On Mon, Mar 23, 2026 at 07:01:49PM +0100, Laurent Vivier wrote:

According to the virtio-net specification, when the VIRTIO_NET_F_GUEST_CSUM is negotiated, the device can set VIRTIO_NET_HDR_F_DATA_VALID in the virtio-net header to indicate that packet checksums have been validated, allowing the guest to skip verification. Without this feature, the device must provide fully checksummed packets.

The vhost-user TCP and UDP paths were unconditionally skipping checksum computation, regardless of whether GUEST_CSUM was negotiated. This went undetected with Linux guests because Linux's virtio-net driver honours VIRTIO_NET_HDR_F_DATA_VALID regardless of whether VIRTIO_NET_F_GUEST_CSUM was negotiated, marking such packets as CHECKSUM_UNNECESSARY and skipping verification.

iPXE, however, does not negotiate GUEST_CSUM, ignores the DATA_VALID flag entirely, and always verifies checksums. This caused TCP connections to fail: the SYN-ACK had a zero TCP checksum, iPXE rejected it, and the connection timed out in SYN_RCVD.

Adding --pcap happened to mask the bug, because the pcap code path forces checksum computation to ensure correct captures.

Offer VIRTIO_NET_F_GUEST_CSUM in the device features, and only skip checksum computation when the guest has actually negotiated it. When GUEST_CSUM is not negotiated, always compute valid checksums as required by the specification.

We keep setting VIRTIO_NET_HDR_F_DATA_VALID unconditionally in VU_HEADER: when GUEST_CSUM is negotiated, the flag lets the guest skip checksum verification; when it is not, the spec says the guest should ignore the flags field, so setting it is harmless.

Signed-off-by: Laurent Vivier

Reviewed-by: David Gibson

---

Notes: Based-on: 20260323165259.1253482-1-lvivier@redhat.com

tcp_vu.c | 8 ++++++-- udp_vu.c | 6 ++++-- vhost_user.c | 1 + 3 files changed, 11 insertions(+), 4 deletions(-)

diff --git a/tcp_vu.c b/tcp_vu.c index 776b47aea18c..2ab56f25a9be 100644 --- a/tcp_vu.c +++ b/tcp_vu.c @@ -172,7 +172,9 @@ int tcp_vu_send_flag(const struct ctx *c, struct tcp_tap_conn *conn, int flags) seq--;

payload = IOV_TAIL(flags_elem[0].in_sg, iov_cnt, VNET_HLEN); - tcp_fill_headers(c, conn, CONN_V4(conn), &payload, -1, seq, !*c->pcap); + tcp_fill_headers(c, conn, CONN_V4(conn), &payload, -1, seq, + vu_has_feature(vdev, VIRTIO_NET_F_GUEST_CSUM) && + !*c->pcap);

if (*c->pcap) pcap_iov(flags_elem[0].in_sg, iov_cnt, VNET_HLEN); @@ -512,7 +514,9 @@ int tcp_vu_data_from_sock(const struct ctx *c, struct tcp_tap_conn *conn) check = -1; previous_dlen = dlen;

- tcp_vu_prepare(c, conn, iov, iov_cnt, &check, !*c->pcap, push); + tcp_vu_prepare(c, conn, iov, iov_cnt, &check, + vu_has_feature(vdev, VIRTIO_NET_F_GUEST_CSUM) && + !*c->pcap, push);

if (*c->pcap) pcap_iov(iov, iov_cnt, VNET_HLEN); diff --git a/udp_vu.c b/udp_vu.c index 80391b4f8788..7ed271403481 100644 --- a/udp_vu.c +++ b/udp_vu.c @@ -223,10 +223,12 @@ void udp_vu_sock_to_tap(const struct ctx *c, int s, int n, flow_sidx_t tosidx) vu_set_vnethdr(iov_vu[0].iov_base, elem_used); iov_drop_header(&data, VNET_HLEN); udp_vu_prepare(c, &data, toside); - if (*c->pcap) { + if (!vu_has_feature(vdev, VIRTIO_NET_F_GUEST_CSUM) || + *c->pcap) { udp_vu_csum(toside, &data); - pcap_iov(data.iov, data.cnt, data.off); } + if (*c->pcap) + pcap_iov(data.iov, data.cnt, data.off); vu_flush(vdev, vq, elem, elem_used); } } diff --git a/vhost_user.c b/vhost_user.c index 75665ec6522f..08a7b2d74099 100644 --- a/vhost_user.c +++ b/vhost_user.c @@ -322,6 +322,7 @@ static bool vu_get_features_exec(struct vu_dev *vdev, { uint64_t features = 1ULL << VIRTIO_F_VERSION_1 | + 1ULL << VIRTIO_NET_F_GUEST_CSUM | 1ULL << VIRTIO_NET_F_MRG_RXBUF | 1ULL << VHOST_F_LOG_ALL | 1ULL << VHOST_USER_F_PROTOCOL_FEATURES; -- 2.53.0

-- 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