ping: support ipv6 ping socket flow labels

Ping sockets don't appear to make any attempt to preserve flow labels
created and set by userspace using IPV6_FLOWINFO_SEND. Instead they are
clobbered by autolabels (if enabled) or zero.

Grab the flowlabel out of the msghdr similar to how rawv6_sendmsg does
it and move the memset up so it doesn't get zeroed after.
Signed-off-by: Alan Brady <alan.brady@intel.com>
Tested-by: Gurucharan <gurucharanx.g@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/ipv6/ping.c

@@ -64,6 +64,8 @@ static int ping_v6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
 	if (err)
 		return err;
 
+	memset(&fl6, 0, sizeof(fl6));
+
 	if (msg->msg_name) {
 		DECLARE_SOCKADDR(struct sockaddr_in6 *, u, msg->msg_name);
 		if (msg->msg_namelen < sizeof(*u))
@@ -72,12 +74,15 @@ static int ping_v6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
 			return -EAFNOSUPPORT;
 		}
 		daddr = &(u->sin6_addr);
+		if (np->sndflow)
+			fl6.flowlabel = u->sin6_flowinfo & IPV6_FLOWINFO_MASK;
 		if (__ipv6_addr_needs_scope_id(ipv6_addr_type(daddr)))
 			oif = u->sin6_scope_id;
 	} else {
 		if (sk->sk_state != TCP_ESTABLISHED)
 			return -EDESTADDRREQ;
 		daddr = &sk->sk_v6_daddr;
+		fl6.flowlabel = np->flow_label;
 	}
 
 	if (!oif)
@@ -101,7 +106,6 @@ static int ping_v6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
 	ipc6.sockc.tsflags = sk->sk_tsflags;
 	ipc6.sockc.mark = sk->sk_mark;
 
-	memset(&fl6, 0, sizeof(fl6));
 	fl6.flowi6_oif = oif;
 
 	if (msg->msg_controllen) {

tools/testing/selftests/net/ipv6_flowlabel.c

@@ -9,6 +9,7 @@
 #include <errno.h>
 #include <fcntl.h>
 #include <limits.h>
+#include <linux/icmpv6.h>
 #include <linux/in6.h>
 #include <stdbool.h>
 #include <stdio.h>
@@ -29,26 +30,48 @@
 #ifndef IPV6_FLOWLABEL_MGR
 #define IPV6_FLOWLABEL_MGR 32
 #endif
+#ifndef IPV6_FLOWINFO_SEND
+#define IPV6_FLOWINFO_SEND 33
+#endif
 
 #define FLOWLABEL_WILDCARD	((uint32_t) -1)
 
 static const char cfg_data[]	= "a";
 static uint32_t cfg_label	= 1;
+static bool use_ping;
+static bool use_flowinfo_send;
+
+static struct icmp6hdr icmp6 = {
+	.icmp6_type = ICMPV6_ECHO_REQUEST
+};
+
+static struct sockaddr_in6 addr = {
+	.sin6_family = AF_INET6,
+	.sin6_addr = IN6ADDR_LOOPBACK_INIT,
+};
 
 static void do_send(int fd, bool with_flowlabel, uint32_t flowlabel)
 {
 	char control[CMSG_SPACE(sizeof(flowlabel))] = {0};
 	struct msghdr msg = {0};
-	struct iovec iov = {0};
+	struct iovec iov = {
+		.iov_base = (char *)cfg_data,
+		.iov_len = sizeof(cfg_data)
+	};
 	int ret;
 
-	iov.iov_base = (char *)cfg_data;
-	iov.iov_len = sizeof(cfg_data);
+	if (use_ping) {
+		iov.iov_base = &icmp6;
+		iov.iov_len = sizeof(icmp6);
+	}
 
 	msg.msg_iov = &iov;
 	msg.msg_iovlen = 1;
 
-	if (with_flowlabel) {
+	if (use_flowinfo_send) {
+		msg.msg_name = &addr;
+		msg.msg_namelen = sizeof(addr);
+	} else if (with_flowlabel) {
 		struct cmsghdr *cm;
 
 		cm = (void *)control;
@@ -94,6 +117,8 @@ static void do_recv(int fd, bool with_flowlabel, uint32_t expect)
 	ret = recvmsg(fd, &msg, 0);
 	if (ret == -1)
 		error(1, errno, "recv");
+	if (use_ping)
+		goto parse_cmsg;
 	if (msg.msg_flags & (MSG_TRUNC | MSG_CTRUNC))
 		error(1, 0, "recv: truncated");
 	if (ret != sizeof(cfg_data))
@@ -101,6 +126,7 @@ static void do_recv(int fd, bool with_flowlabel, uint32_t expect)
 	if (memcmp(data, cfg_data, sizeof(data)))
 		error(1, 0, "recv: data mismatch");
 
+parse_cmsg:
 	cm = CMSG_FIRSTHDR(&msg);
 	if (with_flowlabel) {
 		if (!cm)
@@ -114,9 +140,11 @@ static void do_recv(int fd, bool with_flowlabel, uint32_t expect)
 		flowlabel = ntohl(*(uint32_t *)CMSG_DATA(cm));
 		fprintf(stderr, "recv with label %u\n", flowlabel);
 
-		if (expect != FLOWLABEL_WILDCARD && expect != flowlabel)
+		if (expect != FLOWLABEL_WILDCARD && expect != flowlabel) {
 			fprintf(stderr, "recv: incorrect flowlabel %u != %u\n",
 					flowlabel, expect);
+			error(1, 0, "recv: flowlabel is wrong");
+		}
 
 	} else {
 		fprintf(stderr, "recv without label\n");
@@ -165,11 +193,17 @@ static void parse_opts(int argc, char **argv)
 {
 	int c;
 
-	while ((c = getopt(argc, argv, "l:")) != -1) {
+	while ((c = getopt(argc, argv, "l:ps")) != -1) {
 		switch (c) {
 		case 'l':
 			cfg_label = strtoul(optarg, NULL, 0);
 			break;
+		case 'p':
+			use_ping = true;
+			break;
+		case 's':
+			use_flowinfo_send = true;
+			break;
 		default:
 			error(1, 0, "%s: parse error", argv[0]);
 		}
@@ -178,27 +212,30 @@ static void parse_opts(int argc, char **argv)
 
 int main(int argc, char **argv)
 {
-	struct sockaddr_in6 addr = {
-		.sin6_family = AF_INET6,
-		.sin6_port = htons(8000),
-		.sin6_addr = IN6ADDR_LOOPBACK_INIT,
-	};
 	const int one = 1;
 	int fdt, fdr;
+	int prot = 0;
+
+	addr.sin6_port = htons(8000);
 
 	parse_opts(argc, argv);
 
-	fdt = socket(PF_INET6, SOCK_DGRAM, 0);
+	if (use_ping) {
+		fprintf(stderr, "attempting to use ping sockets\n");
+		prot = IPPROTO_ICMPV6;
+	}
+
+	fdt = socket(PF_INET6, SOCK_DGRAM, prot);
 	if (fdt == -1)
 		error(1, errno, "socket t");
 
-	fdr = socket(PF_INET6, SOCK_DGRAM, 0);
+	fdr = use_ping ? fdt : socket(PF_INET6, SOCK_DGRAM, 0);
 	if (fdr == -1)
 		error(1, errno, "socket r");
 
 	if (connect(fdt, (void *)&addr, sizeof(addr)))
 		error(1, errno, "connect");
-	if (bind(fdr, (void *)&addr, sizeof(addr)))
+	if (!use_ping && bind(fdr, (void *)&addr, sizeof(addr)))
 		error(1, errno, "bind");
 
 	flowlabel_get(fdt, cfg_label, IPV6_FL_S_EXCL, IPV6_FL_F_CREATE);
@@ -216,13 +253,21 @@ int main(int argc, char **argv)
 		do_recv(fdr, false, 0);
 	}
 
+	if (use_flowinfo_send) {
+		fprintf(stderr, "using IPV6_FLOWINFO_SEND to send label\n");
+		addr.sin6_flowinfo = htonl(cfg_label);
+		if (setsockopt(fdt, SOL_IPV6, IPV6_FLOWINFO_SEND, &one,
+			       sizeof(one)) == -1)
+			error(1, errno, "setsockopt flowinfo_send");
+	}
+
 	fprintf(stderr, "send label\n");
 	do_send(fdt, true, cfg_label);
 	do_recv(fdr, true, cfg_label);
 
 	if (close(fdr))
 		error(1, errno, "close r");
-	if (close(fdt))
+	if (!use_ping && close(fdt))
 		error(1, errno, "close t");
 
 	return 0;

tools/testing/selftests/net/ipv6_flowlabel.sh

@@ -18,4 +18,20 @@ echo "TEST datapath (with auto-flowlabels)"
 ./in_netns.sh \
   sh -c 'sysctl -q -w net.ipv6.auto_flowlabels=1 && ./ipv6_flowlabel -l 1'
 
+echo "TEST datapath (with ping-sockets)"
+./in_netns.sh \
+  sh -c 'sysctl -q -w net.ipv6.flowlabel_reflect=4 && \
+    sysctl -q -w net.ipv4.ping_group_range="0 2147483647" && \
+    ./ipv6_flowlabel -l 1 -p'
+
+echo "TEST datapath (with flowinfo-send)"
+./in_netns.sh \
+  sh -c './ipv6_flowlabel -l 1 -s'
+
+echo "TEST datapath (with ping-sockets flowinfo-send)"
+./in_netns.sh \
+  sh -c 'sysctl -q -w net.ipv6.flowlabel_reflect=4 && \
+    sysctl -q -w net.ipv4.ping_group_range="0 2147483647" && \
+    ./ipv6_flowlabel -l 1 -p -s'
+
 echo OK. All tests passed