2
0
mirror of https://github.com/openvswitch/ovs synced 2025-08-22 01:51:26 +00:00
ovs/lib/conntrack-tp.c
Peng He 1116459b3b conntrack: Remove nat_conn introducing key directionality.
The patch avoids the extra allocation for nat_conn.
Currently, when doing NAT, the userspace conntrack will use an extra
conn for the two directions in a flow. However, each conn has actually
the two keys for both orig and rev directions. This patch introduces a
key_node[CT_DIRS] member as per Aaron's suggestion in the conn which
consists of a key, direction, and a cmap_node for hash lookup so
addressing the feedback received by the original patch [0].

With this adjustment, we also remove the assertion that connections in
the table are DEFAULT while updating connection state and/or removing
connections.

[0] https://patchwork.ozlabs.org/project/openvswitch/patch/20201129033255.64647-2-hepeng.0320@bytedance.com/

Reported-by: Michael Plato <michael.plato@tu-berlin.de>
Reported-at: https://mail.openvswitch.org/pipermail/ovs-discuss/2022-September/052065.html
Signed-off-by: Peng He <hepeng.0320@bytedance.com>
Co-authored-by: Paolo Valerio <pvalerio@redhat.com>
Signed-off-by: Paolo Valerio <pvalerio@redhat.com>
Tested-by: Frode Nordahl <frode.nordahl@canonical.com>
Acked-by: Ilya Maximets <i.maximets@ovn.org>
Acked-by: Aaron Conole <aconole@redhat.com>
Signed-off-by: Aaron Conole <aconole@redhat.com>
2023-08-31 13:41:08 -04:00

282 lines
7.5 KiB
C

/*
* Copyright (c) 2020 VMware, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at:
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <config.h>
#include <errno.h>
#include "conntrack-private.h"
#include "conntrack-tp.h"
#include "ct-dpif.h"
#include "openvswitch/vlog.h"
VLOG_DEFINE_THIS_MODULE(conntrack_tp);
static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 5);
static const char *ct_timeout_str[] = {
#define CT_TIMEOUT(NAME) #NAME,
CT_TIMEOUTS
#undef CT_TIMEOUT
};
/* Default timeout policy in seconds. */
static unsigned int ct_dpif_netdev_tp_def[] = {
[CT_DPIF_TP_ATTR_TCP_SYN_SENT] = 30,
[CT_DPIF_TP_ATTR_TCP_SYN_RECV] = 30,
[CT_DPIF_TP_ATTR_TCP_ESTABLISHED] = 24 * 60 * 60,
[CT_DPIF_TP_ATTR_TCP_FIN_WAIT] = 15 * 60,
[CT_DPIF_TP_ATTR_TCP_TIME_WAIT] = 45,
[CT_DPIF_TP_ATTR_TCP_CLOSE] = 30,
[CT_DPIF_TP_ATTR_UDP_FIRST] = 60,
[CT_DPIF_TP_ATTR_UDP_SINGLE] = 60,
[CT_DPIF_TP_ATTR_UDP_MULTIPLE] = 30,
[CT_DPIF_TP_ATTR_ICMP_FIRST] = 60,
[CT_DPIF_TP_ATTR_ICMP_REPLY] = 30,
};
static struct timeout_policy *
timeout_policy_lookup_protected(struct conntrack *ct, int32_t tp_id)
OVS_REQUIRES(ct->ct_lock)
{
struct timeout_policy *tp;
uint32_t hash;
hash = hash_int(tp_id, ct->hash_basis);
CMAP_FOR_EACH_WITH_HASH_PROTECTED (tp, node, hash,
&ct->timeout_policies) {
if (tp->policy.id == tp_id) {
return tp;
}
}
return NULL;
}
static struct timeout_policy *
timeout_policy_lookup(struct conntrack *ct, int32_t tp_id)
{
struct timeout_policy *tp;
uint32_t hash;
hash = hash_int(tp_id, ct->hash_basis);
CMAP_FOR_EACH_WITH_HASH (tp, node, hash, &ct->timeout_policies) {
if (tp->policy.id == tp_id) {
return tp;
}
}
return NULL;
}
struct timeout_policy *
timeout_policy_get(struct conntrack *ct, int32_t tp_id)
{
return timeout_policy_lookup(ct, tp_id);
}
static void
update_existing_tp(struct timeout_policy *tp_dst,
const struct timeout_policy *tp_src)
{
struct ct_dpif_timeout_policy *dst;
const struct ct_dpif_timeout_policy *src;
int i;
dst = &tp_dst->policy;
src = &tp_src->policy;
/* Set the value and present bit to dst if present
* bit in src is set.
*/
for (i = 0; i < ARRAY_SIZE(dst->attrs); i++) {
if (src->present & (1 << i)) {
dst->attrs[i] = src->attrs[i];
dst->present |= (1 << i);
}
}
}
static void
init_default_tp(struct timeout_policy *tp, uint32_t tp_id)
{
tp->policy.id = tp_id;
/* Initialize the timeout value to default, but not
* setting the present bit.
*/
tp->policy.present = 0;
memcpy(tp->policy.attrs, ct_dpif_netdev_tp_def,
sizeof tp->policy.attrs);
}
static void
timeout_policy_create(struct conntrack *ct,
struct timeout_policy *new_tp)
OVS_REQUIRES(ct->ct_lock)
{
uint32_t tp_id = new_tp->policy.id;
struct timeout_policy *tp;
uint32_t hash;
tp = xzalloc(sizeof *tp);
init_default_tp(tp, tp_id);
update_existing_tp(tp, new_tp);
hash = hash_int(tp_id, ct->hash_basis);
cmap_insert(&ct->timeout_policies, &tp->node, hash);
}
static void
timeout_policy_clean(struct conntrack *ct, struct timeout_policy *tp)
OVS_REQUIRES(ct->ct_lock)
{
uint32_t hash = hash_int(tp->policy.id, ct->hash_basis);
cmap_remove(&ct->timeout_policies, &tp->node, hash);
ovsrcu_postpone(free, tp);
}
static int
timeout_policy_delete__(struct conntrack *ct, uint32_t tp_id,
bool warn_on_error)
OVS_REQUIRES(ct->ct_lock)
{
struct timeout_policy *tp;
int err = 0;
tp = timeout_policy_lookup_protected(ct, tp_id);
if (tp) {
timeout_policy_clean(ct, tp);
} else if (warn_on_error) {
VLOG_WARN_RL(&rl, "Failed to delete a non-existent timeout "
"policy: id=%d", tp_id);
err = ENOENT;
}
return err;
}
int
timeout_policy_delete(struct conntrack *ct, uint32_t tp_id)
{
int err;
ovs_mutex_lock(&ct->ct_lock);
err = timeout_policy_delete__(ct, tp_id, true);
ovs_mutex_unlock(&ct->ct_lock);
return err;
}
void
timeout_policy_init(struct conntrack *ct)
OVS_REQUIRES(ct->ct_lock)
{
struct timeout_policy tp;
cmap_init(&ct->timeout_policies);
/* Create default timeout policy. */
memset(&tp, 0, sizeof tp);
tp.policy.id = DEFAULT_TP_ID;
timeout_policy_create(ct, &tp);
}
int
timeout_policy_update(struct conntrack *ct,
struct timeout_policy *new_tp)
{
uint32_t tp_id = new_tp->policy.id;
int err = 0;
ovs_mutex_lock(&ct->ct_lock);
timeout_policy_delete__(ct, tp_id, false);
timeout_policy_create(ct, new_tp);
ovs_mutex_unlock(&ct->ct_lock);
return err;
}
static enum ct_dpif_tp_attr
tm_to_ct_dpif_tp(enum ct_timeout tm)
{
switch (tm) {
case CT_TM_TCP_FIRST_PACKET:
return CT_DPIF_TP_ATTR_TCP_SYN_SENT;
case CT_TM_TCP_OPENING:
return CT_DPIF_TP_ATTR_TCP_SYN_RECV;
case CT_TM_TCP_ESTABLISHED:
return CT_DPIF_TP_ATTR_TCP_ESTABLISHED;
case CT_TM_TCP_CLOSING:
return CT_DPIF_TP_ATTR_TCP_FIN_WAIT;
case CT_TM_TCP_FIN_WAIT:
return CT_DPIF_TP_ATTR_TCP_TIME_WAIT;
case CT_TM_TCP_CLOSED:
return CT_DPIF_TP_ATTR_TCP_CLOSE;
case CT_TM_OTHER_FIRST:
return CT_DPIF_TP_ATTR_UDP_FIRST;
case CT_TM_OTHER_BIDIR:
return CT_DPIF_TP_ATTR_UDP_MULTIPLE;
case CT_TM_OTHER_MULTIPLE:
return CT_DPIF_TP_ATTR_UDP_SINGLE;
case CT_TM_ICMP_FIRST:
return CT_DPIF_TP_ATTR_ICMP_FIRST;
case CT_TM_ICMP_REPLY:
return CT_DPIF_TP_ATTR_ICMP_REPLY;
case N_CT_TM:
default:
OVS_NOT_REACHED();
break;
}
OVS_NOT_REACHED();
return CT_DPIF_TP_ATTR_MAX;
}
/* The conn entry lock must be held on entry and exit. */
void
conn_update_expiration(struct conntrack *ct, struct conn *conn,
enum ct_timeout tm, long long now)
OVS_REQUIRES(conn->lock)
{
struct timeout_policy *tp;
uint32_t val;
tp = timeout_policy_lookup(ct, conn->tp_id);
if (tp) {
val = tp->policy.attrs[tm_to_ct_dpif_tp(tm)];
} else {
val = ct_dpif_netdev_tp_def[tm_to_ct_dpif_tp(tm)];
}
VLOG_DBG_RL(&rl, "Update timeout %s zone=%u with policy id=%d "
"val=%u sec.",
ct_timeout_str[tm], conn->key_node[CT_DIR_FWD].key.zone,
conn->tp_id, val);
atomic_store_relaxed(&conn->expiration, now + val * 1000);
}
void
conn_init_expiration(struct conntrack *ct, struct conn *conn,
enum ct_timeout tm, long long now)
{
struct timeout_policy *tp;
uint32_t val;
tp = timeout_policy_lookup(ct, conn->tp_id);
if (tp) {
val = tp->policy.attrs[tm_to_ct_dpif_tp(tm)];
} else {
val = ct_dpif_netdev_tp_def[tm_to_ct_dpif_tp(tm)];
}
VLOG_DBG_RL(&rl, "Init timeout %s zone=%u with policy id=%d val=%u sec.",
ct_timeout_str[tm], conn->key_node[CT_DIR_FWD].key.zone,
conn->tp_id, val);
conn->expiration = now + val * 1000;
}