On 11/04, Stanislav Fomichev wrote:
> On 11/04, Jakub Kicinski wrote:
> > On Mon, 4 Nov 2024 10:14:24 -0800 Stanislav Fomichev wrote:
> > > -static int configure_flow_steering(void)
> > > +static int configure_flow_steering(struct sockaddr_in6 *server_sin)
> > > {
> > > - return run_command("sudo ethtool -N %s flow-type tcp4 %s %s dst-ip %s %s %s dst-port %s queue %d >&2",
> > > + const char *type = "tcp6";
> > > + const char *server_addr;
> > > + char buf[256];
> > > +
> > > + inet_ntop(AF_INET6, &server_sin->sin6_addr, buf, sizeof(buf));
> > > + server_addr = buf;
> > > +
> > > + if (IN6_IS_ADDR_V4MAPPED(&server_sin->sin6_addr)) {
> > > + type = "tcp4";
> > > + server_addr = strrchr(server_addr, ':') + 1;
> > > + }
> > > +
> > > + return run_command("sudo ethtool -N %s flow-type %s %s %s dst-ip %s %s %s dst-port %s queue %d >&2",
> > > ifname,
> > > + type,
> > > client_ip ? "src-ip" : "",
> > > client_ip ?: "",
> > > - server_ip,
> > > + server_addr,
> > > client_ip ? "src-port" : "",
> > > client_ip ? port : "",
> > > port, start_queue);
> >
> > nit: I think this generate a truncation warning, not sure if it's easy
> > to fix:
> >
> > ncdevmem.c:259:28: warning: ‘%s’ directive output may be truncated writing up to 255 bytes into a region of size between 209 and 215 [-Wformat-truncation=]
> > 259 | return run_command("sudo ethtool -N %s flow-type %s %s %s dst-ip %s %s %s dst-port %s queue %d >&2",
> > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >
> > maybe make buf smaller? 🤔️
>
> Are you having some extra -W arguments to make it trigger on the build
> bot? Wonder why I don't see the warning locally.. (will try to pass
> -Wformat-truncation)
It is indeed -Wformat-truncation but only gcc complains about it, not clang.
Setting buf size to 40 calms it down.
