With vhost-user we can disable the checksum of UDP and TCP.
Add a generic parameter for each of them to disable the
checksum.
v3: align parameters description with tab
s/O/0/
v2: s/UPD/UDP/
Add David's R-b in PATCH 2
Laurent Vivier (2):
udp: Allow checksum to be disabled
tcp: Allow checksum to be disabled
tcp.c | 52 ++++++++++++++++++++++++++------------------
tcp_buf.c | 8 ++++---
tcp_internal.h | 3 ++-
udp.c | 58 ++++++++++++++++++++++++++++++++++----------------
4 files changed, 78 insertions(+), 43 deletions(-)
--
2.46.0
We can need not to set the UDP checksum. Add a parameter to
udp_update_hdr4() and udp_update_hdr6() to disable it.
Signed-off-by: Laurent Vivier <lvivier(a)redhat.com>
---
Notes:
v3: align parameters description with tab
s/O/0/
v2: s/UPD/UDP/
udp.c | 58 ++++++++++++++++++++++++++++++++++++++++------------------
1 file changed, 40 insertions(+), 18 deletions(-)
diff --git a/udp.c b/udp.c
index 2ba00c9c20a8..7b2831386db8 100644
--- a/udp.c
+++ b/udp.c
@@ -294,15 +294,17 @@ static void udp_splice_send(const struct ctx *c, size_t start,
size_t n,
/**
* 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
- * @toside: Flowside for destination side
- * @dlen: Length of UDP payload
+ * @ip4h: Pre-filled IPv4 header (except for tot_len and saddr)
+ * @bp: Pointer to udp_payload_t to update
+ * @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)
*/
static size_t udp_update_hdr4(struct iphdr *ip4h, struct udp_payload_t *bp,
- const struct flowside *toside, size_t dlen)
+ const struct flowside *toside, size_t dlen,
+ bool no_udp_csum)
{
const struct in_addr *src = inany_v4(&toside->oaddr);
const struct in_addr *dst = inany_v4(&toside->eaddr);
@@ -319,22 +321,28 @@ static size_t udp_update_hdr4(struct iphdr *ip4h, struct
udp_payload_t *bp,
bp->uh.source = htons(toside->oport);
bp->uh.dest = htons(toside->eport);
bp->uh.len = htons(l4len);
- csum_udp4(&bp->uh, *src, *dst, bp->data, dlen);
+ if (no_udp_csum)
+ bp->uh.check = 0;
+ else
+ csum_udp4(&bp->uh, *src, *dst, bp->data, dlen);
return l4len;
}
/**
* 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
- * @toside: Flowside for destination side
- * @dlen: Length of UDP payload
+ * @ip6h: Pre-filled IPv6 header (except for payload_len and
+ * addresses)
+ * @bp: Pointer to udp_payload_t to update
+ * @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)
*/
static size_t udp_update_hdr6(struct ipv6hdr *ip6h, struct udp_payload_t *bp,
- const struct flowside *toside, size_t dlen)
+ const struct flowside *toside, size_t dlen,
+ bool no_udp_csum)
{
uint16_t l4len = dlen + sizeof(bp->uh);
@@ -348,7 +356,16 @@ static size_t udp_update_hdr6(struct ipv6hdr *ip6h, struct
udp_payload_t *bp,
bp->uh.source = htons(toside->oport);
bp->uh.dest = htons(toside->eport);
bp->uh.len = ip6h->payload_len;
- csum_udp6(&bp->uh, &toside->oaddr.a6, &toside->eaddr.a6,
bp->data, dlen);
+ 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 {
+ csum_udp6(&bp->uh, &toside->oaddr.a6, &toside->eaddr.a6,
+ bp->data, dlen);
+ }
return l4len;
}
@@ -358,9 +375,11 @@ static size_t udp_update_hdr6(struct ipv6hdr *ip6h, struct
udp_payload_t *bp,
* @mmh: Receiving mmsghdr array
* @idx: Index of the datagram to prepare
* @toside: Flowside for destination side
+ * @no_udp_csum: Do not set UDP checksum
*/
-static void udp_tap_prepare(const struct mmsghdr *mmh, unsigned idx,
- const struct flowside *toside)
+static void udp_tap_prepare(const struct mmsghdr *mmh,
+ unsigned idx, const struct flowside *toside,
+ bool no_udp_csum)
{
struct iovec (*tap_iov)[UDP_NUM_IOVS] = &udp_l2_iov[idx];
struct udp_payload_t *bp = &udp_payload[idx];
@@ -368,13 +387,15 @@ static void udp_tap_prepare(const struct mmsghdr *mmh, unsigned
idx,
size_t l4len;
if (!inany_v4(&toside->eaddr) || !inany_v4(&toside->oaddr)) {
- l4len = udp_update_hdr6(&bm->ip6h, bp, toside, mmh[idx].msg_len);
+ l4len = udp_update_hdr6(&bm->ip6h, bp, toside,
+ mmh[idx].msg_len, no_udp_csum);
tap_hdr_update(&bm->taph, l4len + sizeof(bm->ip6h) +
sizeof(udp6_eth_hdr));
(*tap_iov)[UDP_IOV_ETH] = IOV_OF_LVALUE(udp6_eth_hdr);
(*tap_iov)[UDP_IOV_IP] = IOV_OF_LVALUE(bm->ip6h);
} else {
- l4len = udp_update_hdr4(&bm->ip4h, bp, toside, mmh[idx].msg_len);
+ l4len = udp_update_hdr4(&bm->ip4h, bp, toside,
+ mmh[idx].msg_len, no_udp_csum);
tap_hdr_update(&bm->taph, l4len + sizeof(bm->ip4h) +
sizeof(udp4_eth_hdr));
(*tap_iov)[UDP_IOV_ETH] = IOV_OF_LVALUE(udp4_eth_hdr);
@@ -565,7 +586,8 @@ void udp_listen_sock_handler(const struct ctx *c, union epoll_ref
ref,
udp_splice_prepare(udp_mh_recv, i);
} else if (batchpif == PIF_TAP) {
udp_tap_prepare(udp_mh_recv, i,
- flowside_at_sidx(batchsidx));
+ flowside_at_sidx(batchsidx),
+ false);
}
if (++i >= n)
@@ -636,7 +658,7 @@ void udp_reply_sock_handler(const struct ctx *c, union epoll_ref ref,
if (pif_is_socket(topif))
udp_splice_prepare(udp_mh_recv, i);
else if (topif == PIF_TAP)
- udp_tap_prepare(udp_mh_recv, i, toside);
+ udp_tap_prepare(udp_mh_recv, i, toside, false);
/* Restore sockaddr length clobbered by recvmsg() */
udp_mh_recv[i].msg_hdr.msg_namelen = sizeof(udp_meta[i].s_in);
}
--
2.46.0
We can need not to set TCP checksum. Add a parameter to
tcp_fill_headers4() and tcp_fill_headers6() to disable it.
Signed-off-by: Laurent Vivier <lvivier(a)redhat.com>
Reviewed-by: David Gibson <david(a)gibson.dropbear.id.au>
---
Notes:
v3: align parameters description with tab
tcp.c | 52 ++++++++++++++++++++++++++++++--------------------
tcp_buf.c | 8 +++++---
tcp_internal.h | 3 ++-
3 files changed, 38 insertions(+), 25 deletions(-)
diff --git a/tcp.c b/tcp.c
index f9fe1b9a1330..7307ea74f25c 100644
--- a/tcp.c
+++ b/tcp.c
@@ -896,13 +896,14 @@ static void tcp_fill_header(struct tcphdr *th,
/**
* tcp_fill_headers4() - Fill 802.3, IPv4, TCP headers in pre-cooked buffers
- * @conn: Connection pointer
- * @taph: tap backend specific header
- * @iph: Pointer to IPv4 header
- * @th: Pointer to TCP header
- * @dlen: TCP payload length
- * @check: Checksum, if already known
- * @seq: Sequence number for this segment
+ * @conn: Connection pointer
+ * @taph: tap backend specific header
+ * @iph: Pointer to IPv4 header
+ * @th: Pointer to TCP header
+ * @dlen: TCP payload length
+ * @check: Checksum, if already known
+ * @seq: Sequence number for this segment
+ * @no_tcp_csum: Do not set TCP checksum
*
* Return: The IPv4 payload length, host order
*/
@@ -910,7 +911,7 @@ static size_t tcp_fill_headers4(const struct tcp_tap_conn *conn,
struct tap_hdr *taph,
struct iphdr *iph, struct tcphdr *th,
size_t dlen, const uint16_t *check,
- uint32_t seq)
+ uint32_t seq, bool no_tcp_csum)
{
const struct flowside *tapside = TAPFLOW(conn);
const struct in_addr *src4 = inany_v4(&tapside->oaddr);
@@ -929,7 +930,10 @@ static size_t tcp_fill_headers4(const struct tcp_tap_conn *conn,
tcp_fill_header(th, conn, seq);
- tcp_update_check_tcp4(iph, th);
+ if (no_tcp_csum)
+ th->check = 0;
+ else
+ tcp_update_check_tcp4(iph, th);
tap_hdr_update(taph, l3len + sizeof(struct ethhdr));
@@ -938,20 +942,21 @@ static size_t tcp_fill_headers4(const struct tcp_tap_conn *conn,
/**
* tcp_fill_headers6() - Fill 802.3, IPv6, TCP headers in pre-cooked buffers
- * @conn: Connection pointer
- * @taph: tap backend specific header
- * @ip6h: Pointer to IPv6 header
- * @th: Pointer to TCP header
- * @dlen: TCP payload length
- * @check: Checksum, if already known
- * @seq: Sequence number for this segment
+ * @conn: Connection pointer
+ * @taph: tap backend specific header
+ * @ip6h: Pointer to IPv6 header
+ * @th: Pointer to TCP header
+ * @dlen: TCP payload length
+ * @check: Checksum, if already known
+ * @seq: Sequence number for this segment
+ * @no_tcp_csum: Do not set TCP checksum
*
* Return: The IPv6 payload length, host order
*/
static size_t tcp_fill_headers6(const struct tcp_tap_conn *conn,
struct tap_hdr *taph,
struct ipv6hdr *ip6h, struct tcphdr *th,
- size_t dlen, uint32_t seq)
+ size_t dlen, uint32_t seq, bool no_tcp_csum)
{
const struct flowside *tapside = TAPFLOW(conn);
size_t l4len = dlen + sizeof(*th);
@@ -970,7 +975,10 @@ static size_t tcp_fill_headers6(const struct tcp_tap_conn *conn,
tcp_fill_header(th, conn, seq);
- tcp_update_check_tcp6(ip6h, th);
+ if (no_tcp_csum)
+ th->check = 0;
+ else
+ tcp_update_check_tcp6(ip6h, th);
tap_hdr_update(taph, l4len + sizeof(*ip6h) + sizeof(struct ethhdr));
@@ -984,12 +992,14 @@ static size_t tcp_fill_headers6(const struct tcp_tap_conn *conn,
* @dlen: TCP payload length
* @check: Checksum, if already known
* @seq: Sequence number for this segment
+ * @no_tcp_csum: Do not set TCP checksum
*
* Return: IP payload length, host order
*/
size_t tcp_l2_buf_fill_headers(const struct tcp_tap_conn *conn,
struct iovec *iov, size_t dlen,
- const uint16_t *check, uint32_t seq)
+ const uint16_t *check, uint32_t seq,
+ bool no_tcp_csum)
{
const struct flowside *tapside = TAPFLOW(conn);
const struct in_addr *a4 = inany_v4(&tapside->oaddr);
@@ -998,13 +1008,13 @@ size_t tcp_l2_buf_fill_headers(const struct tcp_tap_conn *conn,
return tcp_fill_headers4(conn, iov[TCP_IOV_TAP].iov_base,
iov[TCP_IOV_IP].iov_base,
iov[TCP_IOV_PAYLOAD].iov_base, dlen,
- check, seq);
+ check, seq, no_tcp_csum);
}
return tcp_fill_headers6(conn, iov[TCP_IOV_TAP].iov_base,
iov[TCP_IOV_IP].iov_base,
iov[TCP_IOV_PAYLOAD].iov_base, dlen,
- seq);
+ seq, no_tcp_csum);
}
/**
diff --git a/tcp_buf.c b/tcp_buf.c
index 1a398461a34b..10a663bdfc26 100644
--- a/tcp_buf.c
+++ b/tcp_buf.c
@@ -320,7 +320,7 @@ int tcp_buf_send_flag(struct ctx *c, struct tcp_tap_conn *conn, int
flags)
return ret;
}
- l4len = tcp_l2_buf_fill_headers(conn, iov, optlen, NULL, seq);
+ l4len = tcp_l2_buf_fill_headers(conn, iov, optlen, NULL, seq, false);
iov[TCP_IOV_PAYLOAD].iov_len = l4len;
if (flags & DUP_ACK) {
@@ -381,7 +381,8 @@ static void tcp_data_to_tap(struct ctx *c, struct tcp_tap_conn *conn,
tcp4_frame_conns[tcp4_payload_used] = conn;
iov = tcp4_l2_iov[tcp4_payload_used++];
- l4len = tcp_l2_buf_fill_headers(conn, iov, dlen, check, seq);
+ l4len = tcp_l2_buf_fill_headers(conn, iov, dlen, check, seq,
+ false);
iov[TCP_IOV_PAYLOAD].iov_len = l4len;
if (tcp4_payload_used > TCP_FRAMES_MEM - 1)
tcp_payload_flush(c);
@@ -389,7 +390,8 @@ static void tcp_data_to_tap(struct ctx *c, struct tcp_tap_conn *conn,
tcp6_frame_conns[tcp6_payload_used] = conn;
iov = tcp6_l2_iov[tcp6_payload_used++];
- l4len = tcp_l2_buf_fill_headers(conn, iov, dlen, NULL, seq);
+ l4len = tcp_l2_buf_fill_headers(conn, iov, dlen, NULL, seq,
+ false);
iov[TCP_IOV_PAYLOAD].iov_len = l4len;
if (tcp6_payload_used > TCP_FRAMES_MEM - 1)
tcp_payload_flush(c);
diff --git a/tcp_internal.h b/tcp_internal.h
index aa8bb64f1f33..e7fe735bfcb4 100644
--- a/tcp_internal.h
+++ b/tcp_internal.h
@@ -91,7 +91,8 @@ void tcp_rst_do(struct ctx *c, struct tcp_tap_conn *conn);
size_t tcp_l2_buf_fill_headers(const struct tcp_tap_conn *conn,
struct iovec *iov, size_t dlen,
- const uint16_t *check, uint32_t seq);
+ const uint16_t *check, uint32_t seq,
+ bool no_tcp_csum);
int tcp_update_seqack_wnd(const struct ctx *c, struct tcp_tap_conn *conn,
int force_seq, struct tcp_info *tinfo);
int tcp_prepare_flags(struct ctx *c, struct tcp_tap_conn *conn, int flags,
--
2.46.0
On Wed, 18 Sep 2024 15:13:26 +0200 Laurent Vivier <lvivier(a)redhat.com> wrote:With vhost-user we can disable the checksum of UDP and TCP. Add a generic parameter for each of them to disable the checksum. v3: align parameters description with tab s/O/0/ v2: s/UPD/UDP/ Add David's R-b in PATCH 2 Laurent Vivier (2): udp: Allow checksum to be disabled tcp: Allow checksum to be disabledApplied. -- Stefano