#include <linux/netfilter_bridge.h>
#include <linux/inetdevice.h>
#include <linux/list.h>
+#include <linux/rculist.h>
#include "openflow-netlink.h"
#include "datapath.h"
/* It's hard to imagine wanting more than one datapath, but... */
#define DP_MAX 32
- /* datapaths. Protected on the read side by rcu_read_lock, on the write side
- * by dp_mutex.
+ /* Datapaths. Protected on the read side by rcu_read_lock, on the write side
+ * by dp_mutex. dp_mutex is almost completely redundant with genl_mutex
+ * maintained by the Generic Netlink code, but the timeout path needs mutual
+ * exclusion too.
*
* It is safe to access the datapath and net_bridge_port structures with just
- * the dp_mutex, but to access the chain you need to take the rcu_read_lock
- * also (because dp_mutex doesn't prevent flows from being destroyed).
+ * dp_mutex.
*/
static struct datapath *dps[DP_MAX];
- static DEFINE_MUTEX(dp_mutex);
+ DEFINE_MUTEX(dp_mutex);
static int dp_maint_func(void *data);
static int send_port_status(struct net_bridge_port *p, uint8_t status);
}
/* Creates a new datapath numbered 'dp_idx'. Returns 0 for success or a
- * negative error code.
- *
- * Not called with any locks. */
+ * negative error code. */
static int new_dp(int dp_idx)
{
struct datapath *dp;
if (!try_module_get(THIS_MODULE))
return -ENODEV;
- mutex_lock(&dp_mutex);
- dp = rcu_dereference(dps[dp_idx]);
- if (dp != NULL) {
+ /* Exit early if a datapath with that number already exists. */
+ if (dps[dp_idx]) {
err = -EEXIST;
goto err_unlock;
}
if (IS_ERR(dp->dp_task))
goto err_destroy_chain;
- rcu_assign_pointer(dps[dp_idx], dp);
- mutex_unlock(&dp_mutex);
+ dps[dp_idx] = dp;
return 0;
err_free_dp:
kfree(dp);
err_unlock:
- mutex_unlock(&dp_mutex);
module_put(THIS_MODULE);
return err;
}
- /* Find and return a free port number under 'dp'. Called under dp_mutex. */
+ /* Find and return a free port number under 'dp'. */
static int find_portno(struct datapath *dp)
{
int i;
return p;
}
- /* Called with dp_mutex. */
int add_switch_port(struct datapath *dp, struct net_device *dev)
{
struct net_bridge_port *p;
return 0;
}
- /* Delete 'p' from switch.
- * Called with dp_mutex. */
+ /* Delete 'p' from switch. */
static int del_switch_port(struct net_bridge_port *p)
{
/* First drop references to device. */
return 0;
}
- /* Called with dp_mutex. */
static void del_dp(struct datapath *dp)
{
struct net_bridge_port *p, *n;
}
#else
/* NB: This has only been tested on 2.4.35 */
-
- /* Called without any locks (?) */
static void dp_frame_hook(struct sk_buff *skb)
{
struct net_bridge_port *p = skb->dev->br_port;
if (!info->attrs[DP_GENL_A_DP_IDX])
return -EINVAL;
- mutex_lock(&dp_mutex);
dp = dp_get(nla_get_u32((info->attrs[DP_GENL_A_DP_IDX])));
if (!dp)
err = -ENOENT;
del_dp(dp);
err = 0;
}
- mutex_unlock(&dp_mutex);
return err;
}
return -EINVAL;
/* Get datapath. */
- mutex_lock(&dp_mutex);
dp = dp_get(nla_get_u32(info->attrs[DP_GENL_A_DP_IDX]));
if (!dp) {
err = -ENOENT;
out_put:
dev_put(port);
out:
- mutex_unlock(&dp_mutex);
return err;
}
if (!info->attrs[DP_GENL_A_DP_IDX] || !va)
return -EINVAL;
- rcu_read_lock();
dp = dp_get(nla_get_u32(info->attrs[DP_GENL_A_DP_IDX]));
- if (!dp) {
- err = -ENOENT;
- goto out;
- }
+ if (!dp)
+ return -ENOENT;
- if (nla_len(va) < sizeof(struct ofp_header)) {
- err = -EINVAL;
- goto out;
- }
+ if (nla_len(va) < sizeof(struct ofp_header))
+ return -EINVAL;
oh = nla_data(va);
sender.xid = oh->xid;
sender.pid = info->snd_pid;
sender.seq = info->snd_seq;
- err = fwd_control_input(dp->chain, &sender, nla_data(va), nla_len(va));
- out:
- rcu_read_unlock();
+ mutex_lock(&dp_mutex);
+ err = fwd_control_input(dp->chain, &sender,
+ nla_data(va), nla_len(va));
+ mutex_unlock(&dp_mutex);
return err;
}
* struct genl_ops. This kluge supports earlier versions also. */
cb->done = dp_genl_openflow_done;
- rcu_read_lock();
if (!cb->args[0]) {
struct nlattr *attrs[DP_GENL_A_MAX + 1];
struct ofp_stats_request *rq;
if (err < 0)
return err;
- err = -EINVAL;
-
if (!attrs[DP_GENL_A_DP_IDX])
- goto out;
+ return -EINVAL;
dp_idx = nla_get_u16(attrs[DP_GENL_A_DP_IDX]);
dp = dp_get(dp_idx);
- if (!dp) {
- err = -ENOENT;
- goto out;
- }
+ if (!dp)
+ return -ENOENT;
va = attrs[DP_GENL_A_OPENFLOW];
len = nla_len(va);
if (!va || len < sizeof *rq)
- goto out;
+ return -EINVAL;
rq = nla_data(va);
type = ntohs(rq->type);
|| ntohs(rq->header.length) != len
|| type >= ARRAY_SIZE(stats)
|| !stats[type].dump)
- goto out;
+ return -EINVAL;
s = &stats[type];
body_len = len - offsetof(struct ofp_stats_request, body);
if (body_len < s->min_body || body_len > s->max_body)
- goto out;
+ return -EINVAL;
cb->args[0] = 1;
cb->args[1] = dp_idx;
void *state;
err = s->init(dp, rq->body, body_len, &state);
if (err)
- goto out;
+ return err;
cb->args[4] = (long) state;
}
} else if (cb->args[0] == 1) {
s = &stats[cb->args[2]];
dp = dp_get(dp_idx);
- if (!dp) {
- err = -ENOENT;
- goto out;
- }
+ if (!dp)
+ return -ENOENT;
} else {
- err = 0;
- goto out;
+ return 0;
}
sender.xid = cb->args[3];
osr = put_openflow_headers(dp, skb, OFPT_STATS_REPLY, &sender,
&max_openflow_len);
- if (IS_ERR(osr)) {
- err = PTR_ERR(osr);
- goto out;
- }
+ if (IS_ERR(osr))
+ return PTR_ERR(osr);
osr->type = htons(s - stats);
osr->flags = 0;
resize_openflow_skb(skb, &osr->header, max_openflow_len);
err = skb->len;
}
- out:
- rcu_read_unlock();
return err;
}
#include <linux/rcupdate.h>
#include <linux/slab.h>
#include <linux/list.h>
+#include <linux/rculist.h>
#include <linux/delay.h>
#include <linux/if_arp.h>
#define DUMMY_MAX_FLOW 8192
- /* xxx Explain need for this separate list because of RCU */
- static spinlock_t pending_free_lock;
- static struct list_head pending_free_list;
-
/* sw_flow private data for dummy table entries. */
struct sw_flow_dummy {
struct list_head node;
struct sw_table_dummy {
struct sw_table swt;
- spinlock_t lock;
unsigned int max_flows;
- atomic_t n_flows;
+ unsigned int n_flows;
struct list_head flows;
struct list_head iter_flows;
unsigned long int next_serial;
};
- static void table_dummy_sfw_destroy(struct sw_flow_dummy *sfw)
- {
- /* xxx Remove the entry from hardware. If you need to do any other
- * xxx clean-up associated with the entry, do it here.
- */
-
- kfree(sfw);
- }
-
- static void table_dummy_rcu_callback(struct rcu_head *rcu)
- {
- struct sw_flow *flow = container_of(rcu, struct sw_flow, rcu);
-
- spin_lock(&pending_free_lock);
- if (flow->private) {
- struct sw_flow_dummy *sfw = flow->private;
- list_add(&sfw->node, &pending_free_list);
- flow->private = NULL;
- }
- spin_unlock(&pending_free_lock);
- flow_free(flow);
- }
-
- static void table_dummy_flow_deferred_free(struct sw_flow *flow)
- {
- call_rcu(&flow->rcu, table_dummy_rcu_callback);
- }
-
static struct sw_flow *table_dummy_lookup(struct sw_table *swt,
const struct sw_flow_key *key)
{
/* xxx Do whatever needs to be done to insert an entry in hardware.
* xxx If the entry can't be inserted, return 0. This stub code
* xxx doesn't do anything yet, so we're going to return 0...you
- * xxx shouldn't.
+ * xxx shouldn't (and you should update n_flows in struct
+ * xxx sw_table_dummy, too).
*/
kfree(flow->private);
return 0;
static int do_delete(struct sw_table *swt, struct sw_flow *flow)
{
- if (flow_del(flow)) {
- list_del_rcu(&flow->node);
- list_del_rcu(&flow->iter_node);
- table_dummy_flow_deferred_free(flow);
- return 1;
- }
- return 0;
+ /* xxx Remove the entry from hardware. If you need to do any other
+ * xxx clean-up associated with the entry, do it here.
+ */
+ list_del_rcu(&flow->node);
+ list_del_rcu(&flow->iter_node);
+ return 1;
}
static int table_dummy_delete(struct sw_table *swt,
struct sw_flow *flow;
unsigned int count = 0;
- list_for_each_entry_rcu (flow, &td->flows, node) {
+ list_for_each_entry (flow, &td->flows, node) {
if (flow_del_matches(&flow->key, key, strict)
&& (!strict || (flow->priority == priority)))
count += do_delete(swt, flow);
}
- if (count)
- atomic_sub(count, &td->n_flows);
+ td->n_flows -= count;
return count;
}
{
struct sw_table_dummy *td = (struct sw_table_dummy *) swt;
struct sw_flow *flow;
- struct sw_flow_dummy *sfw, *n;
int del_count = 0;
uint64_t packet_count = 0;
int i = 0;
- list_for_each_entry_rcu (flow, &td->flows, node) {
+ mutex_lock(&dp_mutex);
+ list_for_each_entry (flow, &td->flows, node) {
/* xxx Retrieve the packet count associated with this entry
* xxx and store it in "packet_count".
*/
}
del_count += do_delete(swt, flow);
}
- if ((i % 50) == 0) {
- msleep_interruptible(1);
- }
i++;
}
+ mutex_unlock(&dp_mutex);
- /* Remove any entries queued for removal */
- spin_lock_bh(&pending_free_lock);
- list_for_each_entry_safe (sfw, n, &pending_free_list, node) {
- list_del(&sfw->node);
- table_dummy_sfw_destroy(sfw);
- }
- spin_unlock_bh(&pending_free_lock);
-
- if (del_count)
- atomic_sub(del_count, &td->n_flows);
+ td->n_flows -= del_count;
return del_count;
}
unsigned long start;
start = ~position->private[0];
- list_for_each_entry_rcu (flow, &tl->iter_flows, iter_node) {
+ list_for_each_entry (flow, &tl->iter_flows, iter_node) {
if (flow->serial <= start && flow_matches(key, &flow->key)) {
int error = callback(flow, private);
if (error) {
{
struct sw_table_dummy *td = (struct sw_table_dummy *) swt;
stats->name = "dummy";
- stats->n_flows = atomic_read(&td->n_flows);
+ stats->n_flows = td->n_flows;
stats->max_flows = td->max_flows;
}
swt->stats = table_dummy_stats;
td->max_flows = DUMMY_MAX_FLOW;
- atomic_set(&td->n_flows, 0);
+ td->n_flows = 0;
INIT_LIST_HEAD(&td->flows);
INIT_LIST_HEAD(&td->iter_flows);
- spin_lock_init(&td->lock);
td->next_serial = 0;
- INIT_LIST_HEAD(&pending_free_list);
- spin_lock_init(&pending_free_lock);
-
return swt;
}
#include <linux/rcupdate.h>
#include <linux/slab.h>
-#include <linux/list.h>
+#include <linux/rculist.h>
struct sw_table_linear {
struct sw_table swt;
- spinlock_t lock;
unsigned int max_flows;
- atomic_t n_flows;
+ unsigned int n_flows;
struct list_head flows;
struct list_head iter_flows;
unsigned long int next_serial;
static int table_linear_insert(struct sw_table *swt, struct sw_flow *flow)
{
struct sw_table_linear *tl = (struct sw_table_linear *) swt;
- unsigned long int flags;
struct sw_flow *f;
* always be placed behind those with equal priority. Just replace
* any flows that match exactly.
*/
- spin_lock_irqsave(&tl->lock, flags);
- list_for_each_entry_rcu (f, &tl->flows, node) {
+ list_for_each_entry (f, &tl->flows, node) {
if (f->priority == flow->priority
&& f->key.wildcards == flow->key.wildcards
- && flow_matches(&f->key, &flow->key)
- && flow_del(f)) {
+ && flow_matches(&f->key, &flow->key)) {
flow->serial = f->serial;
list_replace_rcu(&f->node, &flow->node);
list_replace_rcu(&f->iter_node, &flow->iter_node);
- spin_unlock_irqrestore(&tl->lock, flags);
flow_deferred_free(f);
return 1;
}
}
/* Make sure there's room in the table. */
- if (atomic_read(&tl->n_flows) >= tl->max_flows) {
- spin_unlock_irqrestore(&tl->lock, flags);
+ if (tl->n_flows >= tl->max_flows) {
return 0;
}
- atomic_inc(&tl->n_flows);
+ tl->n_flows++;
/* Insert the entry immediately in front of where we're pointing. */
flow->serial = tl->next_serial++;
list_add_tail_rcu(&flow->node, &f->node);
list_add_rcu(&flow->iter_node, &tl->iter_flows);
- spin_unlock_irqrestore(&tl->lock, flags);
return 1;
}
static int do_delete(struct sw_table *swt, struct sw_flow *flow)
{
- if (flow_del(flow)) {
- list_del_rcu(&flow->node);
- list_del_rcu(&flow->iter_node);
- flow_deferred_free(flow);
- return 1;
- }
- return 0;
+ list_del_rcu(&flow->node);
+ list_del_rcu(&flow->iter_node);
+ flow_deferred_free(flow);
+ return 1;
}
static int table_linear_delete(struct sw_table *swt,
struct sw_flow *flow;
unsigned int count = 0;
- list_for_each_entry_rcu (flow, &tl->flows, node) {
+ list_for_each_entry (flow, &tl->flows, node) {
if (flow_del_matches(&flow->key, key, strict)
&& (!strict || (flow->priority == priority)))
count += do_delete(swt, flow);
}
- if (count)
- atomic_sub(count, &tl->n_flows);
+ tl->n_flows -= count;
return count;
}
struct sw_flow *flow;
int count = 0;
- list_for_each_entry_rcu (flow, &tl->flows, node) {
+ mutex_lock(&dp_mutex);
+ list_for_each_entry (flow, &tl->flows, node) {
if (flow_timeout(flow)) {
count += do_delete(swt, flow);
if (dp->flags & OFPC_SEND_FLOW_EXP)
dp_send_flow_expired(dp, flow);
}
}
- if (count)
- atomic_sub(count, &tl->n_flows);
+ tl->n_flows -= count;
+ mutex_unlock(&dp_mutex);
return count;
}
unsigned long start;
start = position->private[0];
- list_for_each_entry_rcu (flow, &tl->iter_flows, iter_node) {
+ list_for_each_entry (flow, &tl->iter_flows, iter_node) {
if (flow->serial >= start && flow_matches(key, &flow->key)) {
int error = callback(flow, private);
if (error) {
{
struct sw_table_linear *tl = (struct sw_table_linear *) swt;
stats->name = "linear";
- stats->n_flows = atomic_read(&tl->n_flows);
+ stats->n_flows = tl->n_flows;
stats->max_flows = tl->max_flows;
}
swt->stats = table_linear_stats;
tl->max_flows = max_flows;
- atomic_set(&tl->n_flows, 0);
+ tl->n_flows = 0;
INIT_LIST_HEAD(&tl->flows);
INIT_LIST_HEAD(&tl->iter_flows);
- spin_lock_init(&tl->lock);
tl->next_serial = 0;
return swt;