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 06/16] udp: Split splice field in udp_epoll_ref into (mostly) independent bits
Date: Fri, 25 Nov 2022 18:29:06 +1100 [thread overview]
Message-ID: <20221125072916.3060938-7-david@gibson.dropbear.id.au> (raw)
In-Reply-To: <20221125072916.3060938-1-david@gibson.dropbear.id.au>
The @splice field in union udp_epoll_ref can have a number of values for
different types of "spliced" packet flows. Split it into several single
bit fields with more or less independent meanings. The new @splice field
is just a boolean indicating whether the socket is associated with a
spliced flow, making it identical to the @splice fiend in tcp_epoll_ref.
The new bit @orig, indicates whether this is a socket which can originate
new udp packet flows (created with -u or -U) or a socket created on the
fly to handle reply socket. @ns indicates whether the socket lives in the
init namespace or the pasta namespace.
Making these bits more orthogonal to each other will simplify some future
cleanups.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
passt.h | 2 ++
udp.c | 53 ++++++++++++++++++++++++++---------------------------
udp.h | 15 +++++++--------
3 files changed, 35 insertions(+), 35 deletions(-)
diff --git a/passt.h b/passt.h
index 6649c0a..01d8893 100644
--- a/passt.h
+++ b/passt.h
@@ -31,6 +31,8 @@ struct tap_l4_msg {
union epoll_ref;
+#include <stdbool.h>
+
#include "packet.h"
#include "icmp.h"
#include "port_fwd.h"
diff --git a/udp.c b/udp.c
index da0728a..8c5deee 100644
--- a/udp.c
+++ b/udp.c
@@ -46,19 +46,20 @@
* - from init to namespace:
*
* - forward direction: 127.0.0.1:5000 -> 127.0.0.1:80 in init from socket s,
- * with epoll reference: index = 80, splice = UDP_TO_NS
+ * with epoll reference: index = 80, splice = 1, orig = 1, ns = 0
* - if udp_splice_to_ns[V4][5000].target_sock:
* - send packet to udp_splice_to_ns[V4][5000].target_sock, with
* destination port 80
* - otherwise:
* - create new socket udp_splice_to_ns[V4][5000].target_sock
* - bind in namespace to 127.0.0.1:5000
- * - add to epoll with reference: index = 5000, splice: UDP_BACK_TO_INIT
+ * - add to epoll with reference: index = 5000, splice = 1, orig = 0,
+ * ns = 1
* - set udp_splice_to_ns[V4][5000].orig_sock to s
* - update udp_splice_to_ns[V4][5000].ts with current time
*
* - reverse direction: 127.0.0.1:80 -> 127.0.0.1:5000 in namespace socket s,
- * having epoll reference: index = 5000, splice = UDP_BACK_TO_INIT
+ * having epoll reference: index = 5000, splice = 1, orig = 0, ns = 1
* - if udp_splice_to_ns[V4][5000].orig_sock:
* - send to udp_splice_to_ns[V4][5000].orig_sock, with destination port
* 5000
@@ -67,19 +68,20 @@
* - from namespace to init:
*
* - forward direction: 127.0.0.1:2000 -> 127.0.0.1:22 in namespace from
- * socket s, with epoll reference: index = 22, splice = UDP_TO_INIT
+ * socket s, with epoll reference: index = 22, splice = 1, orig = 1, ns = 1
* - if udp4_splice_to_init[V4][2000].target_sock:
* - send packet to udp_splice_to_init[V4][2000].target_sock, with
* destination port 22
* - otherwise:
* - create new socket udp_splice_to_init[V4][2000].target_sock
* - bind in init to 127.0.0.1:2000
- * - add to epoll with reference: index = 2000, splice = UDP_BACK_TO_NS
+ * - add to epoll with reference: index = 2000, splice = 1, orig = 0,
+ * ns = 0
* - set udp_splice_to_init[V4][2000].orig_sock to s
* - update udp_splice_to_init[V4][2000].ts with current time
*
* - reverse direction: 127.0.0.1:22 -> 127.0.0.1:2000 in init from socket s,
- * having epoll reference: index = 2000, splice = UDP_BACK_TO_NS
+ * having epoll reference: index = 2000, splice = 1, orig = 0, ns = 0
* - if udp_splice_to_init[V4][2000].orig_sock:
* - send to udp_splice_to_init[V4][2000].orig_sock, with destination port
* 2000
@@ -404,17 +406,17 @@ static void udp_sock6_iov_init(void)
* #syscalls:pasta getsockname
*/
int udp_splice_new(const struct ctx *c, int v6, int bound_sock, in_port_t src,
- int splice)
+ bool ns)
{
struct epoll_event ev = { .events = EPOLLIN | EPOLLRDHUP | EPOLLHUP };
union epoll_ref ref = { .r.proto = IPPROTO_UDP,
- .r.p.udp.udp = { .splice = splice, .v6 = v6,
- .port = src }
+ .r.p.udp.udp = { .splice = true, .ns = ns,
+ .v6 = v6, .port = src }
};
struct udp_splice_flow *flow;
int act, s;
- if (splice == UDP_BACK_TO_INIT) {
+ if (ns) {
flow = &udp_splice_to_ns[v6 ? V6 : V4][src];
act = UDP_ACT_SPLICE_NS;
} else {
@@ -498,8 +500,7 @@ static int udp_splice_new_ns(void *arg)
if (ns_enter(a->c))
return 0;
- a->s = udp_splice_new(a->c, a->v6, a->bound_sock, a->src,
- UDP_BACK_TO_INIT);
+ a->s = udp_splice_new(a->c, a->v6, a->bound_sock, a->src, true);
return 0;
}
@@ -537,8 +538,8 @@ static void udp_sock_handler_splice(const struct ctx *c, union epoll_ref ref,
src = ntohs(sa->sin_port);
}
- switch (ref.r.p.udp.udp.splice) {
- case UDP_TO_NS:
+
+ if (ref.r.p.udp.udp.orig && !ref.r.p.udp.udp.ns) {
src += c->udp.fwd_out.rdelta[src];
if (!(s = udp_splice_to_ns[v6][src].target_sock)) {
@@ -550,27 +551,24 @@ static void udp_sock_handler_splice(const struct ctx *c, union epoll_ref ref,
if ((s = arg.s) < 0)
return;
}
+
udp_splice_to_ns[v6][src].ts = now->tv_sec;
- break;
- case UDP_BACK_TO_INIT:
+ } else if (!ref.r.p.udp.udp.orig && ref.r.p.udp.udp.ns) {
if (!(s = udp_splice_to_ns[v6][dst].orig_sock))
return;
- break;
- case UDP_TO_INIT:
+ } else if (ref.r.p.udp.udp.orig && ref.r.p.udp.udp.ns) {
src += c->udp.fwd_in.rdelta[src];
if (!(s = udp_splice_to_init[v6][src].target_sock)) {
- s = udp_splice_new(c, v6, ref.r.s, src, UDP_BACK_TO_NS);
+ s = udp_splice_new(c, v6, ref.r.s, src, false);
if (s < 0)
return;
}
udp_splice_to_init[v6][src].ts = now->tv_sec;
- break;
- case UDP_BACK_TO_NS:
+ } else if (!ref.r.p.udp.udp.orig && !ref.r.p.udp.udp.ns) {
if (!(s = udp_splice_to_init[v6][dst].orig_sock))
return;
- break;
- default:
+ } else {
return;
}
@@ -1096,15 +1094,16 @@ void udp_sock_init(const struct ctx *c, int ns, sa_family_t af,
if (c->mode == MODE_PASTA) {
bind_addr = &(uint32_t){ htonl(INADDR_LOOPBACK) };
- uref.udp.splice = UDP_TO_NS;
+ uref.udp.splice = uref.udp.orig = true;
sock_l4(c, AF_INET, IPPROTO_UDP, bind_addr, ifname,
port, uref.u32);
}
if (ns) {
+ uref.udp.splice = uref.udp.orig = uref.udp.ns = true;
+
bind_addr = &(uint32_t){ htonl(INADDR_LOOPBACK) };
- uref.udp.splice = UDP_TO_INIT;
sock_l4(c, AF_INET, IPPROTO_UDP, bind_addr, ifname,
port, uref.u32);
@@ -1129,7 +1128,7 @@ void udp_sock_init(const struct ctx *c, int ns, sa_family_t af,
if (c->mode == MODE_PASTA) {
bind_addr = &in6addr_loopback;
- uref.udp.splice = UDP_TO_NS;
+ uref.udp.splice = uref.udp.orig = true;
sock_l4(c, AF_INET6, IPPROTO_UDP, bind_addr, ifname,
port, uref.u32);
@@ -1137,7 +1136,7 @@ void udp_sock_init(const struct ctx *c, int ns, sa_family_t af,
if (ns) {
bind_addr = &in6addr_loopback;
- uref.udp.splice = UDP_TO_INIT;
+ uref.udp.splice = uref.udp.orig = uref.udp.ns = true;
sock_l4(c, AF_INET6, IPPROTO_UDP, bind_addr, ifname,
port, uref.u32);
diff --git a/udp.h b/udp.h
index 43bd28a..053991e 100644
--- a/udp.h
+++ b/udp.h
@@ -23,20 +23,19 @@ void udp_update_l2_buf(const unsigned char *eth_d, const unsigned char *eth_s,
* union udp_epoll_ref - epoll reference portion for TCP connections
* @bound: Set if this file descriptor is a bound socket
* @splice: Set if descriptor is associated to "spliced" connection
+ * @orig: Set if a spliced socket which can originate "connections"
+ * @ns: Set if this is a socket in the pasta network namespace
* @v6: Set for IPv6 sockets or connections
* @port: Source port for connected sockets, bound port otherwise
* @u32: Opaque u32 value of reference
*/
union udp_epoll_ref {
struct {
- uint32_t splice:3,
-#define UDP_TO_NS 1
-#define UDP_TO_INIT 2
-#define UDP_BACK_TO_NS 3
-#define UDP_BACK_TO_INIT 4
-
- v6:1,
- port:16;
+ bool splice:1,
+ orig:1,
+ ns:1,
+ v6:1;
+ uint32_t port:16;
} udp;
uint32_t u32;
};
--
@@ -23,20 +23,19 @@ void udp_update_l2_buf(const unsigned char *eth_d, const unsigned char *eth_s,
* union udp_epoll_ref - epoll reference portion for TCP connections
* @bound: Set if this file descriptor is a bound socket
* @splice: Set if descriptor is associated to "spliced" connection
+ * @orig: Set if a spliced socket which can originate "connections"
+ * @ns: Set if this is a socket in the pasta network namespace
* @v6: Set for IPv6 sockets or connections
* @port: Source port for connected sockets, bound port otherwise
* @u32: Opaque u32 value of reference
*/
union udp_epoll_ref {
struct {
- uint32_t splice:3,
-#define UDP_TO_NS 1
-#define UDP_TO_INIT 2
-#define UDP_BACK_TO_NS 3
-#define UDP_BACK_TO_INIT 4
-
- v6:1,
- port:16;
+ bool splice:1,
+ orig:1,
+ ns:1,
+ v6:1;
+ uint32_t port:16;
} udp;
uint32_t u32;
};
--
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 ` David Gibson [this message]
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 ` [PATCH v3 14/16] udp: Unify buffers for tap and splice paths David Gibson
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=20221125072916.3060938-7-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).