1 ==================================
2 VDUSE - "vDPA Device in Userspace"
3 ==================================
5 vDPA (virtio data path acceleration) device is a device that uses a
6 datapath which complies with the virtio specifications with vendor
7 specific control path. vDPA devices can be both physically located on
8 the hardware or emulated by software. VDUSE is a framework that makes it
9 possible to implement software-emulated vDPA devices in userspace. And
10 to make the device emulation more secure, the emulated vDPA device's
11 control path is handled in the kernel and only the data path is
12 implemented in the userspace.
14 Note that only virtio block device is supported by VDUSE framework now,
15 which can reduce security risks when the userspace process that implements
16 the data path is run by an unprivileged user. The support for other device
17 types can be added after the security issue of corresponding device driver
18 is clarified or fixed in the future.
20 Create/Destroy VDUSE devices
21 ----------------------------
23 VDUSE devices are created as follows:
25 1. Create a new VDUSE instance with ioctl(VDUSE_CREATE_DEV) on
28 2. Setup each virtqueue with ioctl(VDUSE_VQ_SETUP) on /dev/vduse/$NAME.
30 3. Begin processing VDUSE messages from /dev/vduse/$NAME. The first
31 messages will arrive while attaching the VDUSE instance to vDPA bus.
33 4. Send the VDPA_CMD_DEV_NEW netlink message to attach the VDUSE
36 VDUSE devices are destroyed as follows:
38 1. Send the VDPA_CMD_DEV_DEL netlink message to detach the VDUSE
39 instance from vDPA bus.
41 2. Close the file descriptor referring to /dev/vduse/$NAME.
43 3. Destroy the VDUSE instance with ioctl(VDUSE_DESTROY_DEV) on
46 The netlink messages can be sent via vdpa tool in iproute2 or use the
51 static int netlink_add_vduse(const char *name, enum vdpa_command cmd)
53 struct nl_sock *nlsock;
57 nlsock = nl_socket_alloc();
61 if (genl_connect(nlsock))
64 famid = genl_ctrl_resolve(nlsock, VDPA_GENL_NAME);
72 if (!genlmsg_put(msg, NL_AUTO_PORT, NL_AUTO_SEQ, famid, 0, 0, cmd, 0))
75 NLA_PUT_STRING(msg, VDPA_ATTR_DEV_NAME, name);
76 if (cmd == VDPA_CMD_DEV_NEW)
77 NLA_PUT_STRING(msg, VDPA_ATTR_MGMTDEV_DEV_NAME, "vduse");
79 if (nl_send_sync(nlsock, msg))
83 nl_socket_free(nlsock);
91 nl_socket_free(nlsock);
98 As mentioned above, a VDUSE device is created by ioctl(VDUSE_CREATE_DEV) on
99 /dev/vduse/control. With this ioctl, userspace can specify some basic configuration
100 such as device name (uniquely identify a VDUSE device), virtio features, virtio
101 configuration space, the number of virtqueues and so on for this emulated device.
102 Then a char device interface (/dev/vduse/$NAME) is exported to userspace for device
103 emulation. Userspace can use the VDUSE_VQ_SETUP ioctl on /dev/vduse/$NAME to
104 add per-virtqueue configuration such as the max size of virtqueue to the device.
106 After the initialization, the VDUSE device can be attached to vDPA bus via
107 the VDPA_CMD_DEV_NEW netlink message. Userspace needs to read()/write() on
108 /dev/vduse/$NAME to receive/reply some control messages from/to VDUSE kernel
113 static int vduse_message_handler(int dev_fd)
116 struct vduse_dev_request req;
117 struct vduse_dev_response resp;
119 len = read(dev_fd, &req, sizeof(req));
120 if (len != sizeof(req))
123 resp.request_id = req.request_id;
127 /* handle different types of messages */
131 len = write(dev_fd, &resp, sizeof(resp));
132 if (len != sizeof(resp))
138 There are now three types of messages introduced by VDUSE framework:
140 - VDUSE_GET_VQ_STATE: Get the state for virtqueue, userspace should return
141 avail index for split virtqueue or the device/driver ring wrap counters and
142 the avail and used index for packed virtqueue.
144 - VDUSE_SET_STATUS: Set the device status, userspace should follow
145 the virtio spec: https://docs.oasis-open.org/virtio/virtio/v1.1/virtio-v1.1.html
146 to process this message. For example, fail to set the FEATURES_OK device
147 status bit if the device can not accept the negotiated virtio features
148 get from the VDUSE_DEV_GET_FEATURES ioctl.
150 - VDUSE_UPDATE_IOTLB: Notify userspace to update the memory mapping for specified
151 IOVA range, userspace should firstly remove the old mapping, then setup the new
152 mapping via the VDUSE_IOTLB_GET_FD ioctl.
154 After DRIVER_OK status bit is set via the VDUSE_SET_STATUS message, userspace is
155 able to start the dataplane processing as follows:
157 1. Get the specified virtqueue's information with the VDUSE_VQ_GET_INFO ioctl,
158 including the size, the IOVAs of descriptor table, available ring and used ring,
159 the state and the ready status.
161 2. Pass the above IOVAs to the VDUSE_IOTLB_GET_FD ioctl so that those IOVA regions
162 can be mapped into userspace. Some sample codes is shown below:
166 static int perm_to_prot(uint8_t perm)
171 case VDUSE_ACCESS_WO:
174 case VDUSE_ACCESS_RO:
177 case VDUSE_ACCESS_RW:
178 prot |= PROT_READ | PROT_WRITE;
185 static void *iova_to_va(int dev_fd, uint64_t iova, uint64_t *len)
190 struct vduse_iotlb_entry entry;
196 * Find the first IOVA region that overlaps with the specified
197 * range [start, last] and return the corresponding file descriptor.
199 fd = ioctl(dev_fd, VDUSE_IOTLB_GET_FD, &entry);
203 size = entry.last - entry.start + 1;
204 *len = entry.last - iova + 1;
205 addr = mmap(0, size, perm_to_prot(entry.perm), MAP_SHARED,
208 if (addr == MAP_FAILED)
212 * Using some data structures such as linked list to store
213 * the iotlb mapping. The munmap(2) should be called for the
214 * cached mapping when the corresponding VDUSE_UPDATE_IOTLB
215 * message is received or the device is reset.
218 return addr + iova - entry.start;
221 3. Setup the kick eventfd for the specified virtqueues with the VDUSE_VQ_SETUP_KICKFD
222 ioctl. The kick eventfd is used by VDUSE kernel module to notify userspace to
223 consume the available ring. This is optional since userspace can choose to poll the
224 available ring instead.
226 4. Listen to the kick eventfd (optional) and consume the available ring. The buffer
227 described by the descriptors in the descriptor table should be also mapped into
228 userspace via the VDUSE_IOTLB_GET_FD ioctl before accessing.
230 5. Inject an interrupt for specific virtqueue with the VDUSE_INJECT_VQ_IRQ ioctl
231 after the used ring is filled.
233 For more details on the uAPI, please see include/uapi/linux/vduse.h.