usize = 1;
const DEFAULT_QUEUE_NUMBER: usize = 2;

#[derive(Versionize)]
pub struct State {
    pub avail_features: u64,
    pub acked_features: u64,
    pub config: VirtioFsConfig,
    pub acked_protocol_features: u64,
    pub vu_num_queues: usize,
    pub slave_req_support: bool,
}

impl VersionMapped for State {}

struct SlaveReqHandler {
    cache_offset: GuestAddress,
    cache_size: u64,
    mmap_cache_addr: u64,
    mem: GuestMemoryAtomic<GuestMemoryMmap>,
}

impl SlaveReqHandler {
    // Make sure request is within cache range
    fn is_req_valid(&self, offset: u64, len: u64) -> bool {
        let end = match offset.checked_add(len) {
            Some(n) => n,
            None => return false,
        };

        !(offset >= self.cache_size || end > self.cache_size)
    }
}

impl VhostUserMasterReqHandler for SlaveReqHandler {
    fn handle_config_change(&self) -> HandlerResult<u64> {
        debug!("handle_config_change");
        Ok(0)
    }

    fn fs_slave_map(&self, fs: &VhostUserFSSlaveMsg, fd: &dyn AsRawFd) -> HandlerResult<u64> {
        debug!("fs_slave_map");

        for i in 0..VHOST_USER_FS_SLAVE_ENTRIES {
            let offset = fs.cache_offset[i];
            let len = fs.len[i];

            // Ignore if the length is 0.
            if len == 0 {
                continue;
            }

            if !self.is_req_valid(offset, len) {
                return Err(io::Error::from_raw_os_error(libc::EINVAL));
            }

            let addr = self.mmap_cache_addr + offset;
            let flags = fs.flags[i];
            let ret = unsafe {
                libc::mmap(
                    addr as *mut libc::c_void,
                    len as usize,
                    flags.bits() as i32,
                    libc::MAP_SHARED | libc::MAP_FIXED,
                    fd.as_raw_fd(),
                    fs.fd_offset[i] as libc::off_t,
                )
            };
            if ret == libc::MAP_FAILED {
                return Err(io::Error::last_os_error());
            }
        }

        Ok(0)
    }

    fn fs_slave_unmap(&self, fs: &VhostUserFSSlaveMsg) -> HandlerResult<u64> {
        debug!("fs_slave_unmap");

        for i in 0..VHOST_USER_FS_SLAVE_ENTRIES {
            let offset = fs.cache_offset[i];
            let mut len = fs.len[i];

            // Ignore if the length is 0.
            if len == 0 {
                continue;
            }

            // Need to handle a special case where the slave ask for the unmapping
            // of the entire mapping.
            if len == 0xffff_ffff_ffff_ffff {
                len = self.cache_size;
            }

            if !self.is_req_valid(offset, len) {
                return Err(io::Error::from_raw_os_error(libc::EINVAL));
            }

            let addr = self.mmap_cache_addr + offset;
            let ret = unsafe {
                libc::mmap(
                    addr as *mut libc::c_void,
                    len as usize,
                    libc::PROT_NONE,
                    libc::MAP_ANONYMOUS | libc::MAP_PRIVATE | libc::MAP_FIXED,
                    -1,
                    0,
                )
            };
            if ret == libc::MAP_FAILED {
                return Err(io::Error::last_os_error());
            }
        }

        Ok(0)
    }

    fn fs_slave_sync(&self, fs: &VhostUserFSSlaveMsg) -> HandlerResult<u64> {
        debug!("fs_slave_sync");

        for i in 0..VHOST_USER_FS_SLAVE_ENTRIES {
            let offset = fs.cache_offset[i];
            let len = fs.len[i];

            // Ignore if the length is 0.
            if len == 0 {
                continue;
            }

            if !self.is_req_valid(offset, len) {
                return Err(io::Error::from_raw_os_error(libc::EINVAL));
            }

            let addr = self.mmap_cache_addr + offset;
            let ret =
                unsafe { libc::msync(addr as *mut libc::c_void, len as usize, libc::MS_SYNC) };
            if ret == -1 {
                return Err(io::Error::last_os_error());
            }
        }

        Ok(0)
    }

    fn fs_slave_io(&self, fs: &VhostUserFSSlaveMsg, fd: &dyn AsRawFd) -> HandlerResult<u64> {
        debug!("fs_slave_io");

        let mut done: u64 = 0;
        for i in 0..VHOST_USER_FS_SLAVE_ENTRIES {
            // Ignore if the length is 0.
            if fs.len[i] == 0 {
                continue;
            }

            let mut foffset = fs.fd_offset[i];
            let mut len = fs.len[i] as usize;
            let gpa = fs.cache_offset[i];
            let cache_end = self.cache_offset.raw_value() + self.cache_size;
            let efault = libc::EFAULT;

            let mut ptr = if gpa >= self.cache_offset.raw_value() && gpa < cache_end {
                let offset = gpa
                    .checked_sub(self.cache_offset.raw_value())
                    .ok_or_else(|| io::Error::from_raw_os_error(efault))?;
                let end = gpa
                    .checked_add(fs.len[i])
                    .ok_or_else(|| io::Error::from_raw_os_error(efault))?;

                if end >= cache_end {
                    return Err(io::Error::from_raw_os_error(efault));
                }

                self.mmap_cache_addr + offset
            } else {
                self.mem
                    .memory()
                    .get_host_address(GuestAddress(gpa))
                    .map_err(|e| {
                        error!(
                            "Failed to find RAM region associated with guest physical address 0x{:x}: {:?}",
                            gpa, e
                        );
                        io::Error::from_raw_os_error(efault)
                    })? as u64
            };

            while len > 0 {
                let ret = if (fs.flags[i] & VhostUserFSSlaveMsgFlags::MAP_W)
                    == VhostUserFSSlaveMsgFlags::MAP_W
                {
                    debug!("write: foffset={}, len={}", foffset, len);
                    unsafe {
                        pwrite64(
                            fd.as_raw_fd(),
                            ptr as *const c_void,
                            len as usize,
                            foffset as off64_t,
                        )
                    }
                } else {
                    debug!("read: foffset={}, len={}", foffset, len);
                    unsafe {
                        pread64(
                            fd.as_raw_fd(),
                            ptr as *mut c_void,
                            len as usize,
                            foffset as off64_t,
                        )
                    }
                };

                if ret < 0 {
                    return Err(io::Error::last_os_error());
                }

                if ret == 0 {
                    // EOF
                    return Err(io::Error::new(
                        io::ErrorKind::UnexpectedEof,
                        "failed to access whole buffer",
                    ));
                }
                len -= ret as usize;
                foffset += ret as u64;
                ptr += ret as u64;
                done += ret as u64;
            }
        }

        Ok(done)
    }
}

#[derive(Copy, Clone, Versionize)]
#[repr(C, packed)]
pub struct VirtioFsConfig {
    pub tag: [u8; 36],
    pub num_request_queues: u32,
}

impl Default for VirtioFsConfig {
    fn default() -> Self {
        VirtioFsConfig {
            tag: [0; 36],
            num_request_queues: 0,
        }
    }
}

unsafe impl ByteValued for VirtioFsConfig {}

pub struct Fs {
    common: VirtioCommon,
    vu_common: VhostUserCommon,
    id: String,
    config: VirtioFsConfig,
    // Hold ownership of the memory that is allocated for the device
    // which will be automatically dropped when the device is dropped
    cache: Option<(VirtioSharedMemoryList, MmapRegion)>,
    slave_req_support: bool,
    seccomp_action: SeccompAction,
    guest_memory: Option<GuestMemoryAtomic<GuestMemoryMmap>>,
    epoll_thread: Option<thread::JoinHandle<()>>,
    exit_evt: EventFd,
    iommu: bool,
}

impl Fs {
    /// Create a new virtio-fs device.
    #[allow(clippy::too_many_arguments)]
    pub fn new(
        id: String,
        path: &str,
        tag: &str,
        req_num_queues: usize,
        queue_size: u16,
        cache: Option<(VirtioSharedMemoryList, MmapRegion)>,
        seccomp_action: SeccompAction,
        restoring: bool,
        exit_evt: EventFd,
        iommu: bool,
    ) -> Result<Fs> {
        let mut slave_req_support = false;

        // Calculate the actual number of queues needed.
        let num_queues = NUM_QUEUE_OFFSET + req_num_queues;

        if restoring {
            // We need 'queue_sizes' to report a number of queues that will be
            // enough to handle all the potential queues. VirtioPciDevice::new()
            // will create the actual queues based on this information.
            return Ok(Fs {
                common: VirtioCommon {
                    device_type: VirtioDeviceType::Fs as u32,
                    queue_sizes: vec![queue_size; num_queues],
                    paused_sync: Some(Arc::new(Barrier::new(2))),
                    min_queues: 1,
                    ..Default::default()
                },
                vu_common: VhostUserCommon {
                    socket_path: path.to_string(),
                    vu_num_queues: num_queues,
                    ..Default::default()
                },
                id,
                config: VirtioFsConfig::default(),
                cache,
                slave_req_support,
                seccomp_action,
                guest_memory: None,
                epoll_thread: None,
                exit_evt,
                iommu,
            });
        }

        // Connect to the vhost-user socket.
        let mut vu = VhostUserHandle::connect_vhost_user(false, path, num_queues as u64, false)?;

        // Filling device and vring features VMM supports.
        let avail_features = DEFAULT_VIRTIO_FEATURES;

        let mut avail_protocol_features = VhostUserProtocolFeatures::MQ
            | VhostUserProtocolFeatures::CONFIGURE_MEM_SLOTS
            | VhostUserProtocolFeatures::REPLY_ACK
            | VhostUserProtocolFeatures::INFLIGHT_SHMFD
            | VhostUserProtocolFeatures::LOG_SHMFD;
        let slave_protocol_features =
            VhostUserProtocolFeatures::SLAVE_REQ | VhostUserProtocolFeatures::SLAVE_SEND_FD;
        if cache.is_some() {
            avail_protocol_features |= slave_protocol_features;
        }

        let (acked_features, acked_protocol_features) =
            vu.negotiate_features_vhost_user(avail_features, avail_protocol_features)?;

        let backend_num_queues =
            if acked_protocol_features & VhostUserProtocolFeatures::MQ.bits() != 0 {
                vu.socket_handle()
                    .get_queue_num()
                    .map_err(Error::VhostUserGetQueueMaxNum)? as usize
            } else {
                DEFAULT_QUEUE_NUMBER
            };

        if num_queues > backend_num_queues {
            error!(
                "vhost-user-fs requested too many queues ({}) since the backend only supports {}\n",
                num_queues, backend_num_queues
            );
            return Err(Error::BadQueueNum);
        }

        if acked_protocol_features & slave_protocol_features.bits()
            == slave_protocol_features.bits()
        {
            slave_req_support = true;
        }

        // Create virtio-fs device configuration.
        let mut config = VirtioFsConfig::default();
        let tag_bytes_vec = tag.to_string().into_bytes();
        config.tag[..tag_bytes_vec.len()].copy_from_slice(tag_bytes_vec.as_slice());
        config.num_request_queues = req_num_queues as u32;

        Ok(Fs {
            common: VirtioCommon {
                device_type: VirtioDeviceType::Fs as u32,
                avail_features: acked_features,
                // If part of the available features that have been acked, the
                // PROTOCOL_FEATURES bit must be already set through the VIRTIO
                // acked features as we know the guest would never ack it, thus
                // the feature would be lost.
                acked_features: acked_features & VhostUserVirtioFeatures::PROTOCOL_FEATURES.bits(),
                queue_sizes: vec![queue_size; num_queues],
                paused_sync: Some(Arc::new(Barrier::new(2))),
                min_queues: 1,
                ..Default::default()
            },
            vu_common: VhostUserCommon {
                vu: Some(Arc::new(Mutex::new(vu))),
                acked_protocol_features,
                socket_path: path.to_string(),
                vu_num_queues: num_queues,
                ..Default::default()
            },
            id,
            config,
            cache,
            slave_req_support,
            seccomp_action,
            guest_memory: None,
            epoll_thread: None,
            exit_evt,
            iommu,
        })
    }

    fn state(&self) -> State {
        State {
            avail_features: self.common.avail_features,
            acked_features: self.common.acked_features,
            config: self.config,
            acked_protocol_features: self.vu_common.acked_protocol_features,
            vu_num_queues: self.vu_common.vu_num_queues,
            slave_req_support: self.slave_req_support,
        }
    }

    fn set_state(&mut self, state: &State) {
        self.common.avail_features = state.avail_features;
        self.common.acked_features = state.acked_features;
        self.config = state.config;
        self.vu_common.acked_protocol_features = state.acked_protocol_features;
        self.vu_common.vu_num_queues = state.vu_num_queues;
        self.slave_req_support = state.slave_req_support;

        if let Err(e) = self
            .vu_common
            .restore_backend_connection(self.common.acked_features)
        {
            error!(
                "Failed restoring connection with vhost-user backend: {:?}",
                e
            );
        }
    }
}

impl Drop for Fs {
    fn drop(&mut self) {
        if let Some(kill_evt) = self.common.kill_evt.take() {
            // Ignore the result because there is nothing we can do about it.
            let _ = kill_evt.write(1);
        }
    }
}

impl VirtioDevice for Fs {
    fn device_type(&self) -> u32 {
        self.common.device_type
    }

    fn queue_max_sizes(&self) -> &[u16] {
        &self.common.queue_sizes
    }

    fn features(&self) -> u64 {
        let mut features = self.common.avail_features;
        if self.iommu {
            features |= 1u64 << VIRTIO_F_IOMMU_PLATFORM;
        }
        features
    }

    fn ack_features(&mut self, value: u64) {
        self.common.ack_features(value)
    }

    fn read_config(&self, offset: u64, data: &mut [u8]) {
        self.read_config_from_slice(self.config.as_slice(), offset, data);
    }

    fn activate(
        &mut self,
        mem: GuestMemoryAtomic<GuestMemoryMmap>,
        interrupt_cb: Arc<dyn VirtioInterrupt>,
        queues: Vec<(usize, Queue, EventFd)>,
    ) -> ActivateResult {
        self.common.activate(&queues, &interrupt_cb)?;
        self.guest_memory = Some(mem.clone());

        // Initialize slave communication.
        let slave_req_handler = if self.slave_req_support {
            if let Some(cache) = self.cache.as_ref() {
                let vu_master_req_handler = Arc::new(SlaveReqHandler {
                    cache_offset: cache.0.addr,
                    cache_size: cache.0.len,
                    mmap_cache_addr: cache.0.host_addr,
                    mem: mem.clone(),
                });

                let mut req_handler =
                    MasterReqHandler::new(vu_master_req_handler).map_err(|e| {
                        ActivateError::VhostUserFsSetup(Error::MasterReqHandlerCreation(e))
                    })?;

                if self.vu_common.acked_protocol_features
                    & VhostUserProtocolFeatures::REPLY_ACK.bits()
                    != 0
                {
                    req_handler.set_reply_ack_flag(true);
                }

                Some(req_handler)
            } else {
                None
            }
        } else {
            None
        };

        // Run a dedicated thread for handling potential reconnections with
        // the backend.
        let (kill_evt, pause_evt) = self.common.dup_eventfds();

        let mut handler = self.vu_common.activate(
            mem,
            queues,
            interrupt_cb,
            self.common.acked_features,
            slave_req_handler,
            kill_evt,
            pause_evt,
        )?;

        let paused = self.common.paused.clone();
        let paused_sync = self.common.paused_sync.clone();

        let mut epoll_threads = Vec::new();
        spawn_virtio_thread(
            &self.id,
            &self.seccomp_action,
            Thread::VirtioVhostFs,
            &mut epoll_threads,
            &self.exit_evt,
            move || {
                if let Err(e) = handler.run(paused, paused_sync.unwrap()) {
                    error!("Error running worker: {:?}", e);
                }
            },
        )?;
        self.epoll_thread = Some(epoll_threads.remove(0));

        event!("virtio-device", "activated", "id", &self.id);
        Ok(())
    }

    fn reset(&mut self) -> Option<Arc<dyn VirtioInterrupt>> {
        // We first must resume the virtio thread if it was paused.
        if self.common.pause_evt.take().is_some() {
            self.common.resume().ok()?;
        }

        if let Some(vu) = &self.vu_common.vu {
            if let Err(e) = vu.lock().unwrap().reset_vhost_user() {
                error!("Failed to reset vhost-user daemon: {:?}", e);
                return None;
            }
        }

        if let Some(kill_evt) = self.common.kill_evt.take() {
            // Ignore the result because there is nothing we can do about it.
            let _ = kill_evt.write(1);
        }

        event!("virtio-device", "reset", "id", &self.id);

        // Return the interrupt
        Some(self.common.interrupt_cb.take().unwrap())
    }

    fn shutdown(&mut self) {
        self.vu_common.shutdown()
    }

    fn get_shm_regions(&self) -> Option<VirtioSharedMemoryList> {
        self.cache.as_ref().map(|cache| cache.0.clone())
    }

    fn set_shm_regions(
        &mut self,
        shm_regions: VirtioSharedMemoryList,
    ) -> std::result::Result<(), crate::Error> {
        if let Some(mut cache) = self.cache.as_mut() {
            cache.0 = shm_regions;
            Ok(())
        } else {
            Err(crate::Error::SetShmRegionsNotSupported)
        }
    }

    fn add_memory_region(
        &mut self,
        region: &Arc<GuestRegionMmap>,
    ) -> std::result::Result<(), crate::Error> {
        self.vu_common.add_memory_region(&self.guest_memory, region)
    }

    fn userspace_mappings(&self) -> Vec<UserspaceMapping> {
        let mut mappings = Vec::new();
        if let Some(cache) = self.cache.as_ref() {
            mappings.push(UserspaceMapping {
                host_addr: cache.0.host_addr,
                mem_slot: cache.0.mem_slot,
                addr: cache.0.addr,
                len: cache.0.len,
                mergeable: false,
            })
        }

        mappings
    }
}

impl Pausable for Fs {
    fn pause(&mut self) -> result::Result<(), MigratableError> {
        self.vu_common.pause()?;
        self.common.pause()
    }

    fn resume(&mut self) -> result::Result<(), MigratableError> {
        self.common.resume()?;

        if let Some(epoll_thread) = &self.epoll_thread {
            epoll_thread.thread().unpark();
        }

        self.vu_common.resume()
    }
}

impl Snapshottable for Fs {
    fn id(&self) -> String {
        self.id.clone()
    }

    fn snapshot(&mut self) -> std::result::Result<Snapshot, MigratableError> {
        self.vu_common.snapshot(&self.id(), &self.state())
    }

    fn restore(&mut self, snapshot: Snapshot) -> std::result::Result<(), MigratableError> {
        self.set_state(&snapshot.to_versioned_state(&self.id)?);
        Ok(())
    }
}
impl Transportable for Fs {}

impl Migratable for Fs {
    fn start_dirty_log(&mut self) -> std::result::Result<(), MigratableError> {
        self.vu_common.start_dirty_log(&self.guest_memory)
    }

    fn stop_dirty_log(&mut self) -> std::result::Result<(), MigratableError> {
        self.vu_common.stop_dirty_log()
    }

    fn dirty_log(&mut self) -> std::result::Result<MemoryRangeTable, MigratableError> {
        self.vu_common.dirty_log(&self.guest_memory)
    }

    fn start_migration(&mut self) -> std::result::Result<(), MigratableError> {
        self.vu_common.start_migration()
    }

    fn complete_migration(&mut self) -> std::result::Result<(), MigratableError> {
        self.vu_common
            .complete_migration(self.common.kill_evt.take())
    }
}