migrate, tcp: More careful marshalling of mss parameter during migration

During migration we extract the limit on segment size using TCP_MAXSEG,
and set it on the other side with TCP_REPAIR_OPTIONS.  However, unlike most
32-bit values we transfer we transfer it in native endian, not network
endian.  This is not correct; add it to the list of endian fixups we make.

In addition, while MAXSEG will be 32-bits in practice, and is given as such
to TCP_REPAIR_OPTIONS, the TCP_MAXSEG sockopt treats it as an 'int'.  It's
not strictly safe to pass a uint32_t to a getsockopt() expecting an int,
although we'll get away with it on most (maybe all) platforms.  Correct
this as well.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
[sbrivio: Minor coding style fix]
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
(cherry picked from commit 28772ee91a60b34786023496ea17c2c2f4e5f7f5)

1 files changed, 6 insertions, 1 deletions

diff --git a/tcp.c b/tcp.c
index d00e8fc..65747c7 100644
--- a/tcp.c
+++ b/tcp.c
@@ -2847,13 +2847,16 @@ static int tcp_flow_dump_tinfo(int s, struct tcp_tap_transfer_ext *t)
 static int tcp_flow_dump_mss(int s, struct tcp_tap_transfer_ext *t)
 {
 	socklen_t sl = sizeof(t->mss);
+	int val;
 
-	if (getsockopt(s, SOL_TCP, TCP_MAXSEG, &t->mss, &sl)) {
+	if (getsockopt(s, SOL_TCP, TCP_MAXSEG, &val, &sl)) {
 		int rc = -errno;
 		err_perror("Getting MSS, socket %i", s);
 		return rc;
 	}
 
+	t->mss = (uint32_t)val;
+
 	return 0;
 }
 
@@ -3288,6 +3291,7 @@ int tcp_flow_migrate_source_ext(int fd, int fidx,
 	t->sndq		= htonl(t->sndq);
 	t->notsent	= htonl(t->notsent);
 	t->rcvq		= htonl(t->rcvq);
+	t->mss		= htonl(t->mss);
 
 	t->snd_wl1	= htonl(t->snd_wl1);
 	t->snd_wnd	= htonl(t->snd_wnd);
@@ -3501,6 +3505,7 @@ int tcp_flow_migrate_target_ext(struct ctx *c, union flow *flow, int fd)
 	t.sndq		= ntohl(t.sndq);
 	t.notsent	= ntohl(t.notsent);
 	t.rcvq		= ntohl(t.rcvq);
+	t.mss		= ntohl(t.mss);
 
 	t.snd_wl1	= ntohl(t.snd_wl1);
 	t.snd_wnd	= ntohl(t.snd_wnd);