bcast.c

linux 内核源代码

/*
 * net/tipc/bcast.c: TIPC broadcast code
 *
 * Copyright (c) 2004-2006, Ericsson AB
 * Copyright (c) 2004, Intel Corporation.
 * Copyright (c) 2005, Wind River Systems
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are met:
 *
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. Neither the names of the copyright holders nor the names of its
 *    contributors may be used to endorse or promote products derived from
 *    this software without specific prior written permission.
 *
 * Alternatively, this software may be distributed under the terms of the
 * GNU General Public License ("GPL") version 2 as published by the Free
 * Software Foundation.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
 * POSSIBILITY OF SUCH DAMAGE.
 */

#include "core.h"
#include "msg.h"
#include "dbg.h"
#include "link.h"
#include "net.h"
#include "node.h"
#include "port.h"
#include "addr.h"
#include "node_subscr.h"
#include "name_distr.h"
#include "bearer.h"
#include "name_table.h"
#include "bcast.h"

#define MAX_PKT_DEFAULT_MCAST 1500	/* bcast link max packet size (fixed) */

#define BCLINK_WIN_DEFAULT 20		/* bcast link window size (default) */

#define BCLINK_LOG_BUF_SIZE 0

/*
 * Loss rate for incoming broadcast frames; used to test retransmission code.
 * Set to N to cause every N'th frame to be discarded; 0 => don't discard any.
 */

#define TIPC_BCAST_LOSS_RATE 0

/**
 * struct bcbearer_pair - a pair of bearers used by broadcast link
 * @primary: pointer to primary bearer
 * @secondary: pointer to secondary bearer
 *
 * Bearers must have same priority and same set of reachable destinations
 * to be paired.
 */

struct bcbearer_pair {
	struct bearer *primary;
	struct bearer *secondary;
};

/**
 * struct bcbearer - bearer used by broadcast link
 * @bearer: (non-standard) broadcast bearer structure
 * @media: (non-standard) broadcast media structure
 * @bpairs: array of bearer pairs
 * @bpairs_temp: temporary array of bearer pairs used by tipc_bcbearer_sort()
 * @remains: temporary node map used by tipc_bcbearer_send()
 * @remains_new: temporary node map used tipc_bcbearer_send()
 *
 * Note: The fields labelled "temporary" are incorporated into the bearer
 * to avoid consuming potentially limited stack space through the use of
 * large local variables within multicast routines.  Concurrent access is
 * prevented through use of the spinlock "bc_lock".
 */

struct bcbearer {
	struct bearer bearer;
	struct media media;
	struct bcbearer_pair bpairs[MAX_BEARERS];
	struct bcbearer_pair bpairs_temp[TIPC_MAX_LINK_PRI + 1];
	struct node_map remains;
	struct node_map remains_new;
};

/**
 * struct bclink - link used for broadcast messages
 * @link: (non-standard) broadcast link structure
 * @node: (non-standard) node structure representing b'cast link's peer node
 *
 * Handles sequence numbering, fragmentation, bundling, etc.
 */

struct bclink {
	struct link link;
	struct node node;
};


static struct bcbearer *bcbearer = NULL;
static struct bclink *bclink = NULL;
static struct link *bcl = NULL;
static DEFINE_SPINLOCK(bc_lock);

char tipc_bclink_name[] = "multicast-link";


static u32 buf_seqno(struct sk_buff *buf)
{
	return msg_seqno(buf_msg(buf));
}

static u32 bcbuf_acks(struct sk_buff *buf)
{
	return (u32)(unsigned long)TIPC_SKB_CB(buf)->handle;
}

static void bcbuf_set_acks(struct sk_buff *buf, u32 acks)
{
	TIPC_SKB_CB(buf)->handle = (void *)(unsigned long)acks;
}

static void bcbuf_decr_acks(struct sk_buff *buf)
{
	bcbuf_set_acks(buf, bcbuf_acks(buf) - 1);
}


/**
 * bclink_set_gap - set gap according to contents of current deferred pkt queue
 *
 * Called with 'node' locked, bc_lock unlocked
 */

static void bclink_set_gap(struct node *n_ptr)
{
	struct sk_buff *buf = n_ptr->bclink.deferred_head;

	n_ptr->bclink.gap_after = n_ptr->bclink.gap_to =
		mod(n_ptr->bclink.last_in);
	if (unlikely(buf != NULL))
		n_ptr->bclink.gap_to = mod(buf_seqno(buf) - 1);
}

/**
 * bclink_ack_allowed - test if ACK or NACK message can be sent at this moment
 *
 * This mechanism endeavours to prevent all nodes in network from trying
 * to ACK or NACK at the same time.
 *
 * Note: TIPC uses a different trigger to distribute ACKs than it does to
 *       distribute NACKs, but tries to use the same spacing (divide by 16).
 */

static int bclink_ack_allowed(u32 n)
{
	return((n % TIPC_MIN_LINK_WIN) == tipc_own_tag);
}


/**
 * bclink_retransmit_pkt - retransmit broadcast packets
 * @after: sequence number of last packet to *not* retransmit
 * @to: sequence number of last packet to retransmit
 *
 * Called with bc_lock locked
 */

static void bclink_retransmit_pkt(u32 after, u32 to)
{
	struct sk_buff *buf;

	buf = bcl->first_out;
	while (buf && less_eq(buf_seqno(buf), after)) {
		buf = buf->next;
	}
	tipc_link_retransmit(bcl, buf, mod(to - after));
}

/**
 * tipc_bclink_acknowledge - handle acknowledgement of broadcast packets
 * @n_ptr: node that sent acknowledgement info
 * @acked: broadcast sequence # that has been acknowledged
 *
 * Node is locked, bc_lock unlocked.
 */

void tipc_bclink_acknowledge(struct node *n_ptr, u32 acked)
{
	struct sk_buff *crs;
	struct sk_buff *next;
	unsigned int released = 0;

	if (less_eq(acked, n_ptr->bclink.acked))
		return;

	spin_lock_bh(&bc_lock);

	/* Skip over packets that node has previously acknowledged */

	crs = bcl->first_out;
	while (crs && less_eq(buf_seqno(crs), n_ptr->bclink.acked)) {
		crs = crs->next;
	}

	/* Update packets that node is now acknowledging */

	while (crs && less_eq(buf_seqno(crs), acked)) {
		next = crs->next;
		bcbuf_decr_acks(crs);
		if (bcbuf_acks(crs) == 0) {
			bcl->first_out = next;
			bcl->out_queue_size--;
			buf_discard(crs);
			released = 1;
		}
		crs = next;
	}
	n_ptr->bclink.acked = acked;

	/* Try resolving broadcast link congestion, if necessary */

	if (unlikely(bcl->next_out))
		tipc_link_push_queue(bcl);
	if (unlikely(released && !list_empty(&bcl->waiting_ports)))
		tipc_link_wakeup_ports(bcl, 0);
	spin_unlock_bh(&bc_lock);
}

/**
 * bclink_send_ack - unicast an ACK msg
 *
 * tipc_net_lock and node lock set
 */

static void bclink_send_ack(struct node *n_ptr)
{
	struct link *l_ptr = n_ptr->active_links[n_ptr->addr & 1];

	if (l_ptr != NULL)
		tipc_link_send_proto_msg(l_ptr, STATE_MSG, 0, 0, 0, 0, 0);
}

/**
 * bclink_send_nack- broadcast a NACK msg
 *
 * tipc_net_lock and node lock set
 */

static void bclink_send_nack(struct node *n_ptr)
{
	struct sk_buff *buf;
	struct tipc_msg *msg;

	if (!less(n_ptr->bclink.gap_after, n_ptr->bclink.gap_to))
		return;

	buf = buf_acquire(INT_H_SIZE);
	if (buf) {
		msg = buf_msg(buf);
		msg_init(msg, BCAST_PROTOCOL, STATE_MSG,
			 TIPC_OK, INT_H_SIZE, n_ptr->addr);
		msg_set_mc_netid(msg, tipc_net_id);
		msg_set_bcast_ack(msg, mod(n_ptr->bclink.last_in));
		msg_set_bcgap_after(msg, n_ptr->bclink.gap_after);
		msg_set_bcgap_to(msg, n_ptr->bclink.gap_to);
		msg_set_bcast_tag(msg, tipc_own_tag);

		if (tipc_bearer_send(&bcbearer->bearer, buf, NULL)) {
			bcl->stats.sent_nacks++;
			buf_discard(buf);
		} else {
			tipc_bearer_schedule(bcl->b_ptr, bcl);
			bcl->proto_msg_queue = buf;
			bcl->stats.bearer_congs++;
		}

		/*
		 * Ensure we doesn't send another NACK msg to the node
		 * until 16 more deferred messages arrive from it
		 * (i.e. helps prevent all nodes from NACK'ing at same time)
		 */

		n_ptr->bclink.nack_sync = tipc_own_tag;
	}
}

/**
 * tipc_bclink_check_gap - send a NACK if a sequence gap exists
 *
 * tipc_net_lock and node lock set
 */

void tipc_bclink_check_gap(struct node *n_ptr, u32 last_sent)
{
	if (!n_ptr->bclink.supported ||
	    less_eq(last_sent, mod(n_ptr->bclink.last_in)))
		return;

	bclink_set_gap(n_ptr);
	if (n_ptr->bclink.gap_after == n_ptr->bclink.gap_to)
		n_ptr->bclink.gap_to = last_sent;
	bclink_send_nack(n_ptr);
}

/**
 * tipc_bclink_peek_nack - process a NACK msg meant for another node
 *
 * Only tipc_net_lock set.
 */

static void tipc_bclink_peek_nack(u32 dest, u32 sender_tag, u32 gap_after, u32 gap_to)
{
	struct node *n_ptr = tipc_node_find(dest);
	u32 my_after, my_to;

	if (unlikely(!n_ptr || !tipc_node_is_up(n_ptr)))
		return;
	tipc_node_lock(n_ptr);
	/*
	 * Modify gap to suppress unnecessary NACKs from this node
	 */
	my_after = n_ptr->bclink.gap_after;
	my_to = n_ptr->bclink.gap_to;

	if (less_eq(gap_after, my_after)) {
		if (less(my_after, gap_to) && less(gap_to, my_to))
			n_ptr->bclink.gap_after = gap_to;
		else if (less_eq(my_to, gap_to))
			n_ptr->bclink.gap_to = n_ptr->bclink.gap_after;
	} else if (less_eq(gap_after, my_to)) {
		if (less_eq(my_to, gap_to))
			n_ptr->bclink.gap_to = gap_after;
	} else {
		/*
		 * Expand gap if missing bufs not in deferred queue:
		 */
		struct sk_buff *buf = n_ptr->bclink.deferred_head;
		u32 prev = n_ptr->bclink.gap_to;

		for (; buf; buf = buf->next) {
			u32 seqno = buf_seqno(buf);

			if (mod(seqno - prev) != 1) {
				buf = NULL;
				break;
			}
			if (seqno == gap_after)
				break;
			prev = seqno;
		}
		if (buf == NULL)
			n_ptr->bclink.gap_to = gap_after;
	}
	/*
	 * Some nodes may send a complementary NACK now:
	 */
	if (bclink_ack_allowed(sender_tag + 1)) {
		if (n_ptr->bclink.gap_to != n_ptr->bclink.gap_after) {
			bclink_send_nack(n_ptr);
			bclink_set_gap(n_ptr);
		}
	}
	tipc_node_unlock(n_ptr);
}

/**
 * tipc_bclink_send_msg - broadcast a packet to all nodes in cluster
 */

int tipc_bclink_send_msg(struct sk_buff *buf)
{
	int res;

	spin_lock_bh(&bc_lock);

	res = tipc_link_send_buf(bcl, buf);
	if (unlikely(res == -ELINKCONG))
		buf_discard(buf);
	else
		bcl->stats.sent_info++;

	if (bcl->out_queue_size > bcl->stats.max_queue_sz)
		bcl->stats.max_queue_sz = bcl->out_queue_size;
	bcl->stats.queue_sz_counts++;
	bcl->stats.accu_queue_sz += bcl->out_queue_size;

	spin_unlock_bh(&bc_lock);
	return res;
}

/**
 * tipc_bclink_recv_pkt - receive a broadcast packet, and deliver upwards
 *
 * tipc_net_lock is read_locked, no other locks set
 */

void tipc_bclink_recv_pkt(struct sk_buff *buf)
{