From: David Gibson <david@gibson.dropbear.id.au>
To: Stefano Brivio <sbrivio@redhat.com>, passt-dev@passt.top
Cc: David Gibson <david@gibson.dropbear.id.au>
Subject: [PATCH v3 14/16] udp: Unify buffers for tap and splice paths
Date: Fri, 25 Nov 2022 18:29:14 +1100 [thread overview]
Message-ID: <20221125072916.3060938-15-david@gibson.dropbear.id.au> (raw)
In-Reply-To: <20221125072916.3060938-1-david@gibson.dropbear.id.au>
We maintain a set of buffers for UDP packets to be forwarded via the tap
interface in udp[46]_l2_buf. We then have a separate set of buffers for
packets to be "spliced" in udp_splice_buf[]. However, we only use one of
these at a time, so we can share the buffer space.
For the receiving splice packets we can not only re-use the data buffers
but also the udp[46]_l2_iov_sock and udp[46]_l2_mh_sock control structures.
For sending the splice packets we keep the same data buffers, but we need
specific control structures. We create udp[46]_iov_splice - we can't
reuse udp_l2_iov_sock[] because we need to write iov_len as we're writing
spliced packets, but the tap path expects iov_len to remain the same (it
only uses it for receive). Likewise we create udp[46]_mh_splice with the
mmsghdr structures for sending spliced packets. As well as needing to
reference different iovs, these need to all reference udp_splice_namebuf
instead of individual msg_name fields for each slot.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
udp.c | 71 ++++++++++++++++++++++++++---------------------------------
1 file changed, 31 insertions(+), 40 deletions(-)
diff --git a/udp.c b/udp.c
index 03e88c5..4fc5c5c 100644
--- a/udp.c
+++ b/udp.c
@@ -219,9 +219,6 @@ struct udp6_l2_buf_t {
#endif
udp6_l2_buf[UDP_MAX_FRAMES];
-static struct sockaddr_storage udp_splice_namebuf;
-static uint8_t udp_splice_buf[UDP_MAX_FRAMES][USHRT_MAX];
-
/* recvmmsg()/sendmmsg() data for tap */
static struct iovec udp4_l2_iov_sock [UDP_MAX_FRAMES];
static struct iovec udp6_l2_iov_sock [UDP_MAX_FRAMES];
@@ -236,11 +233,13 @@ static struct mmsghdr udp4_l2_mh_tap [UDP_MAX_FRAMES];
static struct mmsghdr udp6_l2_mh_tap [UDP_MAX_FRAMES];
/* recvmmsg()/sendmmsg() data for "spliced" connections */
-static struct iovec udp_iov_recv [UDP_MAX_FRAMES];
-static struct mmsghdr udp_mmh_recv [UDP_MAX_FRAMES];
+static struct sockaddr_storage udp_splice_namebuf;
-static struct iovec udp_iov_sendto [UDP_MAX_FRAMES];
-static struct mmsghdr udp_mmh_sendto [UDP_MAX_FRAMES];
+static struct iovec udp4_iov_splice [UDP_MAX_FRAMES];
+static struct iovec udp6_iov_splice [UDP_MAX_FRAMES];
+
+static struct mmsghdr udp4_mh_splice [UDP_MAX_FRAMES];
+static struct mmsghdr udp6_mh_splice [UDP_MAX_FRAMES];
/**
* udp_invert_portmap() - Compute reverse port translations for return packets
@@ -522,16 +521,25 @@ static void udp_sock_handler_splice(const struct ctx *c, union epoll_ref ref,
{
in_port_t src, dst = ref.r.p.udp.udp.port;
int s, v6 = ref.r.p.udp.udp.v6, n, i;
+ struct mmsghdr *mmh_recv, *mmh_send;
if (!(events & EPOLLIN))
return;
- n = recvmmsg(ref.r.s, udp_mmh_recv, UDP_MAX_FRAMES, 0, NULL);
+ if (v6) {
+ mmh_recv = udp6_l2_mh_sock;
+ mmh_send = udp6_mh_splice;
+ } else {
+ mmh_recv = udp4_l2_mh_sock;
+ mmh_send = udp4_mh_splice;
+ }
+
+ n = recvmmsg(ref.r.s, mmh_recv, UDP_MAX_FRAMES, 0, NULL);
if (n <= 0)
return;
- src = sa_port(v6, udp_mmh_recv[0].msg_hdr.msg_name);
+ src = sa_port(v6, mmh_recv[0].msg_hdr.msg_name);
if (ref.r.p.udp.udp.ns) {
src += c->udp.fwd_in.rdelta[src];
@@ -562,11 +570,8 @@ static void udp_sock_handler_splice(const struct ctx *c, union epoll_ref ref,
udp_splice_ns[v6][src].ts = now->tv_sec;
}
- for (i = 0; i < n; i++) {
- struct msghdr *mh_s = &udp_mmh_sendto[i].msg_hdr;
-
- mh_s->msg_iov->iov_len = udp_mmh_recv[i].msg_len;
- }
+ for (i = 0; i < n; i++)
+ mmh_send[i].msg_hdr.msg_iov->iov_len = mmh_recv[i].msg_len;
if (v6) {
*((struct sockaddr_in6 *)&udp_splice_namebuf) =
@@ -586,7 +591,7 @@ static void udp_sock_handler_splice(const struct ctx *c, union epoll_ref ref,
});
}
- sendmmsg(s, udp_mmh_sendto, n, MSG_NOSIGNAL);
+ sendmmsg(s, mmh_send, n, MSG_NOSIGNAL);
}
/**
@@ -1163,37 +1168,23 @@ int udp_sock_init_ns(void *arg)
*/
static void udp_splice_iov_init(void)
{
- struct mmsghdr *h;
- struct iovec *iov;
int i;
- for (i = 0, h = udp_mmh_recv; i < UDP_MAX_FRAMES; i++, h++) {
- struct msghdr *mh = &h->msg_hdr;
-
- if (!i) {
- mh->msg_name = &udp_splice_namebuf;
- mh->msg_namelen = sizeof(udp_splice_namebuf);
- }
-
- mh->msg_iov = &udp_iov_recv[i];
- mh->msg_iovlen = 1;
- }
- for (i = 0, iov = udp_iov_recv; i < UDP_MAX_FRAMES; i++, iov++) {
- iov->iov_base = udp_splice_buf[i];
- iov->iov_len = sizeof(udp_splice_buf[i]);
- }
+ for (i = 0; i < UDP_MAX_FRAMES; i++) {
+ struct msghdr *mh4 = &udp4_mh_splice[i].msg_hdr;
+ struct msghdr *mh6 = &udp6_mh_splice[i].msg_hdr;
- for (i = 0, h = udp_mmh_sendto; i < UDP_MAX_FRAMES; i++, h++) {
- struct msghdr *mh = &h->msg_hdr;
+ mh4->msg_name = mh6->msg_name = &udp_splice_namebuf;
+ mh4->msg_namelen = sizeof(udp_splice_namebuf);
+ mh6->msg_namelen = sizeof(udp_splice_namebuf);
- mh->msg_name = &udp_splice_namebuf;
- mh->msg_namelen = sizeof(udp_splice_namebuf);
+ udp4_iov_splice[i].iov_base = udp4_l2_buf[i].data;
+ udp6_iov_splice[i].iov_base = udp6_l2_buf[i].data;
- mh->msg_iov = &udp_iov_sendto[i];
- mh->msg_iovlen = 1;
+ mh4->msg_iov = &udp4_iov_splice[i];
+ mh6->msg_iov = &udp6_iov_splice[i];
+ mh4->msg_iovlen = mh6->msg_iovlen = 1;
}
- for (i = 0, iov = udp_iov_sendto; i < UDP_MAX_FRAMES; i++, iov++)
- iov->iov_base = udp_splice_buf[i];
}
/**
--
@@ -219,9 +219,6 @@ struct udp6_l2_buf_t {
#endif
udp6_l2_buf[UDP_MAX_FRAMES];
-static struct sockaddr_storage udp_splice_namebuf;
-static uint8_t udp_splice_buf[UDP_MAX_FRAMES][USHRT_MAX];
-
/* recvmmsg()/sendmmsg() data for tap */
static struct iovec udp4_l2_iov_sock [UDP_MAX_FRAMES];
static struct iovec udp6_l2_iov_sock [UDP_MAX_FRAMES];
@@ -236,11 +233,13 @@ static struct mmsghdr udp4_l2_mh_tap [UDP_MAX_FRAMES];
static struct mmsghdr udp6_l2_mh_tap [UDP_MAX_FRAMES];
/* recvmmsg()/sendmmsg() data for "spliced" connections */
-static struct iovec udp_iov_recv [UDP_MAX_FRAMES];
-static struct mmsghdr udp_mmh_recv [UDP_MAX_FRAMES];
+static struct sockaddr_storage udp_splice_namebuf;
-static struct iovec udp_iov_sendto [UDP_MAX_FRAMES];
-static struct mmsghdr udp_mmh_sendto [UDP_MAX_FRAMES];
+static struct iovec udp4_iov_splice [UDP_MAX_FRAMES];
+static struct iovec udp6_iov_splice [UDP_MAX_FRAMES];
+
+static struct mmsghdr udp4_mh_splice [UDP_MAX_FRAMES];
+static struct mmsghdr udp6_mh_splice [UDP_MAX_FRAMES];
/**
* udp_invert_portmap() - Compute reverse port translations for return packets
@@ -522,16 +521,25 @@ static void udp_sock_handler_splice(const struct ctx *c, union epoll_ref ref,
{
in_port_t src, dst = ref.r.p.udp.udp.port;
int s, v6 = ref.r.p.udp.udp.v6, n, i;
+ struct mmsghdr *mmh_recv, *mmh_send;
if (!(events & EPOLLIN))
return;
- n = recvmmsg(ref.r.s, udp_mmh_recv, UDP_MAX_FRAMES, 0, NULL);
+ if (v6) {
+ mmh_recv = udp6_l2_mh_sock;
+ mmh_send = udp6_mh_splice;
+ } else {
+ mmh_recv = udp4_l2_mh_sock;
+ mmh_send = udp4_mh_splice;
+ }
+
+ n = recvmmsg(ref.r.s, mmh_recv, UDP_MAX_FRAMES, 0, NULL);
if (n <= 0)
return;
- src = sa_port(v6, udp_mmh_recv[0].msg_hdr.msg_name);
+ src = sa_port(v6, mmh_recv[0].msg_hdr.msg_name);
if (ref.r.p.udp.udp.ns) {
src += c->udp.fwd_in.rdelta[src];
@@ -562,11 +570,8 @@ static void udp_sock_handler_splice(const struct ctx *c, union epoll_ref ref,
udp_splice_ns[v6][src].ts = now->tv_sec;
}
- for (i = 0; i < n; i++) {
- struct msghdr *mh_s = &udp_mmh_sendto[i].msg_hdr;
-
- mh_s->msg_iov->iov_len = udp_mmh_recv[i].msg_len;
- }
+ for (i = 0; i < n; i++)
+ mmh_send[i].msg_hdr.msg_iov->iov_len = mmh_recv[i].msg_len;
if (v6) {
*((struct sockaddr_in6 *)&udp_splice_namebuf) =
@@ -586,7 +591,7 @@ static void udp_sock_handler_splice(const struct ctx *c, union epoll_ref ref,
});
}
- sendmmsg(s, udp_mmh_sendto, n, MSG_NOSIGNAL);
+ sendmmsg(s, mmh_send, n, MSG_NOSIGNAL);
}
/**
@@ -1163,37 +1168,23 @@ int udp_sock_init_ns(void *arg)
*/
static void udp_splice_iov_init(void)
{
- struct mmsghdr *h;
- struct iovec *iov;
int i;
- for (i = 0, h = udp_mmh_recv; i < UDP_MAX_FRAMES; i++, h++) {
- struct msghdr *mh = &h->msg_hdr;
-
- if (!i) {
- mh->msg_name = &udp_splice_namebuf;
- mh->msg_namelen = sizeof(udp_splice_namebuf);
- }
-
- mh->msg_iov = &udp_iov_recv[i];
- mh->msg_iovlen = 1;
- }
- for (i = 0, iov = udp_iov_recv; i < UDP_MAX_FRAMES; i++, iov++) {
- iov->iov_base = udp_splice_buf[i];
- iov->iov_len = sizeof(udp_splice_buf[i]);
- }
+ for (i = 0; i < UDP_MAX_FRAMES; i++) {
+ struct msghdr *mh4 = &udp4_mh_splice[i].msg_hdr;
+ struct msghdr *mh6 = &udp6_mh_splice[i].msg_hdr;
- for (i = 0, h = udp_mmh_sendto; i < UDP_MAX_FRAMES; i++, h++) {
- struct msghdr *mh = &h->msg_hdr;
+ mh4->msg_name = mh6->msg_name = &udp_splice_namebuf;
+ mh4->msg_namelen = sizeof(udp_splice_namebuf);
+ mh6->msg_namelen = sizeof(udp_splice_namebuf);
- mh->msg_name = &udp_splice_namebuf;
- mh->msg_namelen = sizeof(udp_splice_namebuf);
+ udp4_iov_splice[i].iov_base = udp4_l2_buf[i].data;
+ udp6_iov_splice[i].iov_base = udp6_l2_buf[i].data;
- mh->msg_iov = &udp_iov_sendto[i];
- mh->msg_iovlen = 1;
+ mh4->msg_iov = &udp4_iov_splice[i];
+ mh6->msg_iov = &udp6_iov_splice[i];
+ mh4->msg_iovlen = mh6->msg_iovlen = 1;
}
- for (i = 0, iov = udp_iov_sendto; i < UDP_MAX_FRAMES; i++, iov++)
- iov->iov_base = udp_splice_buf[i];
}
/**
--
2.38.1
next prev parent reply other threads:[~2022-11-25 7:29 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-11-25 7:29 [PATCH v3 00/16] Simplify and correct handling of "spliced" UDP forwarding David Gibson
2022-11-25 7:29 ` [PATCH v3 01/16] udp: Also bind() connected ports for "splice" forwarding David Gibson
2022-11-25 7:29 ` [PATCH v3 02/16] udp: Separate tracking of inbound and outbound packet flows David Gibson
2022-11-25 7:29 ` [PATCH v3 03/16] udp: Always use sendto() rather than send() for forwarding spliced packets David Gibson
2022-11-25 7:29 ` [PATCH v3 04/16] udp: Don't connect "forward" sockets for spliced flows David Gibson
2022-11-25 7:29 ` [PATCH v3 05/16] udp: Remove the @bound field from union udp_epoll_ref David Gibson
2022-11-25 7:29 ` [PATCH v3 06/16] udp: Split splice field in udp_epoll_ref into (mostly) independent bits David Gibson
2022-11-25 7:29 ` [PATCH v3 07/16] udp: Don't create double sockets for -U port David Gibson
2022-11-25 7:29 ` [PATCH v3 08/16] udp: Re-use fixed bound sockets for packet forwarding when possible David Gibson
2022-11-25 7:29 ` [PATCH v3 09/16] udp: Don't explicitly track originating socket for spliced "connections" David Gibson
2022-11-25 7:29 ` [PATCH v3 10/16] udp: Update UDP "connection" timestamps in both directions David Gibson
2022-11-25 7:29 ` [PATCH v3 11/16] udp: Simplify udp_sock_handler_splice David Gibson
2022-11-25 7:29 ` [PATCH v3 12/16] udp: Make UDP_SPLICE_FRAMES and UDP_TAP_FRAMES_MEM the same thing David Gibson
2022-11-25 7:29 ` [PATCH v3 13/16] udp: Add helper to extract port from a sockaddr_in or sockaddr_in6 David Gibson
2022-11-25 7:29 ` David Gibson [this message]
2022-11-25 7:29 ` [PATCH v3 15/16] udp: Split send half of udp_sock_handler_splice() from the receive half David Gibson
2022-11-25 7:29 ` [PATCH v3 16/16] udp: Correct splice forwarding when receiving from multiple sources David Gibson
2022-11-29 5:59 ` 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=20221125072916.3060938-15-david@gibson.dropbear.id.au \
--to=david@gibson.dropbear.id.au \
--cc=passt-dev@passt.top \
--cc=sbrivio@redhat.com \
/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).