[linux-block.git] / net / rxrpc / call_event.c

// SPDX-License-Identifier: GPL-2.0-or-later
/* Management of Tx window, Tx resend, ACKs and out-of-sequence reception
 *
 * Copyright (C) 2007 Red Hat, Inc. All Rights Reserved.
 * Written by David Howells (dhowells@redhat.com)
 */

#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

#include <linux/module.h>
#include <linux/circ_buf.h>
#include <linux/net.h>
#include <linux/skbuff.h>
#include <linux/slab.h>
#include <linux/udp.h>
#include <net/sock.h>
#include <net/af_rxrpc.h>
#include "ar-internal.h"

/*
 * Propose a PING ACK be sent.
 */
void rxrpc_propose_ping(struct rxrpc_call *call, u32 serial,
			enum rxrpc_propose_ack_trace why)
{
	unsigned long now = jiffies;
	unsigned long ping_at = now + rxrpc_idle_ack_delay;

	if (time_before(ping_at, call->ping_at)) {
		WRITE_ONCE(call->ping_at, ping_at);
		rxrpc_reduce_call_timer(call, ping_at, now,
					rxrpc_timer_set_for_ping);
		trace_rxrpc_propose_ack(call, why, RXRPC_ACK_PING, serial);
	}
}

/*
 * Propose a DELAY ACK be sent in the future.
 */
void rxrpc_propose_delay_ACK(struct rxrpc_call *call, rxrpc_serial_t serial,
			     enum rxrpc_propose_ack_trace why)
{
	unsigned long expiry = rxrpc_soft_ack_delay;
	unsigned long now = jiffies, ack_at;

	call->ackr_serial = serial;

	if (rxrpc_soft_ack_delay < expiry)
		expiry = rxrpc_soft_ack_delay;
	if (call->peer->srtt_us != 0)
		ack_at = usecs_to_jiffies(call->peer->srtt_us >> 3);
	else
		ack_at = expiry;

	ack_at += READ_ONCE(call->tx_backoff);
	ack_at += now;
	if (time_before(ack_at, call->delay_ack_at)) {
		WRITE_ONCE(call->delay_ack_at, ack_at);
		rxrpc_reduce_call_timer(call, ack_at, now,
					rxrpc_timer_set_for_ack);
	}

	trace_rxrpc_propose_ack(call, why, RXRPC_ACK_DELAY, serial);
}

/*
 * Queue an ACK for immediate transmission.
 */
void rxrpc_send_ACK(struct rxrpc_call *call, u8 ack_reason,
		    rxrpc_serial_t serial, enum rxrpc_propose_ack_trace why)
{
	struct rxrpc_txbuf *txb;

	if (test_bit(RXRPC_CALL_DISCONNECTED, &call->flags))
		return;

	rxrpc_inc_stat(call->rxnet, stat_tx_acks[ack_reason]);

	txb = rxrpc_alloc_txbuf(call, RXRPC_PACKET_TYPE_ACK,
				rcu_read_lock_held() ? GFP_ATOMIC | __GFP_NOWARN : GFP_NOFS);
	if (!txb) {
		kleave(" = -ENOMEM");
		return;
	}

	txb->ack_why		= why;
	txb->wire.seq		= 0;
	txb->wire.type		= RXRPC_PACKET_TYPE_ACK;
	txb->wire.flags		|= RXRPC_SLOW_START_OK;
	txb->ack.bufferSpace	= 0;
	txb->ack.maxSkew	= 0;
	txb->ack.firstPacket	= 0;
	txb->ack.previousPacket	= 0;
	txb->ack.serial		= htonl(serial);
	txb->ack.reason		= ack_reason;
	txb->ack.nAcks		= 0;

	trace_rxrpc_send_ack(call, why, ack_reason, serial);
	rxrpc_send_ack_packet(call, txb);
	rxrpc_put_txbuf(txb, rxrpc_txbuf_put_ack_tx);
}

/*
 * Handle congestion being detected by the retransmit timeout.
 */
static void rxrpc_congestion_timeout(struct rxrpc_call *call)
{
	set_bit(RXRPC_CALL_RETRANS_TIMEOUT, &call->flags);
}

/*
 * Perform retransmission of NAK'd and unack'd packets.
 */
void rxrpc_resend(struct rxrpc_call *call, struct sk_buff *ack_skb)
{
	struct rxrpc_ackpacket *ack = NULL;
	struct rxrpc_txbuf *txb;
	unsigned long resend_at;
	rxrpc_seq_t transmitted = READ_ONCE(call->tx_transmitted);
	ktime_t now, max_age, oldest, ack_ts;
	bool unacked = false;
	unsigned int i;
	LIST_HEAD(retrans_queue);

	_enter("{%d,%d}", call->acks_hard_ack, call->tx_top);

	now = ktime_get_real();
	max_age = ktime_sub_us(now, jiffies_to_usecs(call->peer->rto_j));
	oldest = now;

	if (list_empty(&call->tx_buffer))
		goto no_resend;

	if (list_empty(&call->tx_buffer))
		goto no_further_resend;

	trace_rxrpc_resend(call, ack_skb);
	txb = list_first_entry(&call->tx_buffer, struct rxrpc_txbuf, call_link);

	/* Scan the soft ACK table without dropping the lock and resend any
	 * explicitly NAK'd packets.
	 */
	if (ack_skb) {
		ack = (void *)ack_skb->data + sizeof(struct rxrpc_wire_header);

		for (i = 0; i < ack->nAcks; i++) {
			rxrpc_seq_t seq;

			if (ack->acks[i] & 1)
				continue;
			seq = ntohl(ack->firstPacket) + i;
			if (after(txb->seq, transmitted))
				break;
			if (after(txb->seq, seq))
				continue; /* A new hard ACK probably came in */
			list_for_each_entry_from(txb, &call->tx_buffer, call_link) {
				if (txb->seq == seq)
					goto found_txb;
			}
			goto no_further_resend;

		found_txb:
			if (after(ntohl(txb->wire.serial), call->acks_highest_serial))
				continue; /* Ack point not yet reached */

			rxrpc_see_txbuf(txb, rxrpc_txbuf_see_unacked);

			if (list_empty(&txb->tx_link)) {
				list_add_tail(&txb->tx_link, &retrans_queue);
				set_bit(RXRPC_TXBUF_RESENT, &txb->flags);
			}

			trace_rxrpc_retransmit(call, txb->seq,
					       ktime_to_ns(ktime_sub(txb->last_sent,
								     max_age)));

			if (list_is_last(&txb->call_link, &call->tx_buffer))
				goto no_further_resend;
			txb = list_next_entry(txb, call_link);
		}
	}

	/* Fast-forward through the Tx queue to the point the peer says it has
	 * seen.  Anything between the soft-ACK table and that point will get
	 * ACK'd or NACK'd in due course, so don't worry about it here; here we
	 * need to consider retransmitting anything beyond that point.
	 *
	 * Note that ACK for a packet can beat the update of tx_transmitted.
	 */
	if (after_eq(READ_ONCE(call->acks_prev_seq), READ_ONCE(call->tx_transmitted)))
		goto no_further_resend;

	list_for_each_entry_from(txb, &call->tx_buffer, call_link) {
		if (before_eq(txb->seq, READ_ONCE(call->acks_prev_seq)))
			continue;
		if (after(txb->seq, READ_ONCE(call->tx_transmitted)))
			break; /* Not transmitted yet */

		if (ack && ack->reason == RXRPC_ACK_PING_RESPONSE &&
		    before(ntohl(txb->wire.serial), ntohl(ack->serial)))
			goto do_resend; /* Wasn't accounted for by a more recent ping. */

		if (ktime_after(txb->last_sent, max_age)) {
			if (ktime_before(txb->last_sent, oldest))
				oldest = txb->last_sent;
			continue;
		}

	do_resend:
		unacked = true;
		if (list_empty(&txb->tx_link)) {
			list_add_tail(&txb->tx_link, &retrans_queue);
			set_bit(RXRPC_TXBUF_RESENT, &txb->flags);
			rxrpc_inc_stat(call->rxnet, stat_tx_data_retrans);
		}
	}

no_further_resend:
no_resend:
	resend_at = nsecs_to_jiffies(ktime_to_ns(ktime_sub(now, oldest)));
	resend_at += jiffies + rxrpc_get_rto_backoff(call->peer,
						     !list_empty(&retrans_queue));
	WRITE_ONCE(call->resend_at, resend_at);

	if (unacked)
		rxrpc_congestion_timeout(call);

	/* If there was nothing that needed retransmission then it's likely
	 * that an ACK got lost somewhere.  Send a ping to find out instead of
	 * retransmitting data.
	 */
	if (list_empty(&retrans_queue)) {
		rxrpc_reduce_call_timer(call, resend_at, jiffies,
					rxrpc_timer_set_for_resend);
		ack_ts = ktime_sub(now, call->acks_latest_ts);
		if (ktime_to_us(ack_ts) < (call->peer->srtt_us >> 3))
			goto out;
		rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
			       rxrpc_propose_ack_ping_for_lost_ack);
		goto out;
	}

	/* Retransmit the queue */
	while ((txb = list_first_entry_or_null(&retrans_queue,
					       struct rxrpc_txbuf, tx_link))) {
		list_del_init(&txb->tx_link);
		rxrpc_transmit_one(call, txb);
	}

out:
	_leave("");
}

/*
 * Start transmitting the reply to a service.  This cancels the need to ACK the
 * request if we haven't yet done so.
 */
static void rxrpc_begin_service_reply(struct rxrpc_call *call)
{
	unsigned long now = jiffies;

	rxrpc_set_call_state(call, RXRPC_CALL_SERVER_SEND_REPLY);
	WRITE_ONCE(call->delay_ack_at, now + MAX_JIFFY_OFFSET);
	if (call->ackr_reason == RXRPC_ACK_DELAY)
		call->ackr_reason = 0;
	trace_rxrpc_timer(call, rxrpc_timer_init_for_send_reply, now);
}

/*
 * Close the transmission phase.  After this point there is no more data to be
 * transmitted in the call.
 */
static void rxrpc_close_tx_phase(struct rxrpc_call *call)
{
	_debug("________awaiting reply/ACK__________");

	switch (__rxrpc_call_state(call)) {
	case RXRPC_CALL_CLIENT_SEND_REQUEST:
		rxrpc_set_call_state(call, RXRPC_CALL_CLIENT_AWAIT_REPLY);
		break;
	case RXRPC_CALL_SERVER_SEND_REPLY:
		rxrpc_set_call_state(call, RXRPC_CALL_SERVER_AWAIT_ACK);
		break;
	default:
		break;
	}
}

static bool rxrpc_tx_window_has_space(struct rxrpc_call *call)
{
	unsigned int winsize = min_t(unsigned int, call->tx_winsize,
				     call->cong_cwnd + call->cong_extra);
	rxrpc_seq_t window = call->acks_hard_ack, wtop = window + winsize;
	rxrpc_seq_t tx_top = call->tx_top;
	int space;

	space = wtop - tx_top;
	return space > 0;
}

/*
 * Decant some if the sendmsg prepared queue into the transmission buffer.
 */
static void rxrpc_decant_prepared_tx(struct rxrpc_call *call)
{
	struct rxrpc_txbuf *txb;

	if (!test_bit(RXRPC_CALL_EXPOSED, &call->flags)) {
		if (list_empty(&call->tx_sendmsg))
			return;
		rxrpc_expose_client_call(call);
	}

	while ((txb = list_first_entry_or_null(&call->tx_sendmsg,
					       struct rxrpc_txbuf, call_link))) {
		spin_lock(&call->tx_lock);
		list_del(&txb->call_link);
		spin_unlock(&call->tx_lock);

		call->tx_top = txb->seq;
		list_add_tail(&txb->call_link, &call->tx_buffer);

		if (txb->wire.flags & RXRPC_LAST_PACKET)
			rxrpc_close_tx_phase(call);

		rxrpc_transmit_one(call, txb);

		if (!rxrpc_tx_window_has_space(call))
			break;
	}
}

static void rxrpc_transmit_some_data(struct rxrpc_call *call)
{
	switch (__rxrpc_call_state(call)) {
	case RXRPC_CALL_SERVER_ACK_REQUEST:
		if (list_empty(&call->tx_sendmsg))
			return;
		rxrpc_begin_service_reply(call);
		fallthrough;

	case RXRPC_CALL_SERVER_SEND_REPLY:
	case RXRPC_CALL_CLIENT_SEND_REQUEST:
		if (!rxrpc_tx_window_has_space(call))
			return;
		if (list_empty(&call->tx_sendmsg)) {
			rxrpc_inc_stat(call->rxnet, stat_tx_data_underflow);
			return;
		}
		rxrpc_decant_prepared_tx(call);
		break;
	default:
		return;
	}
}

/*
 * Ping the other end to fill our RTT cache and to retrieve the rwind
 * and MTU parameters.
 */
static void rxrpc_send_initial_ping(struct rxrpc_call *call)
{
	if (call->peer->rtt_count < 3 ||
	    ktime_before(ktime_add_ms(call->peer->rtt_last_req, 1000),
			 ktime_get_real()))
		rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
			       rxrpc_propose_ack_ping_for_params);
}

/*
 * Handle retransmission and deferred ACK/abort generation.
 */
bool rxrpc_input_call_event(struct rxrpc_call *call, struct sk_buff *skb)
{
	unsigned long now, next, t;
	rxrpc_serial_t ackr_serial;
	bool resend = false, expired = false;
	s32 abort_code;

	rxrpc_see_call(call, rxrpc_call_see_input);

	//printk("\n--------------------\n");
	_enter("{%d,%s,%lx}",
	       call->debug_id, rxrpc_call_states[__rxrpc_call_state(call)],
	       call->events);

	if (__rxrpc_call_is_complete(call))
		goto out;

	/* Handle abort request locklessly, vs rxrpc_propose_abort(). */
	abort_code = smp_load_acquire(&call->send_abort);
	if (abort_code) {
		rxrpc_abort_call(call, 0, call->send_abort, call->send_abort_err,
				 call->send_abort_why);
		goto out;
	}

	if (skb && skb->mark == RXRPC_SKB_MARK_ERROR)
		goto out;

	/* If we see our async-event poke, check for timeout trippage. */
	now = jiffies;
	t = READ_ONCE(call->expect_rx_by);
	if (time_after_eq(now, t)) {
		trace_rxrpc_timer(call, rxrpc_timer_exp_normal, now);
		expired = true;
	}

	t = READ_ONCE(call->expect_req_by);
	if (__rxrpc_call_state(call) == RXRPC_CALL_SERVER_RECV_REQUEST &&
	    time_after_eq(now, t)) {
		trace_rxrpc_timer(call, rxrpc_timer_exp_idle, now);
		expired = true;
	}

	t = READ_ONCE(call->expect_term_by);
	if (time_after_eq(now, t)) {
		trace_rxrpc_timer(call, rxrpc_timer_exp_hard, now);
		expired = true;
	}

	t = READ_ONCE(call->delay_ack_at);
	if (time_after_eq(now, t)) {
		trace_rxrpc_timer(call, rxrpc_timer_exp_ack, now);
		cmpxchg(&call->delay_ack_at, t, now + MAX_JIFFY_OFFSET);
		ackr_serial = xchg(&call->ackr_serial, 0);
		rxrpc_send_ACK(call, RXRPC_ACK_DELAY, ackr_serial,
			       rxrpc_propose_ack_ping_for_lost_ack);
	}

	t = READ_ONCE(call->ack_lost_at);
	if (time_after_eq(now, t)) {
		trace_rxrpc_timer(call, rxrpc_timer_exp_lost_ack, now);
		cmpxchg(&call->ack_lost_at, t, now + MAX_JIFFY_OFFSET);
		set_bit(RXRPC_CALL_EV_ACK_LOST, &call->events);
	}

	t = READ_ONCE(call->keepalive_at);
	if (time_after_eq(now, t)) {
		trace_rxrpc_timer(call, rxrpc_timer_exp_keepalive, now);
		cmpxchg(&call->keepalive_at, t, now + MAX_JIFFY_OFFSET);
		rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
			       rxrpc_propose_ack_ping_for_keepalive);
	}

	t = READ_ONCE(call->ping_at);
	if (time_after_eq(now, t)) {
		trace_rxrpc_timer(call, rxrpc_timer_exp_ping, now);
		cmpxchg(&call->ping_at, t, now + MAX_JIFFY_OFFSET);
		rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
			       rxrpc_propose_ack_ping_for_keepalive);
	}

	t = READ_ONCE(call->resend_at);
	if (time_after_eq(now, t)) {
		trace_rxrpc_timer(call, rxrpc_timer_exp_resend, now);
		cmpxchg(&call->resend_at, t, now + MAX_JIFFY_OFFSET);
		resend = true;
	}

	if (skb)
		rxrpc_input_call_packet(call, skb);

	rxrpc_transmit_some_data(call);

	if (skb) {
		struct rxrpc_skb_priv *sp = rxrpc_skb(skb);

		if (sp->hdr.type == RXRPC_PACKET_TYPE_ACK)
			rxrpc_congestion_degrade(call);
	}

	if (test_and_clear_bit(RXRPC_CALL_EV_INITIAL_PING, &call->events))
		rxrpc_send_initial_ping(call);

	/* Process events */
	if (expired) {
		if (test_bit(RXRPC_CALL_RX_HEARD, &call->flags) &&
		    (int)call->conn->hi_serial - (int)call->rx_serial > 0) {
			trace_rxrpc_call_reset(call);
			rxrpc_abort_call(call, 0, RX_CALL_DEAD, -ECONNRESET,
					 rxrpc_abort_call_reset);
		} else {
			rxrpc_abort_call(call, 0, RX_CALL_TIMEOUT, -ETIME,
					 rxrpc_abort_call_timeout);
		}
		goto out;
	}

	if (test_and_clear_bit(RXRPC_CALL_EV_ACK_LOST, &call->events))
		rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
			       rxrpc_propose_ack_ping_for_lost_ack);

	if (resend && __rxrpc_call_state(call) != RXRPC_CALL_CLIENT_RECV_REPLY)
		rxrpc_resend(call, NULL);

	if (test_and_clear_bit(RXRPC_CALL_RX_IS_IDLE, &call->flags))
		rxrpc_send_ACK(call, RXRPC_ACK_IDLE, 0,
			       rxrpc_propose_ack_rx_idle);

	if (call->ackr_nr_unacked > 2) {
		if (call->peer->rtt_count < 3)
			rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
				       rxrpc_propose_ack_ping_for_rtt);
		else if (ktime_before(ktime_add_ms(call->peer->rtt_last_req, 1000),
				      ktime_get_real()))
			rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
				       rxrpc_propose_ack_ping_for_old_rtt);
		else
			rxrpc_send_ACK(call, RXRPC_ACK_IDLE, 0,
				       rxrpc_propose_ack_input_data);
	}

	/* Make sure the timer is restarted */
	if (!__rxrpc_call_is_complete(call)) {
		next = call->expect_rx_by;

#define set(T) { t = READ_ONCE(T); if (time_before(t, next)) next = t; }

		set(call->expect_req_by);
		set(call->expect_term_by);
		set(call->delay_ack_at);
		set(call->ack_lost_at);
		set(call->resend_at);
		set(call->keepalive_at);
		set(call->ping_at);

		now = jiffies;
		if (time_after_eq(now, next))
			rxrpc_poke_call(call, rxrpc_call_poke_timer_now);

		rxrpc_reduce_call_timer(call, next, now, rxrpc_timer_restart);
	}

out:
	if (__rxrpc_call_is_complete(call)) {
		del_timer_sync(&call->timer);
		if (!test_bit(RXRPC_CALL_DISCONNECTED, &call->flags))
			rxrpc_disconnect_call(call);
		if (call->security)
			call->security->free_call_crypto(call);
	}
	if (call->acks_hard_ack != call->tx_bottom)
		rxrpc_shrink_call_tx_buffer(call);
	_leave("");
	return true;
}
Commit	Line	Data
2874c5fd	1	// SPDX-License-Identifier: GPL-2.0-or-later
17926a79 DH	2	/* Management of Tx window, Tx resend, ACKs and out-of-sequence reception
	3	*
	4	* Copyright (C) 2007 Red Hat, Inc. All Rights Reserved.
	5	* Written by David Howells (dhowells@redhat.com)
17926a79 DH	6	*/
17926a79 DH	7
9b6d5398 JP	8	#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
9b6d5398 JP	9
17926a79 DH	10	#include <linux/module.h>
	11	#include <linux/circ_buf.h>
	12	#include <linux/net.h>
	13	#include <linux/skbuff.h>
5a0e3ad6	14	#include <linux/slab.h>
17926a79 DH	15	#include <linux/udp.h>
	16	#include <net/sock.h>
	17	#include <net/af_rxrpc.h>
	18	#include "ar-internal.h"
	19
a5af7e1f DH	20	/*
	21	* Propose a PING ACK be sent.
	22	*/
72f0c6fb DH	23	void rxrpc_propose_ping(struct rxrpc_call *call, u32 serial,
72f0c6fb DH	24	enum rxrpc_propose_ack_trace why)
a5af7e1f	25	{
72f0c6fb DH	26	unsigned long now = jiffies;
72f0c6fb DH	27	unsigned long ping_at = now + rxrpc_idle_ack_delay;
a5af7e1f	28
72f0c6fb	29	if (time_before(ping_at, call->ping_at)) {
72f0c6fb DH	30	WRITE_ONCE(call->ping_at, ping_at);
	31	rxrpc_reduce_call_timer(call, ping_at, now,
	32	rxrpc_timer_set_for_ping);
530403d9	33	trace_rxrpc_propose_ack(call, why, RXRPC_ACK_PING, serial);
a5af7e1f DH	34	}
	35	}
	36
17926a79	37	/*
530403d9	38	* Propose a DELAY ACK be sent in the future.
17926a79	39	*/
4e76bd40 DH	40	void rxrpc_propose_delay_ACK(struct rxrpc_call *call, rxrpc_serial_t serial,
4e76bd40 DH	41	enum rxrpc_propose_ack_trace why)
17926a79	42	{
beb8e5e4	43	unsigned long expiry = rxrpc_soft_ack_delay;
72f0c6fb	44	unsigned long now = jiffies, ack_at;
17926a79	45
530403d9	46	call->ackr_serial = serial;
72f0c6fb	47
530403d9 DH	48	if (rxrpc_soft_ack_delay < expiry)
530403d9 DH	49	expiry = rxrpc_soft_ack_delay;
72f0c6fb DH	50	if (call->peer->srtt_us != 0)
	51	ack_at = usecs_to_jiffies(call->peer->srtt_us >> 3);
	52	else
	53	ack_at = expiry;
	54
	55	ack_at += READ_ONCE(call->tx_backoff);
	56	ack_at += now;
530403d9 DH	57	if (time_before(ack_at, call->delay_ack_at)) {
530403d9 DH	58	WRITE_ONCE(call->delay_ack_at, ack_at);
72f0c6fb DH	59	rxrpc_reduce_call_timer(call, ack_at, now,
72f0c6fb DH	60	rxrpc_timer_set_for_ack);
17926a79	61	}
9c7ad434	62
530403d9	63	trace_rxrpc_propose_ack(call, why, RXRPC_ACK_DELAY, serial);
17926a79 DH	64	}
17926a79 DH	65
72f0c6fb DH	66	/*
	67	* Queue an ACK for immediate transmission.
	68	*/
	69	void rxrpc_send_ACK(struct rxrpc_call *call, u8 ack_reason,
	70	rxrpc_serial_t serial, enum rxrpc_propose_ack_trace why)
	71	{
72f0c6fb DH	72	struct rxrpc_txbuf *txb;
	73
	74	if (test_bit(RXRPC_CALL_DISCONNECTED, &call->flags))
	75	return;
	76
	77	rxrpc_inc_stat(call->rxnet, stat_tx_acks[ack_reason]);
	78
	79	txb = rxrpc_alloc_txbuf(call, RXRPC_PACKET_TYPE_ACK,
446b3e14	80	rcu_read_lock_held() ? GFP_ATOMIC \| __GFP_NOWARN : GFP_NOFS);
72f0c6fb DH	81	if (!txb) {
	82	kleave(" = -ENOMEM");
	83	return;
	84	}
	85
	86	txb->ack_why = why;
	87	txb->wire.seq = 0;
	88	txb->wire.type = RXRPC_PACKET_TYPE_ACK;
	89	txb->wire.flags \|= RXRPC_SLOW_START_OK;
	90	txb->ack.bufferSpace = 0;
	91	txb->ack.maxSkew = 0;
	92	txb->ack.firstPacket = 0;
	93	txb->ack.previousPacket = 0;
	94	txb->ack.serial = htonl(serial);
	95	txb->ack.reason = ack_reason;
	96	txb->ack.nAcks = 0;
	97
72f0c6fb	98	trace_rxrpc_send_ack(call, why, ack_reason, serial);
b0346843 DH	99	rxrpc_send_ack_packet(call, txb);
b0346843 DH	100	rxrpc_put_txbuf(txb, rxrpc_txbuf_put_ack_tx);
72f0c6fb DH	101	}
72f0c6fb DH	102
57494343 DH	103	/*
	104	* Handle congestion being detected by the retransmit timeout.
	105	*/
	106	static void rxrpc_congestion_timeout(struct rxrpc_call *call)
	107	{
	108	set_bit(RXRPC_CALL_RETRANS_TIMEOUT, &call->flags);
	109	}
	110
17926a79	111	/*
248f219c	112	* Perform retransmission of NAK'd and unack'd packets.
17926a79	113	*/
5e6ef4f1	114	void rxrpc_resend(struct rxrpc_call call, struct sk_buff ack_skb)
17926a79	115	{
d57a3a15	116	struct rxrpc_ackpacket *ack = NULL;
a4ea4c47	117	struct rxrpc_txbuf *txb;
2c13c05c	118	unsigned long resend_at;
a4ea4c47	119	rxrpc_seq_t transmitted = READ_ONCE(call->tx_transmitted);
c410bf01	120	ktime_t now, max_age, oldest, ack_ts;
a4ea4c47	121	bool unacked = false;
d57a3a15	122	unsigned int i;
a4ea4c47	123	LIST_HEAD(retrans_queue);
17926a79	124
a4ea4c47	125	_enter("{%d,%d}", call->acks_hard_ack, call->tx_top);
17926a79	126
a158bdd3	127	now = ktime_get_real();
214a9dc7	128	max_age = ktime_sub_us(now, jiffies_to_usecs(call->peer->rto_j));
d57a3a15 DH	129	oldest = now;
d57a3a15 DH	130
d57a3a15 DH	131	if (list_empty(&call->tx_buffer))
d57a3a15 DH	132	goto no_resend;
50235c4b	133
d57a3a15 DH	134	if (list_empty(&call->tx_buffer))
	135	goto no_further_resend;
	136
5e6ef4f1	137	trace_rxrpc_resend(call, ack_skb);
d57a3a15	138	txb = list_first_entry(&call->tx_buffer, struct rxrpc_txbuf, call_link);
17926a79	139
d57a3a15 DH	140	/* Scan the soft ACK table without dropping the lock and resend any
	141	* explicitly NAK'd packets.
	142	*/
5e6ef4f1 DH	143	if (ack_skb) {
	144	ack = (void *)ack_skb->data + sizeof(struct rxrpc_wire_header);
	145
d57a3a15 DH	146	for (i = 0; i < ack->nAcks; i++) {
d57a3a15 DH	147	rxrpc_seq_t seq;
17926a79	148
d57a3a15	149	if (ack->acks[i] & 1)
248f219c	150	continue;
d57a3a15 DH	151	seq = ntohl(ack->firstPacket) + i;
	152	if (after(txb->seq, transmitted))
	153	break;
	154	if (after(txb->seq, seq))
	155	continue; /* A new hard ACK probably came in */
	156	list_for_each_entry_from(txb, &call->tx_buffer, call_link) {
	157	if (txb->seq == seq)
	158	goto found_txb;
	159	}
	160	goto no_further_resend;
	161
	162	found_txb:
	163	if (after(ntohl(txb->wire.serial), call->acks_highest_serial))
	164	continue; /* Ack point not yet reached */
	165
	166	rxrpc_see_txbuf(txb, rxrpc_txbuf_see_unacked);
	167
	168	if (list_empty(&txb->tx_link)) {
d57a3a15 DH	169	list_add_tail(&txb->tx_link, &retrans_queue);
d57a3a15 DH	170	set_bit(RXRPC_TXBUF_RESENT, &txb->flags);
33c40e24	171	}
d57a3a15 DH	172
	173	trace_rxrpc_retransmit(call, txb->seq,
	174	ktime_to_ns(ktime_sub(txb->last_sent,
	175	max_age)));
	176
	177	if (list_is_last(&txb->call_link, &call->tx_buffer))
	178	goto no_further_resend;
	179	txb = list_next_entry(txb, call_link);
17926a79	180	}
d57a3a15	181	}
17926a79	182
d57a3a15 DH	183	/* Fast-forward through the Tx queue to the point the peer says it has
	184	* seen. Anything between the soft-ACK table and that point will get
	185	* ACK'd or NACK'd in due course, so don't worry about it here; here we
	186	* need to consider retransmitting anything beyond that point.
	187	*
	188	* Note that ACK for a packet can beat the update of tx_transmitted.
	189	*/
	190	if (after_eq(READ_ONCE(call->acks_prev_seq), READ_ONCE(call->tx_transmitted)))
	191	goto no_further_resend;
	192
	193	list_for_each_entry_from(txb, &call->tx_buffer, call_link) {
	194	if (before_eq(txb->seq, READ_ONCE(call->acks_prev_seq)))
	195	continue;
	196	if (after(txb->seq, READ_ONCE(call->tx_transmitted)))
	197	break; /* Not transmitted yet */
	198
	199	if (ack && ack->reason == RXRPC_ACK_PING_RESPONSE &&
	200	before(ntohl(txb->wire.serial), ntohl(ack->serial)))
	201	goto do_resend; /* Wasn't accounted for by a more recent ping. */
	202
	203	if (ktime_after(txb->last_sent, max_age)) {
	204	if (ktime_before(txb->last_sent, oldest))
	205	oldest = txb->last_sent;
	206	continue;
	207	}
	208
	209	do_resend:
	210	unacked = true;
	211	if (list_empty(&txb->tx_link)) {
d57a3a15 DH	212	list_add_tail(&txb->tx_link, &retrans_queue);
	213	set_bit(RXRPC_TXBUF_RESENT, &txb->flags);
	214	rxrpc_inc_stat(call->rxnet, stat_tx_data_retrans);
	215	}
dfa7d920	216	}
248f219c	217
d57a3a15	218	no_further_resend:
d57a3a15	219	no_resend:
59299aa1	220	resend_at = nsecs_to_jiffies(ktime_to_ns(ktime_sub(now, oldest)));
a4ea4c47 DH	221	resend_at += jiffies + rxrpc_get_rto_backoff(call->peer,
a4ea4c47 DH	222	!list_empty(&retrans_queue));
a158bdd3	223	WRITE_ONCE(call->resend_at, resend_at);
248f219c	224
57494343 DH	225	if (unacked)
	226	rxrpc_congestion_timeout(call);
	227
	228	/* If there was nothing that needed retransmission then it's likely
	229	* that an ACK got lost somewhere. Send a ping to find out instead of
	230	* retransmitting data.
	231	*/
a4ea4c47	232	if (list_empty(&retrans_queue)) {
5e6ef4f1	233	rxrpc_reduce_call_timer(call, resend_at, jiffies,
a158bdd3	234	rxrpc_timer_set_for_resend);
57494343	235	ack_ts = ktime_sub(now, call->acks_latest_ts);
c410bf01	236	if (ktime_to_us(ack_ts) < (call->peer->srtt_us >> 3))
57494343	237	goto out;
72f0c6fb DH	238	rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
72f0c6fb DH	239	rxrpc_propose_ack_ping_for_lost_ack);
57494343 DH	240	goto out;
	241	}
	242
5e6ef4f1	243	/* Retransmit the queue */
a4ea4c47 DH	244	while ((txb = list_first_entry_or_null(&retrans_queue,
	245	struct rxrpc_txbuf, tx_link))) {
	246	list_del_init(&txb->tx_link);
5e6ef4f1	247	rxrpc_transmit_one(call, txb);
dfa7d920	248	}
248f219c	249
57494343	250	out:
248f219c	251	_leave("");
17926a79 DH	252	}
17926a79 DH	253
2d689424 DH	254	/*
	255	* Start transmitting the reply to a service. This cancels the need to ACK the
	256	* request if we haven't yet done so.
	257	*/
	258	static void rxrpc_begin_service_reply(struct rxrpc_call *call)
	259	{
96b4059f	260	unsigned long now = jiffies;
2d689424	261
96b4059f DH	262	rxrpc_set_call_state(call, RXRPC_CALL_SERVER_SEND_REPLY);
	263	WRITE_ONCE(call->delay_ack_at, now + MAX_JIFFY_OFFSET);
	264	if (call->ackr_reason == RXRPC_ACK_DELAY)
	265	call->ackr_reason = 0;
	266	trace_rxrpc_timer(call, rxrpc_timer_init_for_send_reply, now);
2d689424 DH	267	}
	268
	269	/*
	270	* Close the transmission phase. After this point there is no more data to be
	271	* transmitted in the call.
	272	*/
	273	static void rxrpc_close_tx_phase(struct rxrpc_call *call)
	274	{
	275	_debug("________awaiting reply/ACK__________");
	276
96b4059f	277	switch (__rxrpc_call_state(call)) {
2d689424	278	case RXRPC_CALL_CLIENT_SEND_REQUEST:
96b4059f	279	rxrpc_set_call_state(call, RXRPC_CALL_CLIENT_AWAIT_REPLY);
2d689424 DH	280	break;
2d689424 DH	281	case RXRPC_CALL_SERVER_SEND_REPLY:
96b4059f	282	rxrpc_set_call_state(call, RXRPC_CALL_SERVER_AWAIT_ACK);
2d689424 DH	283	break;
	284	default:
	285	break;
	286	}
2d689424 DH	287	}
2d689424 DH	288
cf37b598 DH	289	static bool rxrpc_tx_window_has_space(struct rxrpc_call *call)
	290	{
	291	unsigned int winsize = min_t(unsigned int, call->tx_winsize,
	292	call->cong_cwnd + call->cong_extra);
	293	rxrpc_seq_t window = call->acks_hard_ack, wtop = window + winsize;
	294	rxrpc_seq_t tx_top = call->tx_top;
	295	int space;
	296
	297	space = wtop - tx_top;
	298	return space > 0;
	299	}
	300
	301	/*
	302	* Decant some if the sendmsg prepared queue into the transmission buffer.
	303	*/
	304	static void rxrpc_decant_prepared_tx(struct rxrpc_call *call)
	305	{
	306	struct rxrpc_txbuf *txb;
	307
a343b174 DH	308	if (!test_bit(RXRPC_CALL_EXPOSED, &call->flags)) {
	309	if (list_empty(&call->tx_sendmsg))
	310	return;
cf37b598	311	rxrpc_expose_client_call(call);
a343b174	312	}
cf37b598 DH	313
	314	while ((txb = list_first_entry_or_null(&call->tx_sendmsg,
	315	struct rxrpc_txbuf, call_link))) {
	316	spin_lock(&call->tx_lock);
	317	list_del(&txb->call_link);
	318	spin_unlock(&call->tx_lock);
	319
	320	call->tx_top = txb->seq;
	321	list_add_tail(&txb->call_link, &call->tx_buffer);
	322
2d689424 DH	323	if (txb->wire.flags & RXRPC_LAST_PACKET)
	324	rxrpc_close_tx_phase(call);
	325
cf37b598 DH	326	rxrpc_transmit_one(call, txb);
cf37b598 DH	327
cf37b598 DH	328	if (!rxrpc_tx_window_has_space(call))
	329	break;
	330	}
	331	}
	332
	333	static void rxrpc_transmit_some_data(struct rxrpc_call *call)
	334	{
96b4059f	335	switch (__rxrpc_call_state(call)) {
cf37b598 DH	336	case RXRPC_CALL_SERVER_ACK_REQUEST:
	337	if (list_empty(&call->tx_sendmsg))
	338	return;
2d689424	339	rxrpc_begin_service_reply(call);
cf37b598 DH	340	fallthrough;
	341
	342	case RXRPC_CALL_SERVER_SEND_REPLY:
cf37b598	343	case RXRPC_CALL_CLIENT_SEND_REQUEST:
cf37b598 DH	344	if (!rxrpc_tx_window_has_space(call))
cf37b598 DH	345	return;
32cf8edb DH	346	if (list_empty(&call->tx_sendmsg)) {
32cf8edb DH	347	rxrpc_inc_stat(call->rxnet, stat_tx_data_underflow);
cf37b598	348	return;
32cf8edb	349	}
cf37b598 DH	350	rxrpc_decant_prepared_tx(call);
	351	break;
	352	default:
	353	return;
	354	}
	355	}
	356
5e6ef4f1 DH	357	/*
	358	* Ping the other end to fill our RTT cache and to retrieve the rwind
	359	* and MTU parameters.
	360	*/
	361	static void rxrpc_send_initial_ping(struct rxrpc_call *call)
	362	{
	363	if (call->peer->rtt_count < 3 \|\|
	364	ktime_before(ktime_add_ms(call->peer->rtt_last_req, 1000),
	365	ktime_get_real()))
	366	rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
	367	rxrpc_propose_ack_ping_for_params);
	368	}
	369
17926a79	370	/*
248f219c	371	* Handle retransmission and deferred ACK/abort generation.
17926a79	372	*/
57af281e	373	bool rxrpc_input_call_event(struct rxrpc_call call, struct sk_buff skb)
17926a79	374	{
a158bdd3	375	unsigned long now, next, t;
72f0c6fb	376	rxrpc_serial_t ackr_serial;
5e6ef4f1	377	bool resend = false, expired = false;
a343b174	378	s32 abort_code;
17926a79	379
cb0fc0c9	380	rxrpc_see_call(call, rxrpc_call_see_input);
e34d4234	381
17926a79	382	//printk("\n--------------------\n");
248f219c	383	_enter("{%d,%s,%lx}",
96b4059f DH	384	call->debug_id, rxrpc_call_states[__rxrpc_call_state(call)],
96b4059f DH	385	call->events);
17926a79	386
96b4059f	387	if (__rxrpc_call_is_complete(call))
5e6ef4f1	388	goto out;
17926a79	389
a343b174 DH	390	/* Handle abort request locklessly, vs rxrpc_propose_abort(). */
	391	abort_code = smp_load_acquire(&call->send_abort);
	392	if (abort_code) {
57af281e DH	393	rxrpc_abort_call(call, 0, call->send_abort, call->send_abort_err,
57af281e DH	394	call->send_abort_why);
a343b174 DH	395	goto out;
	396	}
	397
5e6ef4f1	398	if (skb && skb->mark == RXRPC_SKB_MARK_ERROR)
3feda9d6	399	goto out;
17926a79	400
5e6ef4f1	401	/* If we see our async-event poke, check for timeout trippage. */
a158bdd3 DH	402	now = jiffies;
	403	t = READ_ONCE(call->expect_rx_by);
	404	if (time_after_eq(now, t)) {
	405	trace_rxrpc_timer(call, rxrpc_timer_exp_normal, now);
5e6ef4f1	406	expired = true;
a158bdd3 DH	407	}
	408
	409	t = READ_ONCE(call->expect_req_by);
96b4059f	410	if (__rxrpc_call_state(call) == RXRPC_CALL_SERVER_RECV_REQUEST &&
a158bdd3 DH	411	time_after_eq(now, t)) {
a158bdd3 DH	412	trace_rxrpc_timer(call, rxrpc_timer_exp_idle, now);
5e6ef4f1	413	expired = true;
a158bdd3 DH	414	}
	415
	416	t = READ_ONCE(call->expect_term_by);
	417	if (time_after_eq(now, t)) {
	418	trace_rxrpc_timer(call, rxrpc_timer_exp_hard, now);
5e6ef4f1	419	expired = true;
a158bdd3 DH	420	}
a158bdd3 DH	421
530403d9	422	t = READ_ONCE(call->delay_ack_at);
a158bdd3 DH	423	if (time_after_eq(now, t)) {
a158bdd3 DH	424	trace_rxrpc_timer(call, rxrpc_timer_exp_ack, now);
530403d9 DH	425	cmpxchg(&call->delay_ack_at, t, now + MAX_JIFFY_OFFSET);
	426	ackr_serial = xchg(&call->ackr_serial, 0);
	427	rxrpc_send_ACK(call, RXRPC_ACK_DELAY, ackr_serial,
	428	rxrpc_propose_ack_ping_for_lost_ack);
a158bdd3 DH	429	}
a158bdd3 DH	430
bd1fdf8c DH	431	t = READ_ONCE(call->ack_lost_at);
	432	if (time_after_eq(now, t)) {
	433	trace_rxrpc_timer(call, rxrpc_timer_exp_lost_ack, now);
	434	cmpxchg(&call->ack_lost_at, t, now + MAX_JIFFY_OFFSET);
	435	set_bit(RXRPC_CALL_EV_ACK_LOST, &call->events);
	436	}
	437
415f44e4 DH	438	t = READ_ONCE(call->keepalive_at);
	439	if (time_after_eq(now, t)) {
	440	trace_rxrpc_timer(call, rxrpc_timer_exp_keepalive, now);
	441	cmpxchg(&call->keepalive_at, t, now + MAX_JIFFY_OFFSET);
72f0c6fb DH	442	rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
72f0c6fb DH	443	rxrpc_propose_ack_ping_for_keepalive);
415f44e4 DH	444	}
415f44e4 DH	445
a158bdd3 DH	446	t = READ_ONCE(call->ping_at);
	447	if (time_after_eq(now, t)) {
	448	trace_rxrpc_timer(call, rxrpc_timer_exp_ping, now);
	449	cmpxchg(&call->ping_at, t, now + MAX_JIFFY_OFFSET);
72f0c6fb DH	450	rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
72f0c6fb DH	451	rxrpc_propose_ack_ping_for_keepalive);
a158bdd3 DH	452	}
	453
	454	t = READ_ONCE(call->resend_at);
	455	if (time_after_eq(now, t)) {
	456	trace_rxrpc_timer(call, rxrpc_timer_exp_resend, now);
	457	cmpxchg(&call->resend_at, t, now + MAX_JIFFY_OFFSET);
5e6ef4f1	458	resend = true;
a158bdd3 DH	459	}
a158bdd3 DH	460
5e6ef4f1 DH	461	if (skb)
	462	rxrpc_input_call_packet(call, skb);
	463
cf37b598 DH	464	rxrpc_transmit_some_data(call);
cf37b598 DH	465
5086d9a9 DH	466	if (skb) {
	467	struct rxrpc_skb_priv *sp = rxrpc_skb(skb);
	468
	469	if (sp->hdr.type == RXRPC_PACKET_TYPE_ACK)
	470	rxrpc_congestion_degrade(call);
	471	}
	472
5e6ef4f1 DH	473	if (test_and_clear_bit(RXRPC_CALL_EV_INITIAL_PING, &call->events))
	474	rxrpc_send_initial_ping(call);
	475
a158bdd3	476	/* Process events */
5e6ef4f1	477	if (expired) {
1a025028 DH	478	if (test_bit(RXRPC_CALL_RX_HEARD, &call->flags) &&
	479	(int)call->conn->hi_serial - (int)call->rx_serial > 0) {
	480	trace_rxrpc_call_reset(call);
57af281e DH	481	rxrpc_abort_call(call, 0, RX_CALL_DEAD, -ECONNRESET,
57af281e DH	482	rxrpc_abort_call_reset);
1a025028	483	} else {
57af281e DH	484	rxrpc_abort_call(call, 0, RX_CALL_TIMEOUT, -ETIME,
57af281e DH	485	rxrpc_abort_call_timeout);
1a025028	486	}
5e6ef4f1	487	goto out;
17926a79 DH	488	}
17926a79 DH	489
5e6ef4f1	490	if (test_and_clear_bit(RXRPC_CALL_EV_ACK_LOST, &call->events))
72f0c6fb DH	491	rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
72f0c6fb DH	492	rxrpc_propose_ack_ping_for_lost_ack);
a5af7e1f	493
96b4059f	494	if (resend && __rxrpc_call_state(call) != RXRPC_CALL_CLIENT_RECV_REPLY)
5e6ef4f1 DH	495	rxrpc_resend(call, NULL);
	496
	497	if (test_and_clear_bit(RXRPC_CALL_RX_IS_IDLE, &call->flags))
	498	rxrpc_send_ACK(call, RXRPC_ACK_IDLE, 0,
	499	rxrpc_propose_ack_rx_idle);
	500
5bbf9533	501	if (call->ackr_nr_unacked > 2) {
84e28aa5 DH	502	if (call->peer->rtt_count < 3)
	503	rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
	504	rxrpc_propose_ack_ping_for_rtt);
	505	else if (ktime_before(ktime_add_ms(call->peer->rtt_last_req, 1000),
	506	ktime_get_real()))
	507	rxrpc_send_ACK(call, RXRPC_ACK_PING, 0,
	508	rxrpc_propose_ack_ping_for_old_rtt);
	509	else
	510	rxrpc_send_ACK(call, RXRPC_ACK_IDLE, 0,
	511	rxrpc_propose_ack_input_data);
	512	}
17926a79	513
a158bdd3	514	/* Make sure the timer is restarted */
96b4059f	515	if (!__rxrpc_call_is_complete(call)) {
5e6ef4f1	516	next = call->expect_rx_by;
a158bdd3 DH	517
a158bdd3 DH	518	#define set(T) { t = READ_ONCE(T); if (time_before(t, next)) next = t; }
3d7682af	519
5e6ef4f1 DH	520	set(call->expect_req_by);
	521	set(call->expect_term_by);
	522	set(call->delay_ack_at);
	523	set(call->ack_lost_at);
	524	set(call->resend_at);
	525	set(call->keepalive_at);
	526	set(call->ping_at);
a158bdd3	527
5e6ef4f1 DH	528	now = jiffies;
	529	if (time_after_eq(now, next))
	530	rxrpc_poke_call(call, rxrpc_call_poke_timer_now);
17926a79	531
5e6ef4f1 DH	532	rxrpc_reduce_call_timer(call, next, now, rxrpc_timer_restart);
5e6ef4f1 DH	533	}
17926a79	534
248f219c	535	out:
96b4059f	536	if (__rxrpc_call_is_complete(call)) {
5e6ef4f1	537	del_timer_sync(&call->timer);
03fc55ad DH	538	if (!test_bit(RXRPC_CALL_DISCONNECTED, &call->flags))
	539	rxrpc_disconnect_call(call);
	540	if (call->security)
	541	call->security->free_call_crypto(call);
	542	}
5e6ef4f1 DH	543	if (call->acks_hard_ack != call->tx_bottom)
5e6ef4f1 DH	544	rxrpc_shrink_call_tx_buffer(call);
17926a79	545	_leave("");
57af281e	546	return true;
17926a79	547	}