Since commit 7d3f3b4367f3 ("net: ipv4: Cache pmtu for all packet paths if multipath enabled") __ip_rt_update_pmtu() updates the exception of every path of a multipath route, but update_or_create_fnhe() overwrites whatever pmtu a path already has. A Packet Too Big from a wider path thus raises the pmtu a narrower path has learned, and flows on that path blackhole until it gets its own Packet Too Big again. The early return in __ip_rt_update_pmtu() already keeps a pmtu from going up on the path the Packet Too Big was looked up on. Do the same for the other paths: leave an unexpired exception with a lower pmtu alone. Signed-off-by: Vladimir Vdovin --- net/ipv4/route.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/net/ipv4/route.c b/net/ipv4/route.c index d0a5f6d77094..596af782c1fb 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -676,6 +676,14 @@ static void update_or_create_fnhe(struct fib_nh_common *nhc, __be32 daddr, } if (fnhe) { + /* __ip_rt_update_pmtu() caches a pmtu on every path of a + * multipath route, don't let a Packet Too Big from one path + * raise the pmtu another path already learned. + */ + if (pmtu && fnhe->fnhe_genid == genid && fnhe->fnhe_pmtu && + pmtu > fnhe->fnhe_pmtu && + time_before(jiffies, fnhe->fnhe_expires)) + goto out_unlock; if (fnhe->fnhe_genid != genid) fnhe->fnhe_genid = genid; if (gw) -- 2.56.0