From c410bf01933e5e09d142c66c3df9ad470a7eec13 Mon Sep 17 00:00:00 2001 From: David Howells Date: Mon, 11 May 2020 14:54:34 +0100 Subject: rxrpc: Fix the excessive initial retransmission timeout rxrpc currently uses a fixed 4s retransmission timeout until the RTT is sufficiently sampled. This can cause problems with some fileservers with calls to the cache manager in the afs filesystem being dropped from the fileserver because a packet goes missing and the retransmission timeout is greater than the call expiry timeout. Fix this by: (1) Copying the RTT/RTO calculation code from Linux's TCP implementation and altering it to fit rxrpc. (2) Altering the various users of the RTT to make use of the new SRTT value. (3) Replacing the use of rxrpc_resend_timeout to use the calculated RTO value instead (which is needed in jiffies), along with a backoff. Notes: (1) rxrpc provides RTT samples by matching the serial numbers on outgoing DATA packets that have the RXRPC_REQUEST_ACK set and PING ACK packets against the reference serial number in incoming REQUESTED ACK and PING-RESPONSE ACK packets. (2) Each packet that is transmitted on an rxrpc connection gets a new per-connection serial number, even for retransmissions, so an ACK can be cross-referenced to a specific trigger packet. This allows RTT information to be drawn from retransmitted DATA packets also. (3) rxrpc maintains the RTT/RTO state on the rxrpc_peer record rather than on an rxrpc_call because many RPC calls won't live long enough to generate more than one sample. (4) The calculated SRTT value is in units of 8ths of a microsecond rather than nanoseconds. The (S)RTT and RTO values are displayed in /proc/net/rxrpc/peers. Fixes: 17926a79320a ([AF_RXRPC]: Provide secure RxRPC sockets for use by userspace and kernel both"") Signed-off-by: David Howells --- include/trace/events/rxrpc.h | 17 +++++++---------- 1 file changed, 7 insertions(+), 10 deletions(-) (limited to 'include/trace') diff --git a/include/trace/events/rxrpc.h b/include/trace/events/rxrpc.h index 191fe447f990..ab75f261f04a 100644 --- a/include/trace/events/rxrpc.h +++ b/include/trace/events/rxrpc.h @@ -1112,18 +1112,17 @@ TRACE_EVENT(rxrpc_rtt_tx, TRACE_EVENT(rxrpc_rtt_rx, TP_PROTO(struct rxrpc_call *call, enum rxrpc_rtt_rx_trace why, rxrpc_serial_t send_serial, rxrpc_serial_t resp_serial, - s64 rtt, u8 nr, s64 avg), + u32 rtt, u32 rto), - TP_ARGS(call, why, send_serial, resp_serial, rtt, nr, avg), + TP_ARGS(call, why, send_serial, resp_serial, rtt, rto), TP_STRUCT__entry( __field(unsigned int, call ) __field(enum rxrpc_rtt_rx_trace, why ) - __field(u8, nr ) __field(rxrpc_serial_t, send_serial ) __field(rxrpc_serial_t, resp_serial ) - __field(s64, rtt ) - __field(u64, avg ) + __field(u32, rtt ) + __field(u32, rto ) ), TP_fast_assign( @@ -1132,18 +1131,16 @@ TRACE_EVENT(rxrpc_rtt_rx, __entry->send_serial = send_serial; __entry->resp_serial = resp_serial; __entry->rtt = rtt; - __entry->nr = nr; - __entry->avg = avg; + __entry->rto = rto; ), - TP_printk("c=%08x %s sr=%08x rr=%08x rtt=%lld nr=%u avg=%lld", + TP_printk("c=%08x %s sr=%08x rr=%08x rtt=%u rto=%u", __entry->call, __print_symbolic(__entry->why, rxrpc_rtt_rx_traces), __entry->send_serial, __entry->resp_serial, __entry->rtt, - __entry->nr, - __entry->avg) + __entry->rto) ); TRACE_EVENT(rxrpc_timer, -- cgit v1.2.3 From d1f129470e6cb79b8b97fecd12689f6eb49e27fe Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 28 Apr 2020 22:06:54 +0100 Subject: rxrpc: Trace discarded ACKs Add a tracepoint to track received ACKs that are discarded due to being outside of the Tx window. Signed-off-by: David Howells --- include/trace/events/rxrpc.h | 35 +++++++++++++++++++++++++++++++++++ net/rxrpc/input.c | 12 ++++++++++-- 2 files changed, 45 insertions(+), 2 deletions(-) (limited to 'include/trace') diff --git a/include/trace/events/rxrpc.h b/include/trace/events/rxrpc.h index ab75f261f04a..ba9efdc848f9 100644 --- a/include/trace/events/rxrpc.h +++ b/include/trace/events/rxrpc.h @@ -1541,6 +1541,41 @@ TRACE_EVENT(rxrpc_notify_socket, __entry->serial) ); +TRACE_EVENT(rxrpc_rx_discard_ack, + TP_PROTO(unsigned int debug_id, rxrpc_serial_t serial, + rxrpc_seq_t first_soft_ack, rxrpc_seq_t call_ackr_first, + rxrpc_seq_t prev_pkt, rxrpc_seq_t call_ackr_prev), + + TP_ARGS(debug_id, serial, first_soft_ack, call_ackr_first, + prev_pkt, call_ackr_prev), + + TP_STRUCT__entry( + __field(unsigned int, debug_id ) + __field(rxrpc_serial_t, serial ) + __field(rxrpc_seq_t, first_soft_ack) + __field(rxrpc_seq_t, call_ackr_first) + __field(rxrpc_seq_t, prev_pkt) + __field(rxrpc_seq_t, call_ackr_prev) + ), + + TP_fast_assign( + __entry->debug_id = debug_id; + __entry->serial = serial; + __entry->first_soft_ack = first_soft_ack; + __entry->call_ackr_first = call_ackr_first; + __entry->prev_pkt = prev_pkt; + __entry->call_ackr_prev = call_ackr_prev; + ), + + TP_printk("c=%08x r=%08x %08x<%08x %08x<%08x", + __entry->debug_id, + __entry->serial, + __entry->first_soft_ack, + __entry->call_ackr_first, + __entry->prev_pkt, + __entry->call_ackr_prev) + ); + #endif /* _TRACE_RXRPC_H */ /* This part must be outside protection */ diff --git a/net/rxrpc/input.c b/net/rxrpc/input.c index e438bfd3fdf5..2f22f082a66c 100644 --- a/net/rxrpc/input.c +++ b/net/rxrpc/input.c @@ -866,8 +866,12 @@ static void rxrpc_input_ack(struct rxrpc_call *call, struct sk_buff *skb) /* Discard any out-of-order or duplicate ACKs (outside lock). */ if (before(first_soft_ack, call->ackr_first_seq) || - before(prev_pkt, call->ackr_prev_seq)) + before(prev_pkt, call->ackr_prev_seq)) { + trace_rxrpc_rx_discard_ack(call->debug_id, sp->hdr.serial, + first_soft_ack, call->ackr_first_seq, + prev_pkt, call->ackr_prev_seq); return; + } buf.info.rxMTU = 0; ioffset = offset + nr_acks + 3; @@ -879,8 +883,12 @@ static void rxrpc_input_ack(struct rxrpc_call *call, struct sk_buff *skb) /* Discard any out-of-order or duplicate ACKs (inside lock). */ if (before(first_soft_ack, call->ackr_first_seq) || - before(prev_pkt, call->ackr_prev_seq)) + before(prev_pkt, call->ackr_prev_seq)) { + trace_rxrpc_rx_discard_ack(call->debug_id, sp->hdr.serial, + first_soft_ack, call->ackr_first_seq, + prev_pkt, call->ackr_prev_seq); goto out; + } call->acks_latest_ts = skb->tstamp; call->ackr_first_seq = first_soft_ack; -- cgit v1.2.3