tcp_retransmit_timer() backs off by doubling icsk_rto, so the backoff only works if icsk_rto is nonzero. __tcp_set_rto() adds srtt_us >> 3 and rttvar_us with no lower bound, and relies on tcp_rtt_estimator() to keep rttvar_us at tcp_rto_min() or above. But tcp_rto_min() returns 0 for a route with a locked RTAX_RTO_MIN of 0. A first RTT sample of 0 usec, which usec TS allows, then gives: srtt_us = max(1U, 0 << 3) = 1 rttvar_us = max(mdev_us, tcp_rto_min_us(sk)) = 0 __tcp_set_rto() = usecs_to_jiffies((1 >> 3) + 0) = 0 Keep the result from 0, so that the backoff recovers from any state of the RTT estimator. Do it in __tcp_set_rto() and not in tcp_set_rto(), since tcp_ld_RTO_revert() uses the same sum. The third caller, the linear timeouts of thin streams in tcp_retransmit_timer(), already clamps the result to tcp_rto_min(). Found by code inspection (Eric Dumazet), not reproduced. Fixed by LLM, and tested with existing kernel selftests. Fixes: 05bb1fad1cde ("[TCP]: Allow minimum RTO to be configurable via routing metrics.") Suggested-by: Eric Dumazet Signed-off-by: Chris Mason --- Notes: v1->v2: no code change, new Fixes: tag (Eric), changelog updates include/net/tcp.h | 4 +++- net/ipv4/tcp_input.c | 5 +++-- 2 files changed, 6 insertions(+), 3 deletions(-) diff --git a/include/net/tcp.h b/include/net/tcp.h index 5c9203aae834..6a714189bcc7 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -910,7 +910,9 @@ static inline void tcp_bound_rto(struct sock *sk) static inline u32 __tcp_set_rto(const struct tcp_sock *tp) { - return usecs_to_jiffies((tp->srtt_us >> 3) + tp->rttvar_us); + u32 rto = usecs_to_jiffies((tp->srtt_us >> 3) + tp->rttvar_us); + + return max(rto, 1U); } static inline unsigned long tcp_reqsk_timeout(struct request_sock *req) diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index a2074a1af422..f65697bc96c9 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -1221,8 +1221,9 @@ void tcp_set_rto(struct sock *sk) * with correct one. It is exactly, which we pretend to do. */ - /* NOTE: clamping at TCP_RTO_MIN is not required, current algo - * guarantees that rto is higher. + /* NOTE: clamping at tcp_rto_min() is not required, tcp_rtt_estimator() + * applies it to rttvar_us. But rto must never be 0, which exponential + * backoff could not leave: __tcp_set_rto() takes care of that. */ tcp_bound_rto(sk); } -- 2.55.0