[mirror_ubuntu-bionic-kernel.git] / net / rxrpc / output.c

/* RxRPC packet transmission
 *
 * Copyright (C) 2007 Red Hat, Inc. All Rights Reserved.
 * Written by David Howells (dhowells@redhat.com)
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License
 * as published by the Free Software Foundation; either version
 * 2 of the License, or (at your option) any later version.
 */

#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

#include <linux/net.h>
#include <linux/gfp.h>
#include <linux/skbuff.h>
#include <linux/export.h>
#include <net/sock.h>
#include <net/af_rxrpc.h>
#include "ar-internal.h"

struct rxrpc_ack_buffer {
	struct rxrpc_wire_header whdr;
	struct rxrpc_ackpacket ack;
	u8 acks[255];
	u8 pad[3];
	struct rxrpc_ackinfo ackinfo;
};

struct rxrpc_abort_buffer {
	struct rxrpc_wire_header whdr;
	__be32 abort_code;
};

/*
 * Arrange for a keepalive ping a certain time after we last transmitted.  This
 * lets the far side know we're still interested in this call and helps keep
 * the route through any intervening firewall open.
 *
 * Receiving a response to the ping will prevent the ->expect_rx_by timer from
 * expiring.
 */
static void rxrpc_set_keepalive(struct rxrpc_call *call)
{
	unsigned long now = jiffies, keepalive_at = call->next_rx_timo / 6;

	keepalive_at += now;
	WRITE_ONCE(call->keepalive_at, keepalive_at);
	rxrpc_reduce_call_timer(call, keepalive_at, now,
				rxrpc_timer_set_for_keepalive);
}

/*
 * Fill out an ACK packet.
 */
static size_t rxrpc_fill_out_ack(struct rxrpc_connection *conn,
				 struct rxrpc_call *call,
				 struct rxrpc_ack_buffer *pkt,
				 rxrpc_seq_t *_hard_ack,
				 rxrpc_seq_t *_top,
				 u8 reason)
{
	rxrpc_serial_t serial;
	rxrpc_seq_t hard_ack, top, seq;
	int ix;
	u32 mtu, jmax;
	u8 *ackp = pkt->acks;

	/* Barrier against rxrpc_input_data(). */
	serial = call->ackr_serial;
	hard_ack = READ_ONCE(call->rx_hard_ack);
	top = smp_load_acquire(&call->rx_top);
	*_hard_ack = hard_ack;
	*_top = top;

	pkt->ack.bufferSpace	= htons(8);
	pkt->ack.maxSkew	= htons(call->ackr_skew);
	pkt->ack.firstPacket	= htonl(hard_ack + 1);
	pkt->ack.previousPacket	= htonl(call->ackr_prev_seq);
	pkt->ack.serial		= htonl(serial);
	pkt->ack.reason		= reason;
	pkt->ack.nAcks		= top - hard_ack;

	if (reason == RXRPC_ACK_PING)
		pkt->whdr.flags |= RXRPC_REQUEST_ACK;

	if (after(top, hard_ack)) {
		seq = hard_ack + 1;
		do {
			ix = seq & RXRPC_RXTX_BUFF_MASK;
			if (call->rxtx_buffer[ix])
				*ackp++ = RXRPC_ACK_TYPE_ACK;
			else
				*ackp++ = RXRPC_ACK_TYPE_NACK;
			seq++;
		} while (before_eq(seq, top));
	}

	mtu = conn->params.peer->if_mtu;
	mtu -= conn->params.peer->hdrsize;
	jmax = (call->nr_jumbo_bad > 3) ? 1 : rxrpc_rx_jumbo_max;
	pkt->ackinfo.rxMTU	= htonl(rxrpc_rx_mtu);
	pkt->ackinfo.maxMTU	= htonl(mtu);
	pkt->ackinfo.rwind	= htonl(call->rx_winsize);
	pkt->ackinfo.jumbo_max	= htonl(jmax);

	*ackp++ = 0;
	*ackp++ = 0;
	*ackp++ = 0;
	return top - hard_ack + 3;
}

/*
 * Send an ACK call packet.
 */
int rxrpc_send_ack_packet(struct rxrpc_call *call, bool ping,
			  rxrpc_serial_t *_serial)
{
	struct rxrpc_connection *conn = NULL;
	struct rxrpc_ack_buffer *pkt;
	struct msghdr msg;
	struct kvec iov[2];
	rxrpc_serial_t serial;
	rxrpc_seq_t hard_ack, top;
	size_t len, n;
	int ret;
	u8 reason;

	spin_lock_bh(&call->lock);
	if (call->conn)
		conn = rxrpc_get_connection_maybe(call->conn);
	spin_unlock_bh(&call->lock);
	if (!conn)
		return -ECONNRESET;

	pkt = kzalloc(sizeof(*pkt), GFP_KERNEL);
	if (!pkt) {
		rxrpc_put_connection(conn);
		return -ENOMEM;
	}

	msg.msg_name	= &call->peer->srx.transport;
	msg.msg_namelen	= call->peer->srx.transport_len;
	msg.msg_control	= NULL;
	msg.msg_controllen = 0;
	msg.msg_flags	= 0;

	pkt->whdr.epoch		= htonl(conn->proto.epoch);
	pkt->whdr.cid		= htonl(call->cid);
	pkt->whdr.callNumber	= htonl(call->call_id);
	pkt->whdr.seq		= 0;
	pkt->whdr.type		= RXRPC_PACKET_TYPE_ACK;
	pkt->whdr.flags		= RXRPC_SLOW_START_OK | conn->out_clientflag;
	pkt->whdr.userStatus	= 0;
	pkt->whdr.securityIndex	= call->security_ix;
	pkt->whdr._rsvd		= 0;
	pkt->whdr.serviceId	= htons(call->service_id);

	spin_lock_bh(&call->lock);
	if (ping) {
		reason = RXRPC_ACK_PING;
	} else {
		reason = call->ackr_reason;
		if (!call->ackr_reason) {
			spin_unlock_bh(&call->lock);
			ret = 0;
			goto out;
		}
		call->ackr_reason = 0;
	}
	n = rxrpc_fill_out_ack(conn, call, pkt, &hard_ack, &top, reason);

	spin_unlock_bh(&call->lock);

	iov[0].iov_base	= pkt;
	iov[0].iov_len	= sizeof(pkt->whdr) + sizeof(pkt->ack) + n;
	iov[1].iov_base = &pkt->ackinfo;
	iov[1].iov_len	= sizeof(pkt->ackinfo);
	len = iov[0].iov_len + iov[1].iov_len;

	serial = atomic_inc_return(&conn->serial);
	pkt->whdr.serial = htonl(serial);
	trace_rxrpc_tx_ack(call, serial,
			   ntohl(pkt->ack.firstPacket),
			   ntohl(pkt->ack.serial),
			   pkt->ack.reason, pkt->ack.nAcks);
	if (_serial)
		*_serial = serial;

	if (ping) {
		call->ping_serial = serial;
		smp_wmb();
		/* We need to stick a time in before we send the packet in case
		 * the reply gets back before kernel_sendmsg() completes - but
		 * asking UDP to send the packet can take a relatively long
		 * time, so we update the time after, on the assumption that
		 * the packet transmission is more likely to happen towards the
		 * end of the kernel_sendmsg() call.
		 */
		call->ping_time = ktime_get_real();
		set_bit(RXRPC_CALL_PINGING, &call->flags);
		trace_rxrpc_rtt_tx(call, rxrpc_rtt_tx_ping, serial);
	}

	ret = kernel_sendmsg(conn->params.local->socket, &msg, iov, 2, len);
	if (ping)
		call->ping_time = ktime_get_real();

	if (call->state < RXRPC_CALL_COMPLETE) {
		if (ret < 0) {
			if (ping)
				clear_bit(RXRPC_CALL_PINGING, &call->flags);
			rxrpc_propose_ACK(call, pkt->ack.reason,
					  ntohs(pkt->ack.maxSkew),
					  ntohl(pkt->ack.serial),
					  true, true,
					  rxrpc_propose_ack_retry_tx);
		} else {
			spin_lock_bh(&call->lock);
			if (after(hard_ack, call->ackr_consumed))
				call->ackr_consumed = hard_ack;
			if (after(top, call->ackr_seen))
				call->ackr_seen = top;
			spin_unlock_bh(&call->lock);
		}

		rxrpc_set_keepalive(call);
	}

out:
	rxrpc_put_connection(conn);
	kfree(pkt);
	return ret;
}

/*
 * Send an ABORT call packet.
 */
int rxrpc_send_abort_packet(struct rxrpc_call *call)
{
	struct rxrpc_connection *conn = NULL;
	struct rxrpc_abort_buffer pkt;
	struct msghdr msg;
	struct kvec iov[1];
	rxrpc_serial_t serial;
	int ret;

	/* Don't bother sending aborts for a client call once the server has
	 * hard-ACK'd all of its request data.  After that point, we're not
	 * going to stop the operation proceeding, and whilst we might limit
	 * the reply, it's not worth it if we can send a new call on the same
	 * channel instead, thereby closing off this call.
	 */
	if (rxrpc_is_client_call(call) &&
	    test_bit(RXRPC_CALL_TX_LAST, &call->flags))
		return 0;

	spin_lock_bh(&call->lock);
	if (call->conn)
		conn = rxrpc_get_connection_maybe(call->conn);
	spin_unlock_bh(&call->lock);
	if (!conn)
		return -ECONNRESET;

	msg.msg_name	= &call->peer->srx.transport;
	msg.msg_namelen	= call->peer->srx.transport_len;
	msg.msg_control	= NULL;
	msg.msg_controllen = 0;
	msg.msg_flags	= 0;

	pkt.whdr.epoch		= htonl(conn->proto.epoch);
	pkt.whdr.cid		= htonl(call->cid);
	pkt.whdr.callNumber	= htonl(call->call_id);
	pkt.whdr.seq		= 0;
	pkt.whdr.type		= RXRPC_PACKET_TYPE_ABORT;
	pkt.whdr.flags		= conn->out_clientflag;
	pkt.whdr.userStatus	= 0;
	pkt.whdr.securityIndex	= call->security_ix;
	pkt.whdr._rsvd		= 0;
	pkt.whdr.serviceId	= htons(call->service_id);
	pkt.abort_code		= htonl(call->abort_code);

	iov[0].iov_base	= &pkt;
	iov[0].iov_len	= sizeof(pkt);

	serial = atomic_inc_return(&conn->serial);
	pkt.whdr.serial = htonl(serial);

	ret = kernel_sendmsg(conn->params.local->socket,
			     &msg, iov, 1, sizeof(pkt));

	rxrpc_put_connection(conn);
	return ret;
}

/*
 * send a packet through the transport endpoint
 */
int rxrpc_send_data_packet(struct rxrpc_call *call, struct sk_buff *skb,
			   bool retrans)
{
	struct rxrpc_connection *conn = call->conn;
	struct rxrpc_wire_header whdr;
	struct rxrpc_skb_priv *sp = rxrpc_skb(skb);
	struct msghdr msg;
	struct kvec iov[2];
	rxrpc_serial_t serial;
	size_t len;
	bool lost = false;
	int ret, opt;

	_enter(",{%d}", skb->len);

	/* Each transmission of a Tx packet needs a new serial number */
	serial = atomic_inc_return(&conn->serial);

	whdr.epoch	= htonl(conn->proto.epoch);
	whdr.cid	= htonl(call->cid);
	whdr.callNumber	= htonl(call->call_id);
	whdr.seq	= htonl(sp->hdr.seq);
	whdr.serial	= htonl(serial);
	whdr.type	= RXRPC_PACKET_TYPE_DATA;
	whdr.flags	= sp->hdr.flags;
	whdr.userStatus	= 0;
	whdr.securityIndex = call->security_ix;
	whdr._rsvd	= htons(sp->hdr._rsvd);
	whdr.serviceId	= htons(call->service_id);

	if (test_bit(RXRPC_CONN_PROBING_FOR_UPGRADE, &conn->flags) &&
	    sp->hdr.seq == 1)
		whdr.userStatus	= RXRPC_USERSTATUS_SERVICE_UPGRADE;

	iov[0].iov_base = &whdr;
	iov[0].iov_len = sizeof(whdr);
	iov[1].iov_base = skb->head;
	iov[1].iov_len = skb->len;
	len = iov[0].iov_len + iov[1].iov_len;

	msg.msg_name = &call->peer->srx.transport;
	msg.msg_namelen = call->peer->srx.transport_len;
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_flags = 0;

	/* If our RTT cache needs working on, request an ACK.  Also request
	 * ACKs if a DATA packet appears to have been lost.
	 */
	if (!(sp->hdr.flags & RXRPC_LAST_PACKET) &&
	    (test_and_clear_bit(RXRPC_CALL_EV_ACK_LOST, &call->events) ||
	     retrans ||
	     call->cong_mode == RXRPC_CALL_SLOW_START ||
	     (call->peer->rtt_usage < 3 && sp->hdr.seq & 1) ||
	     ktime_before(ktime_add_ms(call->peer->rtt_last_req, 1000),
			  ktime_get_real())))
		whdr.flags |= RXRPC_REQUEST_ACK;

	if (IS_ENABLED(CONFIG_AF_RXRPC_INJECT_LOSS)) {
		static int lose;
		if ((lose++ & 7) == 7) {
			ret = 0;
			lost = true;
			goto done;
		}
	}

	_proto("Tx DATA %%%u { #%u }", serial, sp->hdr.seq);

	/* send the packet with the don't fragment bit set if we currently
	 * think it's small enough */
	if (iov[1].iov_len >= call->peer->maxdata)
		goto send_fragmentable;

	down_read(&conn->params.local->defrag_sem);
	/* send the packet by UDP
	 * - returns -EMSGSIZE if UDP would have to fragment the packet
	 *   to go out of the interface
	 *   - in which case, we'll have processed the ICMP error
	 *     message and update the peer record
	 */
	ret = kernel_sendmsg(conn->params.local->socket, &msg, iov, 2, len);

	up_read(&conn->params.local->defrag_sem);
	if (ret == -EMSGSIZE)
		goto send_fragmentable;

done:
	trace_rxrpc_tx_data(call, sp->hdr.seq, serial, whdr.flags,
			    retrans, lost);
	if (ret >= 0) {
		ktime_t now = ktime_get_real();
		skb->tstamp = now;
		smp_wmb();
		sp->hdr.serial = serial;
		if (whdr.flags & RXRPC_REQUEST_ACK) {
			call->peer->rtt_last_req = now;
			trace_rxrpc_rtt_tx(call, rxrpc_rtt_tx_data, serial);
			if (call->peer->rtt_usage > 1) {
				unsigned long nowj = jiffies, ack_lost_at;

				ack_lost_at = nsecs_to_jiffies(2 * call->peer->rtt);
				if (ack_lost_at < 1)
					ack_lost_at = 1;

				ack_lost_at += nowj;
				WRITE_ONCE(call->ack_lost_at, ack_lost_at);
				rxrpc_reduce_call_timer(call, ack_lost_at, nowj,
							rxrpc_timer_set_for_lost_ack);
			}
		}
	}

	rxrpc_set_keepalive(call);

	_leave(" = %d [%u]", ret, call->peer->maxdata);
	return ret;

send_fragmentable:
	/* attempt to send this message with fragmentation enabled */
	_debug("send fragment");

	down_write(&conn->params.local->defrag_sem);

	switch (conn->params.local->srx.transport.family) {
	case AF_INET:
		opt = IP_PMTUDISC_DONT;
		ret = kernel_setsockopt(conn->params.local->socket,
					SOL_IP, IP_MTU_DISCOVER,
					(char *)&opt, sizeof(opt));
		if (ret == 0) {
			ret = kernel_sendmsg(conn->params.local->socket, &msg,
					     iov, 2, len);

			opt = IP_PMTUDISC_DO;
			kernel_setsockopt(conn->params.local->socket, SOL_IP,
					  IP_MTU_DISCOVER,
					  (char *)&opt, sizeof(opt));
		}
		break;

#ifdef CONFIG_AF_RXRPC_IPV6
	case AF_INET6:
		opt = IPV6_PMTUDISC_DONT;
		ret = kernel_setsockopt(conn->params.local->socket,
					SOL_IPV6, IPV6_MTU_DISCOVER,
					(char *)&opt, sizeof(opt));
		if (ret == 0) {
			ret = kernel_sendmsg(conn->params.local->socket, &msg,
					     iov, 1, iov[0].iov_len);

			opt = IPV6_PMTUDISC_DO;
			kernel_setsockopt(conn->params.local->socket,
					  SOL_IPV6, IPV6_MTU_DISCOVER,
					  (char *)&opt, sizeof(opt));
		}
		break;
#endif
	}

	up_write(&conn->params.local->defrag_sem);
	goto done;
}

/*
 * reject packets through the local endpoint
 */
void rxrpc_reject_packets(struct rxrpc_local *local)
{
	struct sockaddr_rxrpc srx;
	struct rxrpc_skb_priv *sp;
	struct rxrpc_wire_header whdr;
	struct sk_buff *skb;
	struct msghdr msg;
	struct kvec iov[2];
	size_t size;
	__be32 code;

	_enter("%d", local->debug_id);

	iov[0].iov_base = &whdr;
	iov[0].iov_len = sizeof(whdr);
	iov[1].iov_base = &code;
	iov[1].iov_len = sizeof(code);
	size = sizeof(whdr) + sizeof(code);

	msg.msg_name = &srx.transport;
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_flags = 0;

	memset(&whdr, 0, sizeof(whdr));
	whdr.type = RXRPC_PACKET_TYPE_ABORT;

	while ((skb = skb_dequeue(&local->reject_queue))) {
		rxrpc_see_skb(skb, rxrpc_skb_rx_seen);
		sp = rxrpc_skb(skb);

		if (rxrpc_extract_addr_from_skb(local, &srx, skb) == 0) {
			msg.msg_namelen = srx.transport_len;

			code = htonl(skb->priority);

			whdr.epoch	= htonl(sp->hdr.epoch);
			whdr.cid	= htonl(sp->hdr.cid);
			whdr.callNumber	= htonl(sp->hdr.callNumber);
			whdr.serviceId	= htons(sp->hdr.serviceId);
			whdr.flags	= sp->hdr.flags;
			whdr.flags	^= RXRPC_CLIENT_INITIATED;
			whdr.flags	&= RXRPC_CLIENT_INITIATED;

			kernel_sendmsg(local->socket, &msg, iov, 2, size);
		}

		rxrpc_free_skb(skb, rxrpc_skb_rx_freed);
	}

	_leave("");
}
Commit	Line	Data
17926a79 DH	1	/* RxRPC packet transmission
	2	*
	3	* Copyright (C) 2007 Red Hat, Inc. All Rights Reserved.
	4	* Written by David Howells (dhowells@redhat.com)
	5	*
	6	* This program is free software; you can redistribute it and/or
	7	* modify it under the terms of the GNU General Public License
	8	* as published by the Free Software Foundation; either version
	9	* 2 of the License, or (at your option) any later version.
	10	*/
	11
9b6d5398 JP	12	#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
9b6d5398 JP	13
17926a79	14	#include <linux/net.h>
5a0e3ad6	15	#include <linux/gfp.h>
17926a79	16	#include <linux/skbuff.h>
bc3b2d7f	17	#include <linux/export.h>
17926a79 DH	18	#include <net/sock.h>
	19	#include <net/af_rxrpc.h>
	20	#include "ar-internal.h"
	21
26cb02aa	22	struct rxrpc_ack_buffer {
8d94aa38	23	struct rxrpc_wire_header whdr;
26cb02aa DH	24	struct rxrpc_ackpacket ack;
	25	u8 acks[255];
	26	u8 pad[3];
8d94aa38 DH	27	struct rxrpc_ackinfo ackinfo;
	28	};
	29
26cb02aa DH	30	struct rxrpc_abort_buffer {
	31	struct rxrpc_wire_header whdr;
	32	__be32 abort_code;
	33	};
	34
415f44e4 DH	35	/*
	36	* Arrange for a keepalive ping a certain time after we last transmitted. This
	37	* lets the far side know we're still interested in this call and helps keep
	38	* the route through any intervening firewall open.
	39	*
	40	* Receiving a response to the ping will prevent the ->expect_rx_by timer from
	41	* expiring.
	42	*/
	43	static void rxrpc_set_keepalive(struct rxrpc_call *call)
	44	{
	45	unsigned long now = jiffies, keepalive_at = call->next_rx_timo / 6;
	46
	47	keepalive_at += now;
	48	WRITE_ONCE(call->keepalive_at, keepalive_at);
	49	rxrpc_reduce_call_timer(call, keepalive_at, now,
	50	rxrpc_timer_set_for_keepalive);
	51	}
	52
8d94aa38 DH	53	/*
	54	* Fill out an ACK packet.
	55	*/
1457cc4c DH	56	static size_t rxrpc_fill_out_ack(struct rxrpc_connection *conn,
1457cc4c DH	57	struct rxrpc_call *call,
26cb02aa	58	struct rxrpc_ack_buffer *pkt,
805b21b9	59	rxrpc_seq_t *_hard_ack,
a5af7e1f DH	60	rxrpc_seq_t *_top,
a5af7e1f DH	61	u8 reason)
8d94aa38	62	{
f3639df2	63	rxrpc_serial_t serial;
248f219c DH	64	rxrpc_seq_t hard_ack, top, seq;
248f219c DH	65	int ix;
8d94aa38 DH	66	u32 mtu, jmax;
	67	u8 *ackp = pkt->acks;
	68
248f219c	69	/* Barrier against rxrpc_input_data(). */
f3639df2	70	serial = call->ackr_serial;
248f219c DH	71	hard_ack = READ_ONCE(call->rx_hard_ack);
248f219c DH	72	top = smp_load_acquire(&call->rx_top);
805b21b9 DH	73	*_hard_ack = hard_ack;
805b21b9 DH	74	*_top = top;
248f219c	75
8d94aa38	76	pkt->ack.bufferSpace = htons(8);
248f219c DH	77	pkt->ack.maxSkew = htons(call->ackr_skew);
248f219c DH	78	pkt->ack.firstPacket = htonl(hard_ack + 1);
8d94aa38	79	pkt->ack.previousPacket = htonl(call->ackr_prev_seq);
f3639df2	80	pkt->ack.serial = htonl(serial);
a5af7e1f	81	pkt->ack.reason = reason;
248f219c DH	82	pkt->ack.nAcks = top - hard_ack;
248f219c DH	83
a5af7e1f	84	if (reason == RXRPC_ACK_PING)
8e83134d DH	85	pkt->whdr.flags \|= RXRPC_REQUEST_ACK;
8e83134d DH	86
248f219c DH	87	if (after(top, hard_ack)) {
	88	seq = hard_ack + 1;
	89	do {
	90	ix = seq & RXRPC_RXTX_BUFF_MASK;
	91	if (call->rxtx_buffer[ix])
	92	*ackp++ = RXRPC_ACK_TYPE_ACK;
	93	else
	94	*ackp++ = RXRPC_ACK_TYPE_NACK;
	95	seq++;
	96	} while (before_eq(seq, top));
	97	}
8d94aa38	98
1457cc4c DH	99	mtu = conn->params.peer->if_mtu;
1457cc4c DH	100	mtu -= conn->params.peer->hdrsize;
75e42126	101	jmax = (call->nr_jumbo_bad > 3) ? 1 : rxrpc_rx_jumbo_max;
8d94aa38 DH	102	pkt->ackinfo.rxMTU = htonl(rxrpc_rx_mtu);
8d94aa38 DH	103	pkt->ackinfo.maxMTU = htonl(mtu);
75e42126	104	pkt->ackinfo.rwind = htonl(call->rx_winsize);
8d94aa38 DH	105	pkt->ackinfo.jumbo_max = htonl(jmax);
	106
	107	*ackp++ = 0;
	108	*ackp++ = 0;
	109	*ackp++ = 0;
248f219c	110	return top - hard_ack + 3;
8d94aa38 DH	111	}
	112
	113	/*
26cb02aa	114	* Send an ACK call packet.
8d94aa38	115	*/
bd1fdf8c DH	116	int rxrpc_send_ack_packet(struct rxrpc_call *call, bool ping,
bd1fdf8c DH	117	rxrpc_serial_t *_serial)
8d94aa38 DH	118	{
8d94aa38 DH	119	struct rxrpc_connection *conn = NULL;
26cb02aa	120	struct rxrpc_ack_buffer *pkt;
8d94aa38 DH	121	struct msghdr msg;
	122	struct kvec iov[2];
	123	rxrpc_serial_t serial;
805b21b9	124	rxrpc_seq_t hard_ack, top;
8d94aa38	125	size_t len, n;
26cb02aa	126	int ret;
a5af7e1f	127	u8 reason;
8d94aa38 DH	128
	129	spin_lock_bh(&call->lock);
	130	if (call->conn)
	131	conn = rxrpc_get_connection_maybe(call->conn);
	132	spin_unlock_bh(&call->lock);
	133	if (!conn)
	134	return -ECONNRESET;
	135
	136	pkt = kzalloc(sizeof(*pkt), GFP_KERNEL);
	137	if (!pkt) {
	138	rxrpc_put_connection(conn);
	139	return -ENOMEM;
	140	}
	141
8d94aa38 DH	142	msg.msg_name = &call->peer->srx.transport;
	143	msg.msg_namelen = call->peer->srx.transport_len;
	144	msg.msg_control = NULL;
	145	msg.msg_controllen = 0;
	146	msg.msg_flags = 0;
	147
	148	pkt->whdr.epoch = htonl(conn->proto.epoch);
	149	pkt->whdr.cid = htonl(call->cid);
	150	pkt->whdr.callNumber = htonl(call->call_id);
	151	pkt->whdr.seq = 0;
26cb02aa DH	152	pkt->whdr.type = RXRPC_PACKET_TYPE_ACK;
26cb02aa DH	153	pkt->whdr.flags = RXRPC_SLOW_START_OK \| conn->out_clientflag;
8d94aa38 DH	154	pkt->whdr.userStatus = 0;
	155	pkt->whdr.securityIndex = call->security_ix;
	156	pkt->whdr._rsvd = 0;
	157	pkt->whdr.serviceId = htons(call->service_id);
	158
26cb02aa	159	spin_lock_bh(&call->lock);
a5af7e1f DH	160	if (ping) {
	161	reason = RXRPC_ACK_PING;
	162	} else {
	163	reason = call->ackr_reason;
	164	if (!call->ackr_reason) {
	165	spin_unlock_bh(&call->lock);
	166	ret = 0;
	167	goto out;
	168	}
	169	call->ackr_reason = 0;
8d94aa38	170	}
1457cc4c	171	n = rxrpc_fill_out_ack(conn, call, pkt, &hard_ack, &top, reason);
26cb02aa DH	172
	173	spin_unlock_bh(&call->lock);
	174
	175	iov[0].iov_base = pkt;
	176	iov[0].iov_len = sizeof(pkt->whdr) + sizeof(pkt->ack) + n;
	177	iov[1].iov_base = &pkt->ackinfo;
	178	iov[1].iov_len = sizeof(pkt->ackinfo);
	179	len = iov[0].iov_len + iov[1].iov_len;
8d94aa38	180
b86e218e DH	181	serial = atomic_inc_return(&conn->serial);
b86e218e DH	182	pkt->whdr.serial = htonl(serial);
26cb02aa DH	183	trace_rxrpc_tx_ack(call, serial,
	184	ntohl(pkt->ack.firstPacket),
	185	ntohl(pkt->ack.serial),
	186	pkt->ack.reason, pkt->ack.nAcks);
bd1fdf8c DH	187	if (_serial)
bd1fdf8c DH	188	*_serial = serial;
b86e218e	189
8e83134d	190	if (ping) {
a5af7e1f	191	call->ping_serial = serial;
8e83134d DH	192	smp_wmb();
	193	/* We need to stick a time in before we send the packet in case
	194	* the reply gets back before kernel_sendmsg() completes - but
	195	* asking UDP to send the packet can take a relatively long
	196	* time, so we update the time after, on the assumption that
	197	* the packet transmission is more likely to happen towards the
	198	* end of the kernel_sendmsg() call.
	199	*/
a5af7e1f	200	call->ping_time = ktime_get_real();
8e83134d DH	201	set_bit(RXRPC_CALL_PINGING, &call->flags);
	202	trace_rxrpc_rtt_tx(call, rxrpc_rtt_tx_ping, serial);
	203	}
26cb02aa DH	204
26cb02aa DH	205	ret = kernel_sendmsg(conn->params.local->socket, &msg, iov, 2, len);
8e83134d	206	if (ping)
a5af7e1f	207	call->ping_time = ktime_get_real();
8d94aa38	208
26cb02aa	209	if (call->state < RXRPC_CALL_COMPLETE) {
805b21b9	210	if (ret < 0) {
a5af7e1f DH	211	if (ping)
a5af7e1f DH	212	clear_bit(RXRPC_CALL_PINGING, &call->flags);
248f219c DH	213	rxrpc_propose_ACK(call, pkt->ack.reason,
	214	ntohs(pkt->ack.maxSkew),
	215	ntohl(pkt->ack.serial),
9c7ad434 DH	216	true, true,
9c7ad434 DH	217	rxrpc_propose_ack_retry_tx);
805b21b9 DH	218	} else {
	219	spin_lock_bh(&call->lock);
	220	if (after(hard_ack, call->ackr_consumed))
	221	call->ackr_consumed = hard_ack;
	222	if (after(top, call->ackr_seen))
	223	call->ackr_seen = top;
	224	spin_unlock_bh(&call->lock);
248f219c	225	}
415f44e4 DH	226
415f44e4 DH	227	rxrpc_set_keepalive(call);
248f219c DH	228	}
248f219c DH	229
8d94aa38 DH	230	out:
	231	rxrpc_put_connection(conn);
	232	kfree(pkt);
	233	return ret;
	234	}
	235
26cb02aa DH	236	/*
	237	* Send an ABORT call packet.
	238	*/
	239	int rxrpc_send_abort_packet(struct rxrpc_call *call)
	240	{
	241	struct rxrpc_connection *conn = NULL;
	242	struct rxrpc_abort_buffer pkt;
	243	struct msghdr msg;
	244	struct kvec iov[1];
	245	rxrpc_serial_t serial;
	246	int ret;
	247
dcbefc30 DH	248	/* Don't bother sending aborts for a client call once the server has
	249	* hard-ACK'd all of its request data. After that point, we're not
	250	* going to stop the operation proceeding, and whilst we might limit
	251	* the reply, it's not worth it if we can send a new call on the same
	252	* channel instead, thereby closing off this call.
	253	*/
	254	if (rxrpc_is_client_call(call) &&
	255	test_bit(RXRPC_CALL_TX_LAST, &call->flags))
	256	return 0;
	257
26cb02aa DH	258	spin_lock_bh(&call->lock);
	259	if (call->conn)
	260	conn = rxrpc_get_connection_maybe(call->conn);
	261	spin_unlock_bh(&call->lock);
	262	if (!conn)
	263	return -ECONNRESET;
	264
	265	msg.msg_name = &call->peer->srx.transport;
	266	msg.msg_namelen = call->peer->srx.transport_len;
	267	msg.msg_control = NULL;
	268	msg.msg_controllen = 0;
	269	msg.msg_flags = 0;
	270
	271	pkt.whdr.epoch = htonl(conn->proto.epoch);
	272	pkt.whdr.cid = htonl(call->cid);
	273	pkt.whdr.callNumber = htonl(call->call_id);
	274	pkt.whdr.seq = 0;
	275	pkt.whdr.type = RXRPC_PACKET_TYPE_ABORT;
	276	pkt.whdr.flags = conn->out_clientflag;
	277	pkt.whdr.userStatus = 0;
	278	pkt.whdr.securityIndex = call->security_ix;
	279	pkt.whdr._rsvd = 0;
	280	pkt.whdr.serviceId = htons(call->service_id);
	281	pkt.abort_code = htonl(call->abort_code);
	282
	283	iov[0].iov_base = &pkt;
	284	iov[0].iov_len = sizeof(pkt);
	285
	286	serial = atomic_inc_return(&conn->serial);
	287	pkt.whdr.serial = htonl(serial);
	288
	289	ret = kernel_sendmsg(conn->params.local->socket,
	290	&msg, iov, 1, sizeof(pkt));
	291
	292	rxrpc_put_connection(conn);
	293	return ret;
	294	}
	295
17926a79 DH	296	/*
	297	* send a packet through the transport endpoint
	298	*/
a1767077 DH	299	int rxrpc_send_data_packet(struct rxrpc_call call, struct sk_buff skb,
a1767077 DH	300	bool retrans)
17926a79	301	{
5a924b89 DH	302	struct rxrpc_connection *conn = call->conn;
	303	struct rxrpc_wire_header whdr;
	304	struct rxrpc_skb_priv *sp = rxrpc_skb(skb);
17926a79	305	struct msghdr msg;
5a924b89 DH	306	struct kvec iov[2];
	307	rxrpc_serial_t serial;
	308	size_t len;
a1767077	309	bool lost = false;
17926a79 DH	310	int ret, opt;
	311
	312	_enter(",{%d}", skb->len);
	313
5a924b89 DH	314	/* Each transmission of a Tx packet needs a new serial number */
5a924b89 DH	315	serial = atomic_inc_return(&conn->serial);
17926a79	316
5a924b89 DH	317	whdr.epoch = htonl(conn->proto.epoch);
	318	whdr.cid = htonl(call->cid);
	319	whdr.callNumber = htonl(call->call_id);
	320	whdr.seq = htonl(sp->hdr.seq);
	321	whdr.serial = htonl(serial);
	322	whdr.type = RXRPC_PACKET_TYPE_DATA;
	323	whdr.flags = sp->hdr.flags;
	324	whdr.userStatus = 0;
	325	whdr.securityIndex = call->security_ix;
	326	whdr._rsvd = htons(sp->hdr._rsvd);
	327	whdr.serviceId = htons(call->service_id);
	328
4e255721 DH	329	if (test_bit(RXRPC_CONN_PROBING_FOR_UPGRADE, &conn->flags) &&
	330	sp->hdr.seq == 1)
	331	whdr.userStatus = RXRPC_USERSTATUS_SERVICE_UPGRADE;
	332
5a924b89 DH	333	iov[0].iov_base = &whdr;
	334	iov[0].iov_len = sizeof(whdr);
	335	iov[1].iov_base = skb->head;
	336	iov[1].iov_len = skb->len;
	337	len = iov[0].iov_len + iov[1].iov_len;
	338
	339	msg.msg_name = &call->peer->srx.transport;
	340	msg.msg_namelen = call->peer->srx.transport_len;
17926a79 DH	341	msg.msg_control = NULL;
	342	msg.msg_controllen = 0;
	343	msg.msg_flags = 0;
	344
57494343 DH	345	/* If our RTT cache needs working on, request an ACK. Also request
	346	* ACKs if a DATA packet appears to have been lost.
	347	*/
bf7d620a	348	if (!(sp->hdr.flags & RXRPC_LAST_PACKET) &&
bd1fdf8c DH	349	(test_and_clear_bit(RXRPC_CALL_EV_ACK_LOST, &call->events) \|\|
bd1fdf8c DH	350	retrans \|\|
bf7d620a DH	351	call->cong_mode == RXRPC_CALL_SLOW_START \|\|
	352	(call->peer->rtt_usage < 3 && sp->hdr.seq & 1) \|\|
	353	ktime_before(ktime_add_ms(call->peer->rtt_last_req, 1000),
	354	ktime_get_real())))
0d4b103c DH	355	whdr.flags \|= RXRPC_REQUEST_ACK;
0d4b103c DH	356
8a681c36 DH	357	if (IS_ENABLED(CONFIG_AF_RXRPC_INJECT_LOSS)) {
	358	static int lose;
	359	if ((lose++ & 7) == 7) {
a1767077 DH	360	ret = 0;
	361	lost = true;
	362	goto done;
8a681c36 DH	363	}
	364	}
	365
5a924b89 DH	366	_proto("Tx DATA %%%u { #%u }", serial, sp->hdr.seq);
5a924b89 DH	367
17926a79 DH	368	/* send the packet with the don't fragment bit set if we currently
17926a79 DH	369	* think it's small enough */
5a924b89 DH	370	if (iov[1].iov_len >= call->peer->maxdata)
	371	goto send_fragmentable;
	372
	373	down_read(&conn->params.local->defrag_sem);
	374	/* send the packet by UDP
	375	* - returns -EMSGSIZE if UDP would have to fragment the packet
	376	* to go out of the interface
	377	* - in which case, we'll have processed the ICMP error
	378	* message and update the peer record
	379	*/
	380	ret = kernel_sendmsg(conn->params.local->socket, &msg, iov, 2, len);
	381
	382	up_read(&conn->params.local->defrag_sem);
	383	if (ret == -EMSGSIZE)
	384	goto send_fragmentable;
	385
	386	done:
a1767077 DH	387	trace_rxrpc_tx_data(call, sp->hdr.seq, serial, whdr.flags,
a1767077 DH	388	retrans, lost);
50235c4b	389	if (ret >= 0) {
0d4b103c DH	390	ktime_t now = ktime_get_real();
0d4b103c DH	391	skb->tstamp = now;
50235c4b	392	smp_wmb();
5a924b89	393	sp->hdr.serial = serial;
0d4b103c DH	394	if (whdr.flags & RXRPC_REQUEST_ACK) {
0d4b103c DH	395	call->peer->rtt_last_req = now;
50235c4b	396	trace_rxrpc_rtt_tx(call, rxrpc_rtt_tx_data, serial);
bd1fdf8c DH	397	if (call->peer->rtt_usage > 1) {
	398	unsigned long nowj = jiffies, ack_lost_at;
	399
	400	ack_lost_at = nsecs_to_jiffies(2 * call->peer->rtt);
	401	if (ack_lost_at < 1)
	402	ack_lost_at = 1;
	403
	404	ack_lost_at += nowj;
	405	WRITE_ONCE(call->ack_lost_at, ack_lost_at);
	406	rxrpc_reduce_call_timer(call, ack_lost_at, nowj,
	407	rxrpc_timer_set_for_lost_ack);
	408	}
0d4b103c	409	}
17926a79	410	}
415f44e4 DH	411
	412	rxrpc_set_keepalive(call);
	413
5a924b89 DH	414	_leave(" = %d [%u]", ret, call->peer->maxdata);
5a924b89 DH	415	return ret;
17926a79 DH	416
	417	send_fragmentable:
	418	/* attempt to send this message with fragmentation enabled */
	419	_debug("send fragment");
	420
985a5c82 DH	421	down_write(&conn->params.local->defrag_sem);
	422
	423	switch (conn->params.local->srx.transport.family) {
	424	case AF_INET:
	425	opt = IP_PMTUDISC_DONT;
	426	ret = kernel_setsockopt(conn->params.local->socket,
	427	SOL_IP, IP_MTU_DISCOVER,
	428	(char *)&opt, sizeof(opt));
	429	if (ret == 0) {
5a924b89 DH	430	ret = kernel_sendmsg(conn->params.local->socket, &msg,
5a924b89 DH	431	iov, 2, len);
985a5c82 DH	432
	433	opt = IP_PMTUDISC_DO;
	434	kernel_setsockopt(conn->params.local->socket, SOL_IP,
	435	IP_MTU_DISCOVER,
	436	(char *)&opt, sizeof(opt));
	437	}
	438	break;
75b54cb5	439
d1912747	440	#ifdef CONFIG_AF_RXRPC_IPV6
75b54cb5 DH	441	case AF_INET6:
	442	opt = IPV6_PMTUDISC_DONT;
	443	ret = kernel_setsockopt(conn->params.local->socket,
	444	SOL_IPV6, IPV6_MTU_DISCOVER,
	445	(char *)&opt, sizeof(opt));
	446	if (ret == 0) {
	447	ret = kernel_sendmsg(conn->params.local->socket, &msg,
	448	iov, 1, iov[0].iov_len);
	449
	450	opt = IPV6_PMTUDISC_DO;
	451	kernel_setsockopt(conn->params.local->socket,
	452	SOL_IPV6, IPV6_MTU_DISCOVER,
	453	(char *)&opt, sizeof(opt));
	454	}
	455	break;
d1912747	456	#endif
17926a79 DH	457	}
17926a79 DH	458
985a5c82	459	up_write(&conn->params.local->defrag_sem);
5a924b89	460	goto done;
17926a79	461	}
248f219c DH	462
	463	/*
	464	* reject packets through the local endpoint
	465	*/
	466	void rxrpc_reject_packets(struct rxrpc_local *local)
	467	{
1c2bc7b9	468	struct sockaddr_rxrpc srx;
248f219c DH	469	struct rxrpc_skb_priv *sp;
	470	struct rxrpc_wire_header whdr;
	471	struct sk_buff *skb;
	472	struct msghdr msg;
	473	struct kvec iov[2];
	474	size_t size;
	475	__be32 code;
	476
	477	_enter("%d", local->debug_id);
	478
	479	iov[0].iov_base = &whdr;
	480	iov[0].iov_len = sizeof(whdr);
	481	iov[1].iov_base = &code;
	482	iov[1].iov_len = sizeof(code);
	483	size = sizeof(whdr) + sizeof(code);
	484
1c2bc7b9	485	msg.msg_name = &srx.transport;
248f219c DH	486	msg.msg_control = NULL;
	487	msg.msg_controllen = 0;
	488	msg.msg_flags = 0;
	489
248f219c DH	490	memset(&whdr, 0, sizeof(whdr));
	491	whdr.type = RXRPC_PACKET_TYPE_ABORT;
	492
	493	while ((skb = skb_dequeue(&local->reject_queue))) {
71f3ca40	494	rxrpc_see_skb(skb, rxrpc_skb_rx_seen);
248f219c	495	sp = rxrpc_skb(skb);
1c2bc7b9	496
7b674e39	497	if (rxrpc_extract_addr_from_skb(local, &srx, skb) == 0) {
1c2bc7b9 DH	498	msg.msg_namelen = srx.transport_len;
1c2bc7b9 DH	499
248f219c DH	500	code = htonl(skb->priority);
	501
	502	whdr.epoch = htonl(sp->hdr.epoch);
	503	whdr.cid = htonl(sp->hdr.cid);
	504	whdr.callNumber = htonl(sp->hdr.callNumber);
	505	whdr.serviceId = htons(sp->hdr.serviceId);
	506	whdr.flags = sp->hdr.flags;
	507	whdr.flags ^= RXRPC_CLIENT_INITIATED;
	508	whdr.flags &= RXRPC_CLIENT_INITIATED;
	509
	510	kernel_sendmsg(local->socket, &msg, iov, 2, size);
248f219c DH	511	}
248f219c DH	512
71f3ca40	513	rxrpc_free_skb(skb, rxrpc_skb_rx_freed);
248f219c DH	514	}
	515
	516	_leave("");
	517	}