1
0
Fork 0
mirror of https://passt.top/passt synced 2025-06-12 10:35:34 +02:00

passt: Spare some syscalls, add some optimisations from profiling

Avoid a bunch of syscalls on forwarding paths by:

- storing minimum and maximum file descriptor numbers for each
  protocol, fall back to SO_PROTOCOL query only on overlaps

- allocating a larger receive buffer -- this can result in more
  coalesced packets than sendmmsg() can take (UIO_MAXIOV, i.e. 1024),
  so make sure we don't exceed that within a single call to protocol
  tap handlers

- nesting the handling loop in tap_handler() in the receive loop,
  so that we have better chances of filling our receive buffer in
  fewer calls

- skipping the recvfrom() in the UDP handler on EPOLLERR -- there's
  nothing to be done in that case

and while at it:

- restore the 20ms timer interval for periodic (TCP) events, I
  accidentally changed that to 100ms in an earlier commit

- attempt using SO_ZEROCOPY for UDP -- if it's not available,
  sendmmsg() will succeed anyway

- fix the handling of the status code from sendmmsg(), if it fails,
  we'll try to discard the first message, hence return 1 from the
  UDP handler

Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
This commit is contained in:
Stefano Brivio 2021-04-23 22:22:37 +02:00
parent 962bc97cf1
commit 38b50dba47
10 changed files with 139 additions and 49 deletions

8
tcp.c
View file

@ -1003,6 +1003,11 @@ static void tcp_conn_from_sock(struct ctx *c, int fd)
if (s == -1)
return;
if (s < c->tcp.fd_min)
c->tcp.fd_min = s;
if (s > c->tcp.fd_max)
c->tcp.fd_max = s;
if (sa_l.ss_family == AF_INET) {
struct sockaddr_in *sa4 = (struct sockaddr_in *)&sa_r;
@ -1445,6 +1450,9 @@ int tcp_sock_init(struct ctx *c)
{
in_port_t port;
c->tcp.fd_min = INT_MAX;
c->tcp.fd_max = 0;
for (port = 0; port < (1 << 15) + (1 << 14); port++) {
if (c->v4 && sock_l4_add(c, 4, IPPROTO_TCP, port) < 0)
return -1;