Re: [PATCH v4 5/5] udp: Pass iov_tail to udp_update_hdr4()/udp_update_hdr6()

On Mon, Mar 23, 2026 at 03:31:51PM +0100, Laurent Vivier wrote:

Change udp_update_hdr4() and udp_update_hdr6() to take an iov_tail covering the full L3 frame (IP header + UDP header + data), instead of separate IP header, udp_payload_t, and data-length parameters. The functions now use with_header() and IOV_DROP_HEADER() to access the IP and UDP headers directly from the iov_tail, and derive sizes via iov_tail_size() rather than an explicit length argument.

This decouples the header update functions from the udp_payload_t memory layout, which assumes all headers and data sit in a single contiguous buffer. The vhost-user path uses virtqueue-provided scatter-gather buffers where this assumption does not hold; passing an iov_tail lets both the tap path and the vhost-user path share the same functions without layout-specific helpers.

On the vhost-user side, udp_vu_prepare() likewise switches to with_header() for the Ethernet header, and its caller now drops the vnet header before calling udp_vu_prepare() instead of having the function deal with it internally.

Signed-off-by: Laurent Vivier --- iov.c | 1 - udp.c | 129 +++++++++++++++++++++++++++---------------------- udp_internal.h | 6 +-- udp_vu.c | 51 +++++++++---------- 4 files changed, 97 insertions(+), 90 deletions(-)

diff --git a/iov.c b/iov.c index 7fc9c3c78a32..c1eda9941f32 100644 --- a/iov.c +++ b/iov.c @@ -316,7 +316,6 @@ void *iov_peek_header_(struct iov_tail *tail, void *v, size_t len, size_t align) * * Return: number of bytes written */ -/* cppcheck-suppress unusedFunction */ size_t iov_put_header_(const struct iov_tail *tail, const void *v, size_t len) { size_t l = len; diff --git a/udp.c b/udp.c index 1fc5a42c5ca7..261f2e1b156c 100644 --- a/udp.c +++ b/udp.c @@ -254,42 +254,42 @@ static void udp_iov_init(const struct ctx *c)

/** * udp_update_hdr4() - Update headers for one IPv4 datagram - * @ip4h: Pre-filled IPv4 header (except for tot_len and saddr) - * @bp: Pointer to udp_payload_t to update

I like removing @bp from this function. I'm less enthusiastic about removing @ip4h as a separate parameter. I kind of like the idea of having this take @ip4h, @uh and @payload as separate parameters - a bit like csum_udp4(). I realise however, that that makes the caller awkward, at least with the current constraints of PUSH_HEADER / with_header().

+ * @payload: UDP payload

This is no longer accurate - this needs to include the IP and UDP headers, and so "payload" is not a good name any more.

* @toside: Flowside for destination side - * @dlen: Length of UDP payload * @no_udp_csum: Do not set UDP checksum * * Return: size of IPv4 payload (UDP header + data) */ -size_t udp_update_hdr4(struct iphdr *ip4h, struct udp_payload_t *bp, - const struct flowside *toside, size_t dlen, +size_t udp_update_hdr4(struct iov_tail *payload, const struct flowside *toside, bool no_udp_csum) { const struct in_addr *src = inany_v4(&toside->oaddr); const struct in_addr *dst = inany_v4(&toside->eaddr); - size_t l4len = dlen + sizeof(bp->uh); - size_t l3len = l4len + sizeof(*ip4h); + size_t l3len = iov_tail_size(payload); + size_t l4len = l3len - sizeof(struct iphdr);

assert(src && dst);

- ip4h->tot_len = htons(l3len); - ip4h->daddr = dst->s_addr; - ip4h->saddr = src->s_addr; - ip4h->check = csum_ip4_header(l3len, IPPROTO_UDP, *src, *dst); + with_header(struct iphdr, ip4h, payload) { + ip4h->tot_len = htons(l3len); + ip4h->daddr = dst->s_addr; + ip4h->saddr = src->s_addr; + ip4h->check = csum_ip4_header(l3len, IPPROTO_UDP, *src, *dst); + } + IOV_DROP_HEADER(payload, struct iphdr); + + with_header(struct udphdr, uh, payload) { + uh->source = htons(toside->oport); + uh->dest = htons(toside->eport); + uh->len = htons(l4len); + if (no_udp_csum) { + uh->check = 0; + } else { + struct iov_tail data = *payload;

- bp->uh.source = htons(toside->oport); - bp->uh.dest = htons(toside->eport); - bp->uh.len = htons(l4len); - if (no_udp_csum) { - bp->uh.check = 0; - } else { - const struct iovec iov = { - .iov_base = bp->data, - .iov_len = dlen - }; - struct iov_tail data = IOV_TAIL(&iov, 1, 0); - csum_udp4(&bp->uh, *src, *dst, &data); + IOV_DROP_HEADER(&data, struct udphdr); + csum_udp4(uh, *src, *dst, &data); + } }

return l4len; @@ -297,44 +297,45 @@ size_t udp_update_hdr4(struct iphdr *ip4h, struct udp_payload_t *bp,

/** * udp_update_hdr6() - Update headers for one IPv6 datagram - * @ip6h: Pre-filled IPv6 header (except for payload_len and - * addresses) - * @bp: Pointer to udp_payload_t to update + * @payload: UDP payload * @toside: Flowside for destination side - * @dlen: Length of UDP payload * @no_udp_csum: Do not set UDP checksum * * Return: size of IPv6 payload (UDP header + data) */ -size_t udp_update_hdr6(struct ipv6hdr *ip6h, struct udp_payload_t *bp, - const struct flowside *toside, size_t dlen, +size_t udp_update_hdr6(struct iov_tail *payload, const struct flowside *toside, bool no_udp_csum) { - uint16_t l4len = dlen + sizeof(bp->uh); - - ip6h->payload_len = htons(l4len); - ip6h->daddr = toside->eaddr.a6; - ip6h->saddr = toside->oaddr.a6; - ip6h->version = 6; - ip6h->nexthdr = IPPROTO_UDP; - ip6h->hop_limit = 255; - - bp->uh.source = htons(toside->oport); - bp->uh.dest = htons(toside->eport); - bp->uh.len = ip6h->payload_len; - if (no_udp_csum) { - /* 0 is an invalid checksum for UDP IPv6 and dropped by - * the kernel stack, even if the checksum is disabled by virtio - * flags. We need to put any non-zero value here. - */ - bp->uh.check = 0xffff; - } else { - const struct iovec iov = { - .iov_base = bp->data, - .iov_len = dlen - }; - struct iov_tail data = IOV_TAIL(&iov, 1, 0); - csum_udp6(&bp->uh, &toside->oaddr.a6, &toside->eaddr.a6, &data); + uint16_t l4len = iov_tail_size(payload) - sizeof(struct ipv6hdr); + + with_header(struct ipv6hdr, ip6h, payload) { + ip6h->payload_len = htons(l4len); + ip6h->daddr = toside->eaddr.a6; + ip6h->saddr = toside->oaddr.a6; + ip6h->version = 6; + ip6h->nexthdr = IPPROTO_UDP; + ip6h->hop_limit = 255; + } + IOV_DROP_HEADER(payload, struct ipv6hdr); + + with_header(struct udphdr, uh, payload) { + uh->source = htons(toside->oport); + uh->dest = htons(toside->eport); + uh->len = htons(l4len); + if (no_udp_csum) { + /* 0 is an invalid checksum for UDP IPv6 and dropped by + * the kernel stack, even if the checksum is disabled + * by virtio flags. We need to put any non-zero value + * here. + */ + uh->check = 0xffff; + } else { + struct iov_tail data = *payload; + + IOV_DROP_HEADER(&data, struct udphdr); + csum_udp6(uh, &toside->oaddr.a6, &toside->eaddr.a6, + &data); + } }

return l4len; @@ -374,12 +375,22 @@ static void udp_tap_prepare(const struct mmsghdr *mmh, struct ethhdr *eh = (*tap_iov)[UDP_IOV_ETH].iov_base; struct udp_payload_t *bp = &udp_payload[idx]; struct udp_meta_t *bm = &udp_meta[idx]; + struct iovec iov[3]; + struct iov_tail payload = IOV_TAIL(iov, ARRAY_SIZE(iov), 0); size_t l4len, l2len;

+ iov[1].iov_base = &bp->uh; + iov[1].iov_len = sizeof(bp->uh); + iov[2].iov_base = bp->data; + iov[2].iov_len = mmh[idx].msg_len; + eth_update_mac(eh, NULL, tap_omac); if (!inany_v4(&toside->eaddr) || !inany_v4(&toside->oaddr)) { - l4len = udp_update_hdr6(&bm->ip6h, bp, toside, - mmh[idx].msg_len, no_udp_csum); + + iov[0].iov_base = &bm->ip6h; + iov[0].iov_len = sizeof(bm->ip6h); + + l4len = udp_update_hdr6(&payload, toside, no_udp_csum);

l2len = MAX(l4len + sizeof(bm->ip6h) + ETH_HLEN, ETH_ZLEN); tap_hdr_update(&bm->taph, l2len); @@ -387,8 +398,10 @@ static void udp_tap_prepare(const struct mmsghdr *mmh, eh->h_proto = htons_constant(ETH_P_IPV6); (*tap_iov)[UDP_IOV_IP] = IOV_OF_LVALUE(bm->ip6h); } else { - l4len = udp_update_hdr4(&bm->ip4h, bp, toside, - mmh[idx].msg_len, no_udp_csum); + iov[0].iov_base = &bm->ip4h; + iov[0].iov_len = sizeof(bm->ip4h); + + l4len = udp_update_hdr4(&payload, toside, no_udp_csum);

l2len = MAX(l4len + sizeof(bm->ip4h) + ETH_HLEN, ETH_ZLEN); tap_hdr_update(&bm->taph, l2len); diff --git a/udp_internal.h b/udp_internal.h index 64e457748324..fb3017ae3251 100644 --- a/udp_internal.h +++ b/udp_internal.h @@ -25,11 +25,9 @@ struct udp_payload_t { } __attribute__ ((packed, aligned(__alignof__(unsigned int)))); #endif

-size_t udp_update_hdr4(struct iphdr *ip4h, struct udp_payload_t *bp, - const struct flowside *toside, size_t dlen, +size_t udp_update_hdr4(struct iov_tail *payload, const struct flowside *toside, bool no_udp_csum); -size_t udp_update_hdr6(struct ipv6hdr *ip6h, struct udp_payload_t *bp, - const struct flowside *toside, size_t dlen, +size_t udp_update_hdr6(struct iov_tail *payload, const struct flowside *toside, bool no_udp_csum); void udp_sock_fwd(const struct ctx *c, int s, int rule_hint, uint8_t frompif, in_port_t port, const struct timespec *now); diff --git a/udp_vu.c b/udp_vu.c index dc949e30a793..80391b4f8788 100644 --- a/udp_vu.c +++ b/udp_vu.c @@ -93,42 +93,39 @@ static ssize_t udp_vu_sock_recv(struct iovec *iov, size_t *cnt, int s, bool v6) * @c: Execution context * @data: IO vector tail for the frame * @toside: Address information for one side of the flow - * @dlen: Packet data length * * Return: Layer-4 length */ static size_t udp_vu_prepare(const struct ctx *c, const struct iov_tail *data, - const struct flowside *toside, ssize_t dlen) + const struct flowside *toside) { - const struct iovec *iov = data->iov; - struct ethhdr *eh; + bool ipv4 = inany_v4(&toside->eaddr) && inany_v4(&toside->oaddr); + struct iov_tail payload = *data; size_t l4len;

/* ethernet header */ - eh = vu_eth(iov[0].iov_base); - - memcpy(eh->h_dest, c->guest_mac, sizeof(eh->h_dest)); - memcpy(eh->h_source, c->our_tap_mac, sizeof(eh->h_source)); + with_header(struct ethhdr, eh, &payload) { + memcpy(eh->h_dest, c->guest_mac, sizeof(eh->h_dest)); + memcpy(eh->h_source, c->our_tap_mac, sizeof(eh->h_source)); + + if (ipv4) + eh->h_proto = htons(ETH_P_IP); + else + eh->h_proto = htons(ETH_P_IPV6); + } + IOV_DROP_HEADER(&payload, struct ethhdr);

/* initialize header */ - if (inany_v4(&toside->eaddr) && inany_v4(&toside->oaddr)) { - struct iphdr *iph = vu_ip(iov[0].iov_base); - struct udp_payload_t *bp = vu_payloadv4(iov[0].iov_base); + if (ipv4) { + with_header(struct iphdr, iph, &payload) + *iph = (struct iphdr)L2_BUF_IP4_INIT(IPPROTO_UDP);

- eh->h_proto = htons(ETH_P_IP); - - *iph = (struct iphdr)L2_BUF_IP4_INIT(IPPROTO_UDP); - - l4len = udp_update_hdr4(iph, bp, toside, dlen, true); + l4len = udp_update_hdr4(&payload, toside, true); } else { - struct ipv6hdr *ip6h = vu_ip(iov[0].iov_base); - struct udp_payload_t *bp = vu_payloadv6(iov[0].iov_base); - - eh->h_proto = htons(ETH_P_IPV6); - - *ip6h = (struct ipv6hdr)L2_BUF_IP6_INIT(IPPROTO_UDP); + with_header(struct ipv6hdr, ip6h, &payload) + *ip6h = (struct ipv6hdr)L2_BUF_IP6_INIT(IPPROTO_UDP);

- l4len = udp_update_hdr6(ip6h, bp, toside, dlen, true); + l4len = udp_update_hdr6(&payload, toside, true); }

return l4len; @@ -137,7 +134,7 @@ static size_t udp_vu_prepare(const struct ctx *c, const struct iov_tail *data, /** * udp_vu_csum() - Calculate and set checksum for a UDP packet * @toside: Address information for one side of the flow - * @data: IO vector tail for the frame (including vnet header) + * @data: IO vector tail for the L2 frame */ static void udp_vu_csum(const struct flowside *toside, const struct iov_tail *data) @@ -148,7 +145,6 @@ static void udp_vu_csum(const struct flowside *toside, struct udphdr *uh, uh_storage; bool ipv4 = src4 && dst4;

- IOV_DROP_HEADER(&payload, struct virtio_net_hdr_mrg_rxbuf); IOV_DROP_HEADER(&payload, struct ethhdr); if (ipv4) IOV_DROP_HEADER(&payload, struct iphdr); @@ -225,10 +221,11 @@ void udp_vu_sock_to_tap(const struct ctx *c, int s, int n, flow_sidx_t tosidx) if (iov_cnt > 0) { struct iov_tail data = IOV_TAIL(iov_vu, iov_cnt, 0); vu_set_vnethdr(iov_vu[0].iov_base, elem_used); - udp_vu_prepare(c, &data, toside, dlen); + iov_drop_header(&data, VNET_HLEN); + udp_vu_prepare(c, &data, toside); if (*c->pcap) { udp_vu_csum(toside, &data); - pcap_iov(data.iov, data.cnt, VNET_HLEN); + pcap_iov(data.iov, data.cnt, data.off); } vu_flush(vdev, vq, elem, elem_used); } -- 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