public inbox for passt-dev@passt.top
 help / color / mirror / code / Atom feed
From: Laurent Vivier <lvivier@redhat.com>
To: passt-dev@passt.top
Cc: Laurent Vivier <lvivier@redhat.com>
Subject: [PATCH 7/7] tcp_vu: Use iov_tail helpers to build headers in tcp_vu_send_flag()
Date: Mon, 23 Mar 2026 17:52:59 +0100	[thread overview]
Message-ID: <20260323165259.1253482-8-lvivier@redhat.com> (raw)
In-Reply-To: <20260323165259.1253482-1-lvivier@redhat.com>

tcp_vu_send_flag() previously used vu_eth(), vu_ip(), vu_payloadv4() and
vu_payloadv6() to access headers via direct pointer arithmetic on a
single contiguous buffer.  Replace these with iov_tail-based
with_header() and IOV_DROP_HEADER() accessors, which work correctly with
multi-buffer virtio descriptors.

Enlarge the flags_iov and flags_elem arrays to accommodate split
descriptors, and pass iov_cnt through vu_collect() so we track
how many iovecs the frame actually spans.

Extract a tcp_vu_send_dup() helper for the DUP_ACK path that uses
iov_memcopy() to duplicate a frame across potentially multi-buffer
descriptors, replacing the previous single-buffer memcpy.

With all callers converted, remove the vu_eth(), vu_ip(),
vu_payloadv4() and vu_payloadv6() inline helpers from vu_common.h,
and drop the cppcheck-suppress annotation on iov_memcopy() which is
now used.

Signed-off-by: Laurent Vivier <lvivier@redhat.com>
---
 iov.c       |   1 -
 tcp_vu.c    | 113 ++++++++++++++++++++++++++++------------------------
 vu_common.h |  20 ----------
 3 files changed, 62 insertions(+), 72 deletions(-)

diff --git a/iov.c b/iov.c
index 55263ec10b44..9638dc6de03d 100644
--- a/iov.c
+++ b/iov.c
@@ -207,7 +207,6 @@ void iov_memset(const struct iovec *iov, size_t iov_cnt, size_t offset, int c,
  *
  * Return: total number of bytes copied
  */
-/* cppcheck-suppress unusedFunction */
 size_t iov_memcopy(struct iovec *dst_iov, size_t dst_iov_cnt, size_t dst_offs,
 		   const struct iovec *iov, size_t iov_cnt, size_t offs)
 {
diff --git a/tcp_vu.c b/tcp_vu.c
index d6bd6a34d6da..776b47aea18c 100644
--- a/tcp_vu.c
+++ b/tcp_vu.c
@@ -74,6 +74,41 @@ static size_t tcp_vu_hdrlen(bool v6)
 	return hdrlen;
 }
 
+/**
+ * tcp_vu_send_dup() - Duplicate a frame into a new virtqueue element
+ * @c:		Execution context
+ * @vq:		Receive virtqueue
+ * @dest_elem:	Destination virtqueue element to collect
+ * @dest_iov:	Destination iovec array for collected buffers
+ * @max_dest_iov: Maximum number of entries in @dest_iov
+ * @src_iov:	Source iovec array containing the frame to duplicate
+ * @src_cnt:	Number of entries in @src_iov
+ *
+ * Return: number of virtqueue elements collected (0 if none available)
+ */
+static int tcp_vu_send_dup(const struct ctx *c, struct vu_virtq *vq,
+			   struct vu_virtq_element *dest_elem,
+			   struct iovec *dest_iov, size_t max_dest_iov,
+			   const struct iovec *src_iov, size_t src_cnt)
+{
+	size_t dest_cnt, frame_size = iov_size(src_iov, src_cnt);
+	const struct vu_dev *vdev = c->vdev;
+	int elem_cnt;
+
+	elem_cnt = vu_collect(vdev, vq, dest_elem, 1, dest_iov, max_dest_iov,
+			      &dest_cnt, frame_size, NULL);
+	if (elem_cnt == 0)
+		return 0;
+
+	vu_pad(dest_iov, dest_cnt, 0, frame_size);
+	iov_memcopy(dest_iov, dest_cnt, 0, src_iov, src_cnt, 0);
+
+	if (*c->pcap)
+		pcap_iov(dest_iov, dest_cnt, VNET_HLEN);
+
+	return elem_cnt;
+}
+
 /**
  * tcp_vu_send_flag() - Send segment with flags to vhost-user (no payload)
  * @c:		Execution context
@@ -87,13 +122,9 @@ int tcp_vu_send_flag(const struct ctx *c, struct tcp_tap_conn *conn, int flags)
 	struct vu_dev *vdev = c->vdev;
 	struct vu_virtq *vq = &vdev->vq[VHOST_USER_RX_QUEUE];
 	struct vu_virtq_element flags_elem[2];
-	struct ipv6hdr *ip6h = NULL;
-	struct iphdr *ip4h = NULL;
-	struct iovec flags_iov[2];
+	size_t optlen, hdrlen, iov_cnt;
+	struct iovec flags_iov[64];
 	struct iov_tail payload;
-	size_t optlen, hdrlen;
-	struct tcphdr *th;
-	struct ethhdr *eh;
 	uint32_t seq;
 	int elem_cnt;
 	int nb_ack;
@@ -102,77 +133,57 @@ int tcp_vu_send_flag(const struct ctx *c, struct tcp_tap_conn *conn, int flags)
 	hdrlen = tcp_vu_hdrlen(CONN_V6(conn));
 
 	elem_cnt = vu_collect(vdev, vq, &flags_elem[0], 1,
-			      &flags_iov[0], 1, NULL,
+			      flags_iov, ARRAY_SIZE(flags_iov), &iov_cnt,
 			      hdrlen + sizeof(struct tcp_syn_opts), NULL);
-	if (elem_cnt != 1)
+	if (elem_cnt == 0)
 		return -1;
 
-	assert(flags_elem[0].in_num == 1);
-	assert(flags_elem[0].in_sg[0].iov_len >=
-	       MAX(hdrlen + sizeof(struct tcp_syn_opts), ETH_ZLEN + VNET_HLEN));
-
-	vu_set_vnethdr(flags_elem[0].in_sg[0].iov_base, 1);
+	vu_set_vnethdr(flags_elem[0].in_sg[0].iov_base, elem_cnt);
 
-	eh = vu_eth(flags_elem[0].in_sg[0].iov_base);
+	payload = IOV_TAIL(flags_elem[0].in_sg, iov_cnt, VNET_HLEN);
 
-	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));
+	IOV_DROP_HEADER(&payload, struct ethhdr);
 
 	if (CONN_V4(conn)) {
-		eh->h_proto = htons(ETH_P_IP);
-
-		ip4h = vu_ip(flags_elem[0].in_sg[0].iov_base);
-		*ip4h = (struct iphdr)L2_BUF_IP4_INIT(IPPROTO_TCP);
-
-		th = vu_payloadv4(flags_elem[0].in_sg[0].iov_base);
+		with_header(struct iphdr, ip4h, &payload)
+			*ip4h = (struct iphdr)L2_BUF_IP4_INIT(IPPROTO_TCP);
+		IOV_DROP_HEADER(&payload, struct iphdr);
 	} else {
-		eh->h_proto = htons(ETH_P_IPV6);
-
-		ip6h = vu_ip(flags_elem[0].in_sg[0].iov_base);
-		*ip6h = (struct ipv6hdr)L2_BUF_IP6_INIT(IPPROTO_TCP);
-		th = vu_payloadv6(flags_elem[0].in_sg[0].iov_base);
+		with_header(struct ipv6hdr, ip6h, &payload)
+			*ip6h = (struct ipv6hdr)L2_BUF_IP6_INIT(IPPROTO_TCP);
+		IOV_DROP_HEADER(&payload, struct ipv6hdr);
 	}
 
-	memset(th, 0, sizeof(*th));
-	th->doff = sizeof(*th) / 4;
-	th->ack = 1;
+	with_header(struct tcphdr, th, &payload)
+		memset(th, 0, sizeof(*th));
 
 	seq = conn->seq_to_tap;
-	payload = IOV_TAIL(flags_elem[0].in_sg, 1, hdrlen - sizeof(*th));
 	ret = tcp_prepare_flags(c, conn, flags, &payload, &optlen);
 	if (ret <= 0) {
-		vu_queue_rewind(vq, 1);
+		vu_queue_rewind(vq, elem_cnt);
 		return ret;
 	}
 
-	vu_pad(&flags_iov[0], 1, 0, hdrlen + optlen);
+	vu_pad(flags_elem[0].in_sg, iov_cnt, 0, hdrlen + optlen);
 
 	if (flags & KEEPALIVE)
 		seq--;
 
-	payload = IOV_TAIL(flags_elem[0].in_sg, 1, VNET_HLEN);
+	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);
 
 	if (*c->pcap)
-		pcap_iov(&flags_elem[0].in_sg[0], 1, VNET_HLEN);
-	nb_ack = 1;
+		pcap_iov(flags_elem[0].in_sg, iov_cnt, VNET_HLEN);
+	nb_ack = elem_cnt;
 
 	if (flags & DUP_ACK) {
-		elem_cnt = vu_collect(vdev, vq, &flags_elem[1], 1,
-				      &flags_iov[1], 1, NULL,
-				      hdrlen + optlen, NULL);
-		if (elem_cnt == 1 &&
-		    flags_elem[1].in_sg[0].iov_len >=
-		    flags_elem[0].in_sg[0].iov_len) {
-			vu_pad(&flags_iov[1], 1, 0, hdrlen + optlen);
-			memcpy(flags_elem[1].in_sg[0].iov_base,
-			       flags_elem[0].in_sg[0].iov_base,
-			       flags_elem[0].in_sg[0].iov_len);
-			nb_ack++;
-
-			if (*c->pcap)
-				pcap_iov(&flags_elem[1].in_sg[0], 1, VNET_HLEN);
-		}
+		elem_cnt = tcp_vu_send_dup(c, vq, &flags_elem[elem_cnt],
+					   &flags_iov[iov_cnt],
+					   ARRAY_SIZE(flags_iov) - iov_cnt,
+					   flags_elem[0].in_sg, iov_cnt);
+		nb_ack += elem_cnt;
 	}
 
 	vu_flush(vdev, vq, flags_elem, nb_ack);
diff --git a/vu_common.h b/vu_common.h
index f09dd25d3d1f..8e42ce88b2ab 100644
--- a/vu_common.h
+++ b/vu_common.h
@@ -15,26 +15,6 @@
 #include "ip.h"
 #include "virtio.h"
 
-static inline void *vu_eth(void *base)
-{
-	return ((char *)base + VNET_HLEN);
-}
-
-static inline void *vu_ip(void *base)
-{
-	return (struct ethhdr *)vu_eth(base) + 1;
-}
-
-static inline void *vu_payloadv4(void *base)
-{
-	return (struct iphdr *)vu_ip(base) + 1;
-}
-
-static inline void *vu_payloadv6(void *base)
-{
-	return (struct ipv6hdr *)vu_ip(base) + 1;
-}
-
 int vu_collect(const struct vu_dev *vdev, struct vu_virtq *vq,
 	       struct vu_virtq_element *elem, int max_elem,
 	       struct iovec *in_sg, size_t max_in_sg, size_t *in_total,
-- 
2.53.0


  parent reply	other threads:[~2026-03-23 16:53 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-03-23 16:52 [PATCH 0/7] vhost-user,tcp: Handle multiple iovec entries per virtqueue element Laurent Vivier
2026-03-23 16:52 ` [PATCH 1/7] tcp: pass ipv4h checksum, not a pointer to the checksum Laurent Vivier
2026-03-24  3:53   ` David Gibson
2026-03-24  7:56     ` Laurent Vivier
2026-03-24 23:49       ` David Gibson
2026-03-23 16:52 ` [PATCH 2/7] tcp: use iov_tail to access headers in tcp_fill_headers() Laurent Vivier
2026-03-24  3:58   ` David Gibson
2026-03-23 16:52 ` [PATCH 3/7] tcp_vu: Use iov_tail helpers to build headers in tcp_vu_prepare() Laurent Vivier
2026-03-25  4:46   ` David Gibson
2026-03-23 16:52 ` [PATCH 4/7] tcp_vu: Support multibuffer frames in tcp_vu_sock_recv() Laurent Vivier
2026-03-25  5:06   ` David Gibson
2026-03-23 16:52 ` [PATCH 5/7] tcp: Use iov_tail to access headers in tcp_prepare_flags() Laurent Vivier
2026-03-23 16:52 ` [PATCH 6/7] iov: introduce iov_memcopy() Laurent Vivier
2026-03-23 16:52 ` Laurent Vivier [this message]
2026-03-25  5:07 ` [PATCH 0/7] vhost-user,tcp: Handle multiple iovec entries per virtqueue element David Gibson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260323165259.1253482-8-lvivier@redhat.com \
    --to=lvivier@redhat.com \
    --cc=passt-dev@passt.top \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://passt.top/passt

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for IMAP folder(s).