It was defined in tap.h, put it on its own so that future callers won't need to depend on all definitions in tap.h Signed-off-by: Ammar Yasser --- tap.h | 9 +-------- tap_hdr.h | 19 +++++++++++++++++++ 2 files changed, 20 insertions(+), 8 deletions(-) create mode 100644 tap_hdr.h diff --git a/tap.h b/tap.h index b335933..1625975 100644 --- a/tap.h +++ b/tap.h @@ -10,6 +10,7 @@ #include #include "passt.h" +#include "tap_hdr.h" /** L2_MAX_LEN_PASTA - Maximum frame length for pasta mode (with L2 header) * @@ -38,14 +39,6 @@ struct udphdr; -/** - * struct tap_hdr - tap backend specific headers - * @vnet_len: Frame length (for qemu socket transport) - */ -struct tap_hdr { - uint32_t vnet_len; -} __attribute__((packed)); - /** * tap_hdr_iov() - struct iovec for a tap header * @c: Execution context diff --git a/tap_hdr.h b/tap_hdr.h new file mode 100644 index 0000000..15bdb5b --- /dev/null +++ b/tap_hdr.h @@ -0,0 +1,19 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later + * Copyright (c) 2021 Red Hat GmbH + * Author: Stefano Brivio + */ + +#ifndef TAP_HDR_H +#define TAP_HDR_H + +#include + +/** + * struct tap_hdr - tap backend specific headers + * @vnet_len: Frame length (for qemu socket transport) + */ +struct tap_hdr { + uint32_t vnet_len; +} __attribute__((packed)); + +#endif /* TAP_HDR_H */ -- 2.34.1

The --vhost controls whether or no to use vhost acceleration for pasta. The EPOLL_TYPE_VHOST_CALL means that kernel wants to notify us about data it has written. EPOLL_TYPE_VHOST_ERROR means the kernel encountered an internal error on vhost and wants to notify us that something went wrong. Add the fd_vhost field on the context which will carry the file descriptor of the device and will indicate that the setup was successful. The vq field contains the kick, call and err file descriptors for both queues. Also add vhost (mark if vhost acceleration was requested) and virtio_features field. Signed-off-by: Ammar Yasser --- conf.c | 6 ++++++ epoll_type.h | 4 ++++ passt.c | 3 +++ passt.h | 15 +++++++++++++++ 4 files changed, 28 insertions(+) diff --git a/conf.c b/conf.c index faf2681..c31546e 100644 --- a/conf.c +++ b/conf.c @@ -736,6 +736,7 @@ pasta_opts: " default: auto\n" " --host-lo-to-ns-lo Translate host-loopback forwards to\n" " namespace loopback\n" + " --vhost\t\tUse vhost-kernel acceleration\n" " --userns NSPATH Target user namespace to join\n" " --netns PATH|NAME Target network namespace to join\n" " --netns-only Don't join existing user namespace\n" @@ -766,6 +767,7 @@ enum passt_modes conf_mode(int argc, char *argv[]) int vhost_user = 0; const struct option optvu[] = { {"vhost-user", no_argument, &vhost_user, 1 }, + {"vhost", no_argument, NULL, 0 }, { 0 }, }; char argv0[PATH_MAX], *basearg0; @@ -1309,6 +1311,7 @@ void conf(struct ctx *c, int argc, char **argv) {"ipv4-only", no_argument, NULL, '4' }, {"ipv6-only", no_argument, NULL, '6' }, {"one-off", no_argument, NULL, '1' }, + {"vhost", no_argument, &c->vhost, 1 }, {"tcp-ports", required_argument, NULL, 't' }, {"udp-ports", required_argument, NULL, 'u' }, {"tcp-ns", required_argument, NULL, 'T' }, @@ -1838,6 +1841,9 @@ void conf(struct ctx *c, int argc, char **argv) die("--no-copy-addrs needs --config-net"); } + if (c->vhost && c->mode != MODE_PASTA) + die("--vhost is only available in pasta mode"); + if (c->mode == MODE_PASTA && c->splice_only) { if (c->no_splice) die("--splice-only is incompatible with --no-splice"); diff --git a/epoll_type.h b/epoll_type.h index 061325a..e3206d1 100644 --- a/epoll_type.h +++ b/epoll_type.h @@ -50,6 +50,10 @@ enum epoll_type { EPOLL_TYPE_CONF_LISTEN, /* Configuration socket */ EPOLL_TYPE_CONF, + /* vhost-kernel call socket */ + EPOLL_TYPE_VHOST_CALL, + /* vhost-kernel error socket */ + EPOLL_TYPE_VHOST_ERROR, EPOLL_NUM_TYPES, }; diff --git a/passt.c b/passt.c index 5054551..865b331 100644 --- a/passt.c +++ b/passt.c @@ -65,6 +65,7 @@ char pkt_buf[PKT_BUF_BYTES] __attribute__ ((aligned(PAGE_SIZE))); struct ctx passt_ctx = { .pidfile_fd = -1, .fd_tap = -1, + .fd_vhost = -1, .fd_tap_listen = -1, .fd_control_listen = -1, .fd_repair_listen = -1, @@ -92,6 +93,8 @@ char *epoll_type_str[] = { [EPOLL_TYPE_NL_NEIGH] = "netlink neighbour notifier socket", [EPOLL_TYPE_CONF_LISTEN] = "configuration listening socket", [EPOLL_TYPE_CONF] = "configuration socket", + [EPOLL_TYPE_VHOST_CALL] = "vhost-kernel call socket", + [EPOLL_TYPE_VHOST_ERROR] = "vhost-kernel error socket", }; static_assert(ARRAY_SIZE(epoll_type_str) == EPOLL_NUM_TYPES, "epoll_type_str[] doesn't match enum epoll_type"); diff --git a/passt.h b/passt.h index 51ccd4f..c729316 100644 --- a/passt.h +++ b/passt.h @@ -183,8 +183,11 @@ struct ip6_ctx { * @fd_repair_listen: File descriptor for listening TCP_REPAIR socket, if any * @fd_repair: Connected AF_UNIX socket for TCP_REPAIR helper * @our_tap_mac: Pasta/passt's MAC on the tap link + * @fd_vhost: File descriptor for /dev/vhost-net, set on vhost setup * @guest_mac: MAC address of guest or namespace, seen or configured * @hash_secret: 128-bit secret for siphash functions + * @virtio_features: Negotiated virtio feature bits + * @vhost: Enable vhost-kernel acceleration for pasta * @ifi4: Template interface for IPv4, -1: none, 0: IPv4 disabled * @ip4: IPv4 configuration * @dns_search: DNS search list @@ -202,6 +205,8 @@ struct ip6_ctx { * @no_udp: Disable UDP operation * @udp: Context for UDP protocol handler * @no_icmp: Disable ICMP operation + * @vq: Per-virtqueue eventfd descriptors for vhost-kernel + * ([0] is RX, [1] is TX; each holds kick_fd, call_fd, err_fd) * @mtu: MTU passed via DHCP/NDP * @no_dns: Do not source/use DNS servers for any purpose * @no_dns_search: Do not source/use domain search lists for any purpose @@ -258,11 +263,14 @@ struct ctx { int fd_control; int fd_repair_listen; int fd_repair; + + int fd_vhost; unsigned char our_tap_mac[ETH_ALEN]; unsigned char guest_mac[ETH_ALEN]; uint16_t mtu; uint64_t hash_secret[2]; + uint64_t virtio_features; int ifi4; struct ip4_ctx ip4; @@ -288,6 +296,12 @@ struct ctx { struct udp_ctx udp; int no_icmp; + struct { + int kick_fd; + int call_fd; + int err_fd; + } vq[2]; + int no_dns; int no_dns_search; int no_dhcp_dns; @@ -300,6 +314,7 @@ struct ctx { int splice_only; int host_lo_to_ns_lo; int freebind; + int vhost; bool chroot_fallback; int low_wmem; -- 2.34.1

- setup_vhost_net: try to open /dev/vhost-net and negotiate the needed features on the device file descriptor. Set the descriptor on c->fd_host on success. - setup_eventfds: setup the call, kick and err files for a given queue_index. - setup_memory_table: create the memory table object that gets shared with the kernel so it writes directly to the global pkt_buf on rx, and when we receive data from the host and eventually write it the protocol specific buffers the kernel is able to read from them. we share only buffers for tcp and udp as they are the only two protocols that will have vhost acceleration. besides, other protocols don't have a global buffer from which they allocate packets to send to the guest and would be trickier to support Signed-off-by: Ammar Yasser --- virtio.c | 283 ++++++++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 282 insertions(+), 1 deletion(-) diff --git a/virtio.c b/virtio.c index d7016cc..d89d485 100644 --- a/virtio.c +++ b/virtio.c @@ -72,19 +72,47 @@ * SUCH DAMAGE. */ +#include "passt.h" #include +#include #include #include #include #include #include +#include #include +#include +#include +#include +#include +#include +#include +#include #include "util.h" #include "virtio.h" #include "vhost_user.h" +#include "tcp_buf.h" +#include "epoll_ctl.h" +#include "udp.h" + + +struct vq_state vqs[2]; + +struct vring_desc vring_desc[2][VHOST_NDESCS] __attribute__((aligned(PAGE_SIZE))); + +#pragma GCC diagnostic push +#pragma GCC diagnostic ignored "-Wpedantic" +union vring_avail_u vring_avail_all[2] __attribute__((aligned(PAGE_SIZE))); +union vring_used_u vring_used_all[2] __attribute__((aligned(PAGE_SIZE))); +#pragma GCC diagnostic pop -#define VIRTQUEUE_MAX_SIZE 1024 +union vhost_memory_u vhost_memory = { + .mem = { + .nregions = N_VHOST_REGIONS, + }, +}; /** * vu_gpa_to_va() - Translate guest physical address to our virtual address. @@ -766,3 +794,256 @@ void vu_queue_flush(const struct vu_dev *vdev, struct vu_virtq *vq, if ((uint16_t)(new - vq->signalled_used) < (uint16_t)(new - old)) vq->signalled_used_valid = false; } + +/** + * setup_vhost_net() - Open and negotiate features on /dev/vhost-net + * @c: Execution context; c->fd_vhost is set on success + * + */ +void setup_vhost_net(struct ctx *c) +{ + static const uint64_t req_features = + (1ULL << VIRTIO_F_VERSION_1) | (1ULL << VHOST_NET_F_VIRTIO_NET_HDR); + int vhost_fd, rc; + + vhost_fd = open("/dev/vhost-net", O_RDWR | O_NONBLOCK | O_CLOEXEC); + if (vhost_fd < 0) + die_perror("Failed to open /dev/vhost-net"); + + rc = ioctl(vhost_fd, VHOST_SET_OWNER, NULL); + if (rc < 0) + die_perror("VHOST_SET_OWNER ioctl on /dev/vhost-net failed"); + + rc = ioctl(vhost_fd, VHOST_GET_FEATURES, &c->virtio_features); + if (rc < 0) + die_perror("VHOST_GET_FEATURES ioctl on /dev/vhost-net failed"); + + debug("vhost features: %lx", c->virtio_features); + debug("req features: %lx", req_features); + + c->virtio_features &= req_features; + if (c->virtio_features != req_features) + die("vhost does not support required features"); + + rc = ioctl(vhost_fd, VHOST_SET_FEATURES, &c->virtio_features); + if (rc < 0) + die_perror("VHOST_SET_FEATURES ioctl on /dev/vhost-net failed"); + + c->fd_vhost = vhost_fd; +} + +/** + * setup_eventfds() - Set up call/kick eventfds and vring size for one queue + * @c: Execution context; c->fd_vhost must already be set + * @queue_idx: Index of the queue (vring) to configure + * + */ +void setup_eventfds(struct ctx *c, int queue_idx) +{ + int vhost_fd = c->fd_vhost; + struct vhost_vring_file call_file = { .index = queue_idx }; + struct vhost_vring_file kick_file = { .index = queue_idx }; + struct vhost_vring_file err_file = { .index = queue_idx }; + + struct vhost_vring_state state = { + .index = queue_idx, + .num = VHOST_NDESCS, + }; + union epoll_ref ref = { + .type = EPOLL_TYPE_VHOST_CALL, + .queue = queue_idx, + }; + struct epoll_event ev; + int rc; + + call_file.fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC); + if (call_file.fd < 0) + die_perror("Failed to create call eventfd"); + ref.fd = call_file.fd; + + rc = ioctl(vhost_fd, VHOST_SET_VRING_CALL, &call_file); + if (rc < 0) + die_perror("VHOST_SET_VRING_CALL ioctl on /dev/vhost-net failed"); + + ev = (struct epoll_event){ .data.u64 = ref.u64, .events = EPOLLIN }; + rc = epoll_ctl(c->epollfd, EPOLL_CTL_ADD, ref.fd, &ev); + if (rc < 0) + die_perror("Failed to add call eventfd to epoll"); + c->vq[queue_idx].call_fd = call_file.fd; + + err_file.fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC); + if (err_file.fd < 0) + die_perror("Failed to create error eventfd"); + + rc = ioctl(vhost_fd, VHOST_SET_VRING_ERR, &err_file); + if (rc < 0) + die_perror("VHOST_SET_VRING_ERR ioctl on /dev/vhost-net failed"); + + ref.type = EPOLL_TYPE_VHOST_ERROR; + ref.fd = err_file.fd; + ev.data.u64 = ref.u64; + rc = epoll_ctl(c->epollfd, EPOLL_CTL_ADD, ref.fd, &ev); + if (rc < 0) + die_perror("Failed to add error eventfd to epoll"); + c->vq[queue_idx].err_fd = err_file.fd; + + rc = ioctl(vhost_fd, VHOST_SET_VRING_NUM, &state); + if (rc < 0) { + die_perror("VHOST_SET_VRING_NUM ioctl on /dev/vhost-net failed (queue %d)", + queue_idx); + } + + kick_file.fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC); + if (kick_file.fd < 0) + die_perror("Failed to create kick eventfd"); + + rc = ioctl(vhost_fd, VHOST_SET_VRING_KICK, &kick_file); + if (rc < 0) { + die_perror("VHOST_SET_VRING_KICK ioctl on /dev/vhost-net failed (queue %d)", + queue_idx); + } + + c->vq[queue_idx].kick_fd = kick_file.fd; + + vqs[queue_idx].num_free = VHOST_NDESCS; +} + +/** + * setup_memory_table() - Register the GPA/HVA translation table + * @c: Execution context; c->fd_vhost must already be set + * + * pasta has no real guest, so container->host addresses are 1:1 and can + * be interpreted directly rather than translated. + */ +int setup_memory_table(struct ctx *c) { +#define VHOST_MEMORY_REGION_PTR(addr, size) \ + (struct vhost_memory_region) { \ + .guest_phys_addr = (uintptr_t)addr, \ + .memory_size = size, \ + .userspace_addr = (uintptr_t)addr, \ + } +#define VHOST_MEMORY_REGION(elem) VHOST_MEMORY_REGION_PTR(&elem, sizeof(elem)) + + /* general purpose buffers */ + vhost_memory.mem.regions[0] = VHOST_MEMORY_REGION(pkt_buf); + vhost_memory.mem.regions[1] = VHOST_MEMORY_REGION(eth_pad); + + /* tcp specific buffers */ + vhost_memory.mem.regions[2] = VHOST_MEMORY_REGION(tcp_payload_tap_hdr); + vhost_memory.mem.regions[3] = VHOST_MEMORY_REGION(tcp4_payload_ip); + vhost_memory.mem.regions[4] = VHOST_MEMORY_REGION(tcp6_payload_ip); + vhost_memory.mem.regions[5] = VHOST_MEMORY_REGION(tcp_payload); + vhost_memory.mem.regions[6] = VHOST_MEMORY_REGION(tcp_eth_hdr); + + /* udp specific buffers */ + vhost_memory.mem.regions[7] = VHOST_MEMORY_REGION(udp_payload); + vhost_memory.mem.regions[8] = VHOST_MEMORY_REGION(udp_eth_hdr); + vhost_memory.mem.regions[9] = VHOST_MEMORY_REGION(udp_iov_recv); + vhost_memory.mem.regions[10] = VHOST_MEMORY_REGION(udp_mh_recv); + vhost_memory.mem.regions[11] = VHOST_MEMORY_REGION(udp_meta); + vhost_memory.mem.nregions = 12; + + return ioctl(c->fd_vhost, VHOST_SET_MEM_TABLE, &vhost_memory.mem); +} + + + +/** + * set_vring_for_queue() - Register a vring's addresses and bind its backend + * @c: Execution context; c->fd_vhost must already be set + * @queue_idx: Index of the queue (vring) to configure + * @tap_fd: Tap fd to bind as this queue's backend + * + */ +void set_vring_for_queue(struct ctx *c, int queue_idx, int tap_fd) +{ + int vhost_fd = c->fd_vhost; + struct vhost_vring_addr addr = { + .index = queue_idx, + .desc_user_addr = (unsigned long)vring_desc[queue_idx], + .avail_user_addr = (unsigned long)&vring_avail_all[queue_idx], + .used_user_addr = (unsigned long)&vring_used_all[queue_idx], + .log_guest_addr = (unsigned long)&vring_used_all[queue_idx], + }; + struct vhost_vring_file file = { + .index = queue_idx, + .fd = tap_fd, + }; + unsigned int i; + int rc; + + debug("qid: %d", queue_idx); + debug("vhost desc addr: 0x%llx", addr.desc_user_addr); + debug("vhost avail addr: 0x%llx", addr.avail_user_addr); + debug("vhost used addr: 0x%llx", addr.used_user_addr); + + rc = ioctl(vhost_fd, VHOST_SET_VRING_ADDR, &addr); + if (rc < 0) + die_perror("VHOST_SET_VRING_ADDR ioctl on /dev/vhost-net failed"); + + if (queue_idx == 0) { + for (i = 0; i < VHOST_NDESCS; ++i) { + vring_desc[0][i].addr = (uintptr_t)pkt_buf + + i * (PKT_BUF_BYTES / VHOST_NDESCS); + vring_desc[0][i].len = PKT_BUF_BYTES / VHOST_NDESCS; + vring_desc[0][i].flags = VRING_DESC_F_WRITE; + } + + for (i = 0; i < VHOST_NDESCS; ++i) + vring_avail_all[0].avail.ring[i] = htole16(i); + + rx_descriptor_handoff(c); + } + + if (queue_idx == 1) { + for (i = 0; i < (VHOST_NDESCS - 1); ++i) { + vring_desc[1][i].next = i+1; + } + } + + debug("qid: %d", file.index); + debug("tap fd: %d", file.fd); + rc = ioctl(vhost_fd, VHOST_NET_SET_BACKEND, &file); + if (rc < 0) + die_perror("VHOST_NET_SET_BACKEND ioctl on /dev/vhost-net failed"); +} + +/** + * rx_descriptor_handoff() - Batch-announce freed RX descriptors to the kernel + * @c: Execution context + * + * Bumps avail.idx by the number of descriptors accumulated in + * vqs[0].num_free (from prior consume_one_rx_descriptor() calls), + * then resets the counter to zero. The kernel will see the new + * avail.idx and consume the freshly-available descriptors. + * + */ +void rx_descriptor_handoff(struct ctx *c) +{ + smp_wmb(); + + if (!vqs[0].num_free) + return; + + vring_avail_all[0].avail.idx += vqs[0].num_free; + vqs[0].num_free = 0; + vhost_kick(&vring_used_all[0].used, c->vq[0].kick_fd); +} + + +/** + * vhost_kick() - Notify the kernel that new descriptors are available + * @used: The vring_used queue. Taken as a parameter to check if the kernel + * virtio thread is actively reading descriptors or no + * @kick_fd: Which fd to notify about (will differ by which queue we are about + * to announce availability in) + */ +void vhost_kick(struct vring_used *used, int kick_fd) { + /* Ensure that the read to used->flags doesn't get reordered to be + * above the avail.idx update + */ + smp_mb(); + + if (!(used->flags & VRING_USED_F_NO_NOTIFY)) + eventfd_write(kick_fd, 1); +} \ No newline at end of file -- 2.34.1

- tap_send_single was changed to directly call tap_send_frames_passt/pasta instead of relying on tap_send_frames because protocols that use tap_send_single won't use vhost acceleration, only tcp and udp will. The function was also moved lower in the file to by the time its called tap_send_frames_* functions are defined and a forward declaration isn't needed - extend udp_meta_t to instead of always carrying a tap_hdr, to carry a union of a tap_hdr and a virtio_net header because if vhost is used the will the latter type of header and then build an iov from this virtio net header in udp.c - create iov_from_virtio_net_hdr to build an iov from the vnet header - in the case that vhost is requested, set the first frame fragment to be a virtio_net header through calling iov_from_virtio_net_hdr - convert the tcp_payload_tap_hdr array to be an array of virtio_net header. in the case where we aren't using vhost, we will just be using the first 4 bytes of it - add a boolean argumen to tap_send_frames_pasta to determine if vhost will be used Signed-off-by: Ammar Yasser --- tap.c | 100 +++++++++++++++++++++++++++++++----------------------- tcp_buf.c | 39 +++++++++++++++++---- tcp_buf.h | 4 +-- udp.c | 28 ++++++++++++++- udp.h | 6 +++- 5 files changed, 125 insertions(+), 52 deletions(-) diff --git a/tap.c b/tap.c index 9d601a2..e177eef 100644 --- a/tap.c +++ b/tap.c @@ -124,45 +124,6 @@ unsigned long tap_l2_max_len(const struct ctx *c) return 0; /* Unreachable, for cppcheck's sake */ } -/** - * tap_send_single() - Send a single frame - * @c: Execution context - * @data: Packet buffer - * @l2len: Total L2 packet length - */ -void tap_send_single(const struct ctx *c, const void *data, size_t l2len) -{ - uint8_t padded[ETH_ZLEN] = { 0 }; - struct iovec iov[2]; - size_t iovcnt = 0; - uint32_t vnet_len; - - if (l2len < ETH_ZLEN) { - memcpy(padded, data, l2len); - data = padded; - l2len = ETH_ZLEN; - } - - vnet_len = htonl(l2len); - - switch (c->mode) { - case MODE_PASST: - iov[iovcnt] = IOV_OF_LVALUE(vnet_len); - iovcnt++; - /* fall through */ - case MODE_PASTA: - iov[iovcnt].iov_base = (void *)data; - iov[iovcnt].iov_len = l2len; - iovcnt++; - - tap_send_frames(c, iov, iovcnt, 1); - break; - case MODE_VU: - vu_send_single(c, data, l2len); - break; - } -} - /** * tap_push_l2h() - Build an L2 header for an inbound packet * @c: Execution context @@ -415,7 +376,7 @@ void tap_icmp6_send(const struct ctx *c, */ static size_t tap_send_frames_pasta(const struct ctx *c, const struct iovec *iov, - size_t bufs_per_frame, size_t nframes) + size_t bufs_per_frame, size_t nframes, bool vhost) { size_t nbufs = bufs_per_frame * nframes; size_t i; @@ -498,6 +459,59 @@ static size_t tap_send_frames_passt(const struct ctx *c, return i / bufs_per_frame; } +/** + * tap_send_single() - Send a single frame + * @c: Execution context + * @data: Packet buffer + * @l2len: Total L2 packet length + */ +void tap_send_single(const struct ctx *c, const void *data, size_t l2len) +{ + uint8_t padded[ETH_ZLEN] = { 0 }; + struct iovec iov[2]; + size_t iovcnt = 0; + int m = 0; + uint32_t vnet_len; + + if (l2len < ETH_ZLEN) { + memcpy(padded, data, l2len); + data = padded; + l2len = ETH_ZLEN; + } + + vnet_len = htonl(l2len); + switch (c->mode) { + case MODE_PASST: + /* create an iov for the length */ + iov[iovcnt] = IOV_OF_LVALUE(vnet_len); + iovcnt++; + /* create the data iov */ + iov[iovcnt].iov_base = (void *)data; + iov[iovcnt].iov_len = l2len; + iovcnt++; + m = tap_send_frames_passt(c, iov, iovcnt, 1); + break; + case MODE_PASTA: + /* don't create a length iov in the case of pasta */ + iov[iovcnt].iov_base = (void *)data; + iov[iovcnt].iov_len = l2len; + iovcnt++; + + m = tap_send_frames_pasta(c, iov, iovcnt, 1, false); + break; + case MODE_VU: + m = vu_send_single(c, data, l2len); + break; + } + + if (m < 1) + debug("tap: failed to send a single frame"); + + pcap_multiple(iov, iovcnt, m, + c->mode == MODE_PASST ? sizeof(uint32_t) : + (c->fd_vhost != -1) ? VNET_HLEN : 0); +} + /** * tap_send_frames() - Send out multiple prepared frames * @c: Execution context @@ -523,7 +537,8 @@ size_t tap_send_frames(const struct ctx *c, const struct iovec *iov, switch (c->mode) { case MODE_PASTA: - m = tap_send_frames_pasta(c, iov, bufs_per_frame, nframes); + /* use vhost in pasta sending only if the vhost setup was actually successful, indicated by a non-zero fd_vhost */ + m = tap_send_frames_pasta(c, iov, bufs_per_frame, nframes, ((c->fd_vhost != -1))); break; case MODE_PASST: m = tap_send_frames_passt(c, iov, bufs_per_frame, nframes); @@ -539,7 +554,8 @@ size_t tap_send_frames(const struct ctx *c, const struct iovec *iov, nframes - m, nframes); pcap_multiple(iov, bufs_per_frame, m, - c->mode == MODE_PASST ? sizeof(uint32_t) : 0); + c->mode == MODE_PASST ? sizeof(uint32_t) : + (c->fd_vhost != -1) ? VNET_HLEN : 0); return m; } diff --git a/tcp_buf.c b/tcp_buf.c index 4452337..080e4cb 100644 --- a/tcp_buf.c +++ b/tcp_buf.c @@ -21,6 +21,7 @@ #include #include +#include #include "util.h" #include "ip.h" @@ -38,7 +39,7 @@ /* Ethernet header for IPv4 and IPv6 frames */ struct ethhdr tcp_eth_hdr[TCP_FRAMES_MEM]; -struct tap_hdr tcp_payload_tap_hdr[TCP_FRAMES_MEM]; +struct virtio_net_hdr_mrg_rxbuf tcp_payload_tap_hdr[TCP_FRAMES_MEM]; /* IP headers for IPv4 and IPv6 */ struct iphdr tcp4_payload_ip[TCP_FRAMES_MEM]; @@ -71,6 +72,14 @@ void tcp_update_l2_buf(const unsigned char *eth_d) eth_update_mac(&tcp_eth_hdr[i], eth_d, NULL); } +static inline struct iovec iov_from_virtio_net_hdr(struct virtio_net_hdr_mrg_rxbuf *hdr) +{ + return (struct iovec){ + .iov_base = hdr, + .iov_len = sizeof(*hdr), + }; +} + /** * tcp_sock_iov_init() - Initialise scatter-gather L2 buffers for IPv4 sockets * @c: Execution context @@ -89,7 +98,14 @@ void tcp_sock_iov_init(const struct ctx *c) for (i = 0; i < TCP_FRAMES_MEM; i++) { struct iovec *iov = tcp_l2_iov[i]; - iov[TCP_IOV_TAP] = tap_hdr_iov(c, &tcp_payload_tap_hdr[i]); + /* If we are using pasta with vhost acceleration, the first entry in the tcp buffers + * should point to a virtio_net header, otherwise a tap header + */ + if ((c->fd_vhost != -1)) + iov[TCP_IOV_TAP] = iov_from_virtio_net_hdr(&tcp_payload_tap_hdr[i]); + else + iov[TCP_IOV_TAP] = tap_hdr_iov(c, (struct tap_hdr *)&tcp_payload_tap_hdr[i]); + iov[TCP_IOV_ETH].iov_len = sizeof(struct ethhdr); iov[TCP_IOV_PAYLOAD].iov_base = &tcp_payload[i]; iov[TCP_IOV_ETH_PAD].iov_base = eth_pad; @@ -135,8 +151,7 @@ void tcp_payload_flush(const struct ctx *c, const struct timespec *now) { size_t m; - m = tap_send_frames(c, &tcp_l2_iov[0][0], TCP_NUM_IOVS, - tcp_payload_used); + m = tap_send_frames(c, &tcp_l2_iov[0][0], TCP_NUM_IOVS, tcp_payload_used); if (m != tcp_payload_used) { tcp_revert_seq(c, &tcp_frame_conns[m], &tcp_l2_iov[m], tcp_payload_used - m, now); @@ -177,7 +192,7 @@ static void tcp_l2_buf_fill_headers(const struct ctx *c, { struct iov_tail tail = IOV_TAIL(&iov[TCP_IOV_PAYLOAD], 1, 0); struct tcphdr th_storage, *th = IOV_REMOVE_HEADER(&tail, th_storage); - struct tap_hdr *taph = iov[TCP_IOV_TAP].iov_base; + const struct flowside *tapside = TAPFLOW(conn); const struct in_addr *a4 = inany_v4(&tapside->oaddr); struct ethhdr *eh = iov[TCP_IOV_ETH].iov_base; @@ -192,7 +207,19 @@ static void tcp_l2_buf_fill_headers(const struct ctx *c, l2len = tcp_fill_headers(c, conn, eh, ip4h, ip6h, th, &tail, iov_tail_size(&tail), csum_flags, seq); - tap_hdr_update(taph, l2len); + + /* when in pasta mode, the length of the tap header iov is zero, so this + * l2len write doesn't do anything. But this function gets called from both + * pasta and passt. Make the l2len write in case we are in passt mode, denoted by + * fd_vhost being -1 since this field won't get initialized at all in passt. + */ + if (c->mode == MODE_PASST) { + iov[TCP_IOV_TAP].iov_len = sizeof(struct tap_hdr); + struct tap_hdr *taph = iov[TCP_IOV_TAP].iov_base; + tap_hdr_update(taph, l2len); + } else if (c->fd_vhost == -1) { /* pasta mode but without vhost */ + iov[TCP_IOV_TAP].iov_len = 0; + } } /** diff --git a/tcp_buf.h b/tcp_buf.h index c749038..da48cd0 100644 --- a/tcp_buf.h +++ b/tcp_buf.h @@ -9,7 +9,7 @@ #include "tcp_conn.h" #include "tcp_internal.h" -void tcp_sock_iov_init(); +void tcp_sock_iov_init(const struct ctx *c); void tcp_payload_flush(const struct ctx *c, const struct timespec *now); int tcp_buf_data_from_sock(const struct ctx *c, struct tcp_tap_conn *conn, uint32_t already_sent, const struct timespec *now); @@ -20,7 +20,7 @@ int tcp_buf_send_flag(const struct ctx *c, struct tcp_tap_conn *conn, int flags, #define TCP_FRAMES \ (c->mode == MODE_PASTA ? 1 : TCP_FRAMES_MEM) -extern struct tap_hdr tcp_payload_tap_hdr[TCP_FRAMES_MEM]; +extern struct virtio_net_hdr_mrg_rxbuf tcp_payload_tap_hdr[TCP_FRAMES_MEM]; extern struct ethhdr tcp_eth_hdr[TCP_FRAMES_MEM]; extern struct tcp_payload_t tcp_payload[TCP_FRAMES_MEM]; diff --git a/udp.c b/udp.c index d05ee66..5f90dce 100644 --- a/udp.c +++ b/udp.c @@ -103,6 +103,7 @@ #include #include #include +#include #include "checksum.h" #include "util.h" @@ -119,6 +120,18 @@ #include "udp_vu.h" #include "epoll_ctl.h" +/* UDP header and data for inbound messages */ +struct udp_payload_t udp_payload[UDP_MAX_FRAMES]; + +/* Ethernet headers for IPv4 and IPv6 frames */ +struct ethhdr udp_eth_hdr[UDP_MAX_FRAMES]; + +/* IOVs and msghdr arrays for receiving datagrams from sockets */ +struct iovec udp_iov_recv [UDP_MAX_FRAMES]; +struct mmsghdr udp_mh_recv [UDP_MAX_FRAMES]; + +/* Pre-cooked headers for UDP packets */ +struct udp_meta_t udp_meta[UDP_MAX_FRAMES]; #define UDP_TIMEOUT "/proc/sys/net/netfilter/nf_conntrack_udp_timeout" #define UDP_TIMEOUT_STREAM \ @@ -216,7 +229,20 @@ static void udp_iov_init_one(const struct ctx *c, size_t i) *siov = IOV_OF_LVALUE(payload->data); tiov[UDP_IOV_ETH] = IOV_OF_LVALUE(udp_eth_hdr[i]); - tiov[UDP_IOV_TAP] = tap_hdr_iov(c, &meta->taph); + /* if the vhost tap fd is initialized, this is a sign for us that + * we will be using virtio transport. make the iov that is supposed + * to point to a tap header point instead to a virtio_net_mrg_rxbuf + */ + if (c->fd_vhost != -1) { + struct iovec vnet_iov = { + .iov_base = (void *)(&meta->vnet_hdr), + .iov_len = sizeof(meta->vnet_hdr) + }; + tiov[UDP_IOV_TAP] = vnet_iov; + } else { + tiov[UDP_IOV_TAP] = tap_hdr_iov(c, &meta->taph); + } + tiov[UDP_IOV_PAYLOAD].iov_base = payload; tiov[UDP_IOV_ETH_PAD].iov_base = eth_pad; diff --git a/udp.h b/udp.h index b7a367c..3464b74 100644 --- a/udp.h +++ b/udp.h @@ -9,6 +9,7 @@ #include #include #include +#include #include "tap_hdr.h" #include "fwd.h" @@ -49,7 +50,10 @@ extern struct mmsghdr udp_mh_recv [UDP_MAX_FRAMES]; struct udp_meta_t { struct ipv6hdr ip6h; struct iphdr ip4h; - struct tap_hdr taph; + union { + struct virtio_net_hdr_mrg_rxbuf vnet_hdr; + struct tap_hdr taph; +}; #ifdef __AVX2__ } __attribute__((aligned(32))); #else -- 2.34.1

On Sun, Aug 02, 2026 at 01:21:48PM +0000, Ammar Yasser wrote:

It was defined in tap.h, put it on its own so that future callers won't need to depend on all definitions in tap.h

Signed-off-by: Ammar Yasser

Reviewed-by: David Gibson

--- tap.h | 9 +-------- tap_hdr.h | 19 +++++++++++++++++++ 2 files changed, 20 insertions(+), 8 deletions(-) create mode 100644 tap_hdr.h

diff --git a/tap.h b/tap.h index b335933..1625975 100644 --- a/tap.h +++ b/tap.h @@ -10,6 +10,7 @@ #include

#include "passt.h" +#include "tap_hdr.h"

/** L2_MAX_LEN_PASTA - Maximum frame length for pasta mode (with L2 header) * @@ -38,14 +39,6 @@

struct udphdr;

-/** - * struct tap_hdr - tap backend specific headers - * @vnet_len: Frame length (for qemu socket transport) - */ -struct tap_hdr { - uint32_t vnet_len; -} __attribute__((packed)); - /** * tap_hdr_iov() - struct iovec for a tap header * @c: Execution context diff --git a/tap_hdr.h b/tap_hdr.h new file mode 100644 index 0000000..15bdb5b --- /dev/null +++ b/tap_hdr.h @@ -0,0 +1,19 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later + * Copyright (c) 2021 Red Hat GmbH + * Author: Stefano Brivio + */ + +#ifndef TAP_HDR_H +#define TAP_HDR_H + +#include + +/** + * struct tap_hdr - tap backend specific headers + * @vnet_len: Frame length (for qemu socket transport) + */ +struct tap_hdr { + uint32_t vnet_len; +} __attribute__((packed)); + +#endif /* TAP_HDR_H */ -- 2.34.1

-- David Gibson (he or they) | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you, not the other way | around. http://www.ozlabs.org/~dgibson

On Sun, Aug 02, 2026 at 01:21:50PM +0000, Ammar Yasser wrote:

The --vhost controls whether or no to use vhost acceleration for pasta. The EPOLL_TYPE_VHOST_CALL means that kernel wants to notify us about data it has written. EPOLL_TYPE_VHOST_ERROR means the kernel encountered an internal error on vhost and wants to notify us that something went wrong.

Add the fd_vhost field on the context which will carry the file descriptor of the device and will indicate that the setup was successful. The vq field contains the kick, call and err file descriptors for both queues.

Also add vhost (mark if vhost acceleration was requested) and virtio_features field.

A couple of concerns about the command line interface. IIUC, once this is all ready the difference between vhost-kernel and regular tuntap should be neither guest-visible nor user-visible. So, assuming it does perform better, we'll probably want to make it the default. So, making this explicitly an opt-in probably isn't what we want. Maybe --vhost on|off|auto would be a better idea, defaulting to off now, but auto (use if available) in future. "on" would be if the user definitely wants vhost, and would rather we exit than fall back if it's not available. A more minor concern is that there's the potential confusion between this and --vhost-user. Not immediately sure what we can do about this, since that's just reflecting the confusingly similar names of the underlying features.

Signed-off-by: Ammar Yasser --- conf.c | 6 ++++++ epoll_type.h | 4 ++++ passt.c | 3 +++ passt.h | 15 +++++++++++++++ 4 files changed, 28 insertions(+)

diff --git a/conf.c b/conf.c index faf2681..c31546e 100644 --- a/conf.c +++ b/conf.c @@ -736,6 +736,7 @@ pasta_opts: " default: auto\n" " --host-lo-to-ns-lo Translate host-loopback forwards to\n" " namespace loopback\n" + " --vhost\t\tUse vhost-kernel acceleration\n" " --userns NSPATH Target user namespace to join\n" " --netns PATH|NAME Target network namespace to join\n" " --netns-only Don't join existing user namespace\n" @@ -766,6 +767,7 @@ enum passt_modes conf_mode(int argc, char *argv[]) int vhost_user = 0; const struct option optvu[] = { {"vhost-user", no_argument, &vhost_user, 1 }, + {"vhost", no_argument, NULL, 0 }, { 0 }, }; char argv0[PATH_MAX], *basearg0; @@ -1309,6 +1311,7 @@ void conf(struct ctx *c, int argc, char **argv) {"ipv4-only", no_argument, NULL, '4' }, {"ipv6-only", no_argument, NULL, '6' }, {"one-off", no_argument, NULL, '1' }, + {"vhost", no_argument, &c->vhost, 1 }, {"tcp-ports", required_argument, NULL, 't' }, {"udp-ports", required_argument, NULL, 'u' }, {"tcp-ns", required_argument, NULL, 'T' }, @@ -1838,6 +1841,9 @@ void conf(struct ctx *c, int argc, char **argv) die("--no-copy-addrs needs --config-net"); }

+ if (c->vhost && c->mode != MODE_PASTA) + die("--vhost is only available in pasta mode"); + if (c->mode == MODE_PASTA && c->splice_only) { if (c->no_splice) die("--splice-only is incompatible with --no-splice"); diff --git a/epoll_type.h b/epoll_type.h index 061325a..e3206d1 100644 --- a/epoll_type.h +++ b/epoll_type.h @@ -50,6 +50,10 @@ enum epoll_type { EPOLL_TYPE_CONF_LISTEN, /* Configuration socket */ EPOLL_TYPE_CONF, + /* vhost-kernel call socket */ + EPOLL_TYPE_VHOST_CALL, + /* vhost-kernel error socket */ + EPOLL_TYPE_VHOST_ERROR,

EPOLL_NUM_TYPES, }; diff --git a/passt.c b/passt.c index 5054551..865b331 100644 --- a/passt.c +++ b/passt.c @@ -65,6 +65,7 @@ char pkt_buf[PKT_BUF_BYTES] __attribute__ ((aligned(PAGE_SIZE))); struct ctx passt_ctx = { .pidfile_fd = -1, .fd_tap = -1, + .fd_vhost = -1, .fd_tap_listen = -1, .fd_control_listen = -1, .fd_repair_listen = -1, @@ -92,6 +93,8 @@ char *epoll_type_str[] = { [EPOLL_TYPE_NL_NEIGH] = "netlink neighbour notifier socket", [EPOLL_TYPE_CONF_LISTEN] = "configuration listening socket", [EPOLL_TYPE_CONF] = "configuration socket", + [EPOLL_TYPE_VHOST_CALL] = "vhost-kernel call socket", + [EPOLL_TYPE_VHOST_ERROR] = "vhost-kernel error socket", }; static_assert(ARRAY_SIZE(epoll_type_str) == EPOLL_NUM_TYPES, "epoll_type_str[] doesn't match enum epoll_type"); diff --git a/passt.h b/passt.h index 51ccd4f..c729316 100644 --- a/passt.h +++ b/passt.h @@ -183,8 +183,11 @@ struct ip6_ctx { * @fd_repair_listen: File descriptor for listening TCP_REPAIR socket, if any * @fd_repair: Connected AF_UNIX socket for TCP_REPAIR helper * @our_tap_mac: Pasta/passt's MAC on the tap link + * @fd_vhost: File descriptor for /dev/vhost-net, set on vhost setup * @guest_mac: MAC address of guest or namespace, seen or configured * @hash_secret: 128-bit secret for siphash functions + * @virtio_features: Negotiated virtio feature bits + * @vhost: Enable vhost-kernel acceleration for pasta * @ifi4: Template interface for IPv4, -1: none, 0: IPv4 disabled * @ip4: IPv4 configuration * @dns_search: DNS search list @@ -202,6 +205,8 @@ struct ip6_ctx { * @no_udp: Disable UDP operation * @udp: Context for UDP protocol handler * @no_icmp: Disable ICMP operation + * @vq: Per-virtqueue eventfd descriptors for vhost-kernel + * ([0] is RX, [1] is TX; each holds kick_fd, call_fd, err_fd) * @mtu: MTU passed via DHCP/NDP * @no_dns: Do not source/use DNS servers for any purpose * @no_dns_search: Do not source/use domain search lists for any purpose @@ -258,11 +263,14 @@ struct ctx { int fd_control; int fd_repair_listen; int fd_repair; + + int fd_vhost; unsigned char our_tap_mac[ETH_ALEN]; unsigned char guest_mac[ETH_ALEN]; uint16_t mtu;

uint64_t hash_secret[2]; + uint64_t virtio_features;

int ifi4; struct ip4_ctx ip4; @@ -288,6 +296,12 @@ struct ctx { struct udp_ctx udp; int no_icmp;

+ struct { + int kick_fd; + int call_fd; + int err_fd; + } vq[2]; + int no_dns; int no_dns_search; int no_dhcp_dns; @@ -300,6 +314,7 @@ struct ctx { int splice_only; int host_lo_to_ns_lo; int freebind; + int vhost; bool chroot_fallback;

int low_wmem;

It might be a bit neater to put the new vhost-kernel related fields into a substructure, rather than spreading them across struct ctx. At some point it would be nice to make the various tap backends a bit more pluggable / independent from each other. Doing that's obviously not in scope for this series, but keeping it's internal data all together will at least not make that job harder in the future.

-- 2.34.1

-- David Gibson (he or they) | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you, not the other way | around. http://www.ozlabs.org/~dgibson

On Sun, Aug 02, 2026 at 01:21:52PM +0000, Ammar Yasser wrote:

- setup_vhost_net: try to open /dev/vhost-net and negotiate the needed features on the device file descriptor. Set the descriptor on c->fd_host on success. - setup_eventfds: setup the call, kick and err files for a given queue_index. - setup_memory_table: create the memory table object that gets shared with the kernel so it writes directly to the global pkt_buf on rx, and when we receive data from the host and eventually write it the protocol specific buffers the kernel is able to read from them. we share only buffers for tcp and udp as they are the only two protocols that will have vhost acceleration. besides, other protocols don't have a global buffer from which they allocate packets to send to the guest and would be trickier to support

Signed-off-by: Ammar Yasser

Similar to notes on the earlier patches, I think this will be clearer with the new vhost-kernel related functions in new .c and .h files, rather than mixed in with the general virtio helpers.

--- virtio.c | 283 ++++++++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 282 insertions(+), 1 deletion(-)

diff --git a/virtio.c b/virtio.c index d7016cc..d89d485 100644 --- a/virtio.c +++ b/virtio.c @@ -72,19 +72,47 @@ * SUCH DAMAGE. */

+#include "passt.h" #include +#include #include #include #include #include #include +#include #include +#include +#include +#include +#include +#include +#include +#include

#include "util.h" #include "virtio.h" #include "vhost_user.h" +#include "tcp_buf.h" +#include "epoll_ctl.h" +#include "udp.h" + + +struct vq_state vqs[2];

'vqs' isn't really an adequate name for a global variable, especially since this is fully global, not 'static'.

+ +struct vring_desc vring_desc[2][VHOST_NDESCS] __attribute__((aligned(PAGE_SIZE))); + +#pragma GCC diagnostic push +#pragma GCC diagnostic ignored "-Wpedantic" +union vring_avail_u vring_avail_all[2] __attribute__((aligned(PAGE_SIZE))); +union vring_used_u vring_used_all[2] __attribute__((aligned(PAGE_SIZE))); +#pragma GCC diagnostic pop

-#define VIRTQUEUE_MAX_SIZE 1024 +union vhost_memory_u vhost_memory = { + .mem = { + .nregions = N_VHOST_REGIONS, + }, +};

/** * vu_gpa_to_va() - Translate guest physical address to our virtual address. @@ -766,3 +794,256 @@ void vu_queue_flush(const struct vu_dev *vdev, struct vu_virtq *vq, if ((uint16_t)(new - vq->signalled_used) < (uint16_t)(new - old)) vq->signalled_used_valid = false; } + +/** + * setup_vhost_net() - Open and negotiate features on /dev/vhost-net + * @c: Execution context; c->fd_vhost is set on success + * + */ +void setup_vhost_net(struct ctx *c) +{ + static const uint64_t req_features = + (1ULL << VIRTIO_F_VERSION_1) | (1ULL << VHOST_NET_F_VIRTIO_NET_HDR);

Since it's also 'const' anyway, I'm not sure the 'static' does anything useful.

+ int vhost_fd, rc; + + vhost_fd = open("/dev/vhost-net", O_RDWR | O_NONBLOCK | O_CLOEXEC); + if (vhost_fd < 0) + die_perror("Failed to open /dev/vhost-net");

We probably want to be able to fall back to the /dev/net/tun character device if vhost doesn't work. So making this setup function fallible would be preferable to using die() on errors.

+ rc = ioctl(vhost_fd, VHOST_SET_OWNER, NULL); + if (rc < 0) + die_perror("VHOST_SET_OWNER ioctl on /dev/vhost-net failed"); + + rc = ioctl(vhost_fd, VHOST_GET_FEATURES, &c->virtio_features); + if (rc < 0) + die_perror("VHOST_GET_FEATURES ioctl on /dev/vhost-net failed"); + + debug("vhost features: %lx", c->virtio_features); + debug("req features: %lx", req_features); + + c->virtio_features &= req_features; + if (c->virtio_features != req_features) + die("vhost does not support required features"); + + rc = ioctl(vhost_fd, VHOST_SET_FEATURES, &c->virtio_features); + if (rc < 0) + die_perror("VHOST_SET_FEATURES ioctl on /dev/vhost-net failed"); + + c->fd_vhost = vhost_fd; +} + +/** + * setup_eventfds() - Set up call/kick eventfds and vring size for one queue + * @c: Execution context; c->fd_vhost must already be set + * @queue_idx: Index of the queue (vring) to configure + * + */ +void setup_eventfds(struct ctx *c, int queue_idx) +{ + int vhost_fd = c->fd_vhost;

We use (when possible) the "reverse christmas tree" convention for ordering locals, which would but this further down (see CONTRIBUTING.md for more details).

+ struct vhost_vring_file call_file = { .index = queue_idx }; + struct vhost_vring_file kick_file = { .index = queue_idx }; + struct vhost_vring_file err_file = { .index = queue_idx }; + + struct vhost_vring_state state = { + .index = queue_idx, + .num = VHOST_NDESCS, + }; + union epoll_ref ref = { + .type = EPOLL_TYPE_VHOST_CALL, + .queue = queue_idx, + }; + struct epoll_event ev; + int rc; + + call_file.fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC); + if (call_file.fd < 0) + die_perror("Failed to create call eventfd");

This error isn't particularly meaningful to an end user, since it doesn't mention vhost-kernel at all.

+ ref.fd = call_file.fd; + + rc = ioctl(vhost_fd, VHOST_SET_VRING_CALL, &call_file); + if (rc < 0) + die_perror("VHOST_SET_VRING_CALL ioctl on /dev/vhost-net failed"); + + ev = (struct epoll_event){ .data.u64 = ref.u64, .events = EPOLLIN }; + rc = epoll_ctl(c->epollfd, EPOLL_CTL_ADD, ref.fd, &ev); + if (rc < 0) + die_perror("Failed to add call eventfd to epoll"); + c->vq[queue_idx].call_fd = call_file.fd; + + err_file.fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC); + if (err_file.fd < 0) + die_perror("Failed to create error eventfd"); + + rc = ioctl(vhost_fd, VHOST_SET_VRING_ERR, &err_file); + if (rc < 0) + die_perror("VHOST_SET_VRING_ERR ioctl on /dev/vhost-net failed"); + + ref.type = EPOLL_TYPE_VHOST_ERROR; + ref.fd = err_file.fd; + ev.data.u64 = ref.u64; + rc = epoll_ctl(c->epollfd, EPOLL_CTL_ADD, ref.fd, &ev); + if (rc < 0) + die_perror("Failed to add error eventfd to epoll"); + c->vq[queue_idx].err_fd = err_file.fd;

We'd generally prefer to use the existing epoll_add() helper, rather than open coding an EPOLL_CTL_ADD call.

+ + rc = ioctl(vhost_fd, VHOST_SET_VRING_NUM, &state); + if (rc < 0) { + die_perror("VHOST_SET_VRING_NUM ioctl on /dev/vhost-net failed (queue %d)", + queue_idx); + } + + kick_file.fd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC); + if (kick_file.fd < 0) + die_perror("Failed to create kick eventfd"); + + rc = ioctl(vhost_fd, VHOST_SET_VRING_KICK, &kick_file); + if (rc < 0) { + die_perror("VHOST_SET_VRING_KICK ioctl on /dev/vhost-net failed (queue %d)", + queue_idx); + } + + c->vq[queue_idx].kick_fd = kick_file.fd; + + vqs[queue_idx].num_free = VHOST_NDESCS; +} + +/** + * setup_memory_table() - Register the GPA/HVA translation table + * @c: Execution context; c->fd_vhost must already be set + * + * pasta has no real guest, so container->host addresses are 1:1 and can + * be interpreted directly rather than translated.

This is a little unclear, I'd say explicitly that we use a mapping where GPA == HVA.

+ */ +int setup_memory_table(struct ctx *c) { +#define VHOST_MEMORY_REGION_PTR(addr, size) \ + (struct vhost_memory_region) { \ + .guest_phys_addr = (uintptr_t)addr, \ + .memory_size = size, \ + .userspace_addr = (uintptr_t)addr, \ + } +#define VHOST_MEMORY_REGION(elem) VHOST_MEMORY_REGION_PTR(&elem, sizeof(elem))

"elem" is probably not a good name here. Here it's any contiguous variable, but in the vhost-user code, "elem" means a specific data structure within the descriptor rings.

+ + /* general purpose buffers */ + vhost_memory.mem.regions[0] = VHOST_MEMORY_REGION(pkt_buf); + vhost_memory.mem.regions[1] = VHOST_MEMORY_REGION(eth_pad); + + /* tcp specific buffers */ + vhost_memory.mem.regions[2] = VHOST_MEMORY_REGION(tcp_payload_tap_hdr); + vhost_memory.mem.regions[3] = VHOST_MEMORY_REGION(tcp4_payload_ip); + vhost_memory.mem.regions[4] = VHOST_MEMORY_REGION(tcp6_payload_ip); + vhost_memory.mem.regions[5] = VHOST_MEMORY_REGION(tcp_payload); + vhost_memory.mem.regions[6] = VHOST_MEMORY_REGION(tcp_eth_hdr); + + /* udp specific buffers */ + vhost_memory.mem.regions[7] = VHOST_MEMORY_REGION(udp_payload); + vhost_memory.mem.regions[8] = VHOST_MEMORY_REGION(udp_eth_hdr); + vhost_memory.mem.regions[9] = VHOST_MEMORY_REGION(udp_iov_recv); + vhost_memory.mem.regions[10] = VHOST_MEMORY_REGION(udp_mh_recv); + vhost_memory.mem.regions[11] = VHOST_MEMORY_REGION(udp_meta);

Not sure if there would be value in delegating these to helpers in tcp.c and udp.c

+ vhost_memory.mem.nregions = 12; + + return ioctl(c->fd_vhost, VHOST_SET_MEM_TABLE, &vhost_memory.mem); +} + + + +/** + * set_vring_for_queue() - Register a vring's addresses and bind its backend + * @c: Execution context; c->fd_vhost must already be set + * @queue_idx: Index of the queue (vring) to configure + * @tap_fd: Tap fd to bind as this queue's backend + * + */ +void set_vring_for_queue(struct ctx *c, int queue_idx, int tap_fd) +{ + int vhost_fd = c->fd_vhost; + struct vhost_vring_addr addr = { + .index = queue_idx, + .desc_user_addr = (unsigned long)vring_desc[queue_idx], + .avail_user_addr = (unsigned long)&vring_avail_all[queue_idx], + .used_user_addr = (unsigned long)&vring_used_all[queue_idx], + .log_guest_addr = (unsigned long)&vring_used_all[queue_idx], + }; + struct vhost_vring_file file = { + .index = queue_idx, + .fd = tap_fd, + }; + unsigned int i; + int rc; + + debug("qid: %d", queue_idx); + debug("vhost desc addr: 0x%llx", addr.desc_user_addr); + debug("vhost avail addr: 0x%llx", addr.avail_user_addr); + debug("vhost used addr: 0x%llx", addr.used_user_addr); + + rc = ioctl(vhost_fd, VHOST_SET_VRING_ADDR, &addr); + if (rc < 0) + die_perror("VHOST_SET_VRING_ADDR ioctl on /dev/vhost-net failed"); + + if (queue_idx == 0) { + for (i = 0; i < VHOST_NDESCS; ++i) { + vring_desc[0][i].addr = (uintptr_t)pkt_buf + + i * (PKT_BUF_BYTES / VHOST_NDESCS); + vring_desc[0][i].len = PKT_BUF_BYTES / VHOST_NDESCS; + vring_desc[0][i].flags = VRING_DESC_F_WRITE; + } + + for (i = 0; i < VHOST_NDESCS; ++i) + vring_avail_all[0].avail.ring[i] = htole16(i); + + rx_descriptor_handoff(c); + } + + if (queue_idx == 1) { + for (i = 0; i < (VHOST_NDESCS - 1); ++i) { + vring_desc[1][i].next = i+1; + } + } + + debug("qid: %d", file.index); + debug("tap fd: %d", file.fd); + rc = ioctl(vhost_fd, VHOST_NET_SET_BACKEND, &file); + if (rc < 0) + die_perror("VHOST_NET_SET_BACKEND ioctl on /dev/vhost-net failed"); +} + +/** + * rx_descriptor_handoff() - Batch-announce freed RX descriptors to the kernel + * @c: Execution context + * + * Bumps avail.idx by the number of descriptors accumulated in + * vqs[0].num_free (from prior consume_one_rx_descriptor() calls), + * then resets the counter to zero. The kernel will see the new + * avail.idx and consume the freshly-available descriptors. + * + */ +void rx_descriptor_handoff(struct ctx *c)

Can this be static? That's the sort of thing that's harder to review when signatures are split from implementations. If not, it should have a properly prefixed name.

+{ + smp_wmb(); + + if (!vqs[0].num_free) + return; + + vring_avail_all[0].avail.idx += vqs[0].num_free; + vqs[0].num_free = 0; + vhost_kick(&vring_used_all[0].used, c->vq[0].kick_fd); +} + + +/** + * vhost_kick() - Notify the kernel that new descriptors are available + * @used: The vring_used queue. Taken as a parameter to check if the kernel + * virtio thread is actively reading descriptors or no + * @kick_fd: Which fd to notify about (will differ by which queue we are about + * to announce availability in) + */ +void vhost_kick(struct vring_used *used, int kick_fd) {

Again, does this need to be global?

+ /* Ensure that the read to used->flags doesn't get reordered to be + * above the avail.idx update + */ + smp_mb(); + + if (!(used->flags & VRING_USED_F_NO_NOTIFY)) + eventfd_write(kick_fd, 1); +} \ No newline at end of file ^ What diff said.

-- David Gibson (he or they) | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you, not the other way | around. http://www.ozlabs.org/~dgibson