[ovs-dev] [patch v2 2/5] conntrack: Add rcu support.
Darrell Ball
dlu998 at gmail.com
Thu Nov 29 04:42:52 UTC 2018
Thanks for looking Aaron
Darrell
On Wed, Nov 28, 2018 at 1:34 PM Aaron Conole <aconole at redhat.com> wrote:
> Hi Darrell,
>
> Finally a moment to look at things, yay! Quick inline comments.
>
> Darrell Ball <dlu998 at gmail.com> writes:
>
> > For performance and code simplification reasons, add rcu support for
> > conntrack. The array of hmaps is replaced by a cmap as part of this
> > conversion. Using a single map also simplifies the handling of NAT
> > and allows the removal of the nat_conn map and friends. Per connection
> > entry locks are introduced, which are needed in a few code paths.
> > A subsequent patch will move the connection entry lock to the protocol
> > specific layer.
> >
> > Signed-off-by: Darrell Ball <dlu998 at gmail.com>
> > ---
> > lib/conntrack-icmp.c | 23 +-
> > lib/conntrack-other.c | 13 +-
> > lib/conntrack-private.h | 120 +++---
> > lib/conntrack-tcp.c | 21 +-
> > lib/conntrack.c | 964
> +++++++++++++++++++-----------------------------
> > lib/conntrack.h | 106 +-----
> > 6 files changed, 471 insertions(+), 776 deletions(-)
>
> Statistics look good.
>
> >
> > diff --git a/lib/conntrack-icmp.c b/lib/conntrack-icmp.c
> > index 40fd1d8..fd10985 100644
> > --- a/lib/conntrack-icmp.c
> > +++ b/lib/conntrack-icmp.c
> > @@ -1,5 +1,5 @@
> > /*
> > - * Copyright (c) 2015, 2016 Nicira, Inc.
> > + * Copyright (c) 2015-2018 Nicira, Inc.
> > *
> > * Licensed under the Apache License, Version 2.0 (the "License");
> > * you may not use this file except in compliance with the License.
> > @@ -46,16 +46,13 @@ conn_icmp_cast(const struct conn *conn)
> > }
> >
> > static enum ct_update_res
> > -icmp_conn_update(struct conn *conn_, struct conntrack_bucket *ctb,
> > - struct dp_packet *pkt OVS_UNUSED, bool reply, long
> long now)
> > +icmp_conn_update(struct conn *conn_, struct dp_packet *pkt OVS_UNUSED,
> > + bool reply, long long now)
> > {
> > struct conn_icmp *conn = conn_icmp_cast(conn_);
> >
> > - if (reply && conn->state != ICMPS_REPLY) {
> > - conn->state = ICMPS_REPLY;
> > - }
> > -
> > - conn_update_expiration(ctb, &conn->up, icmp_timeouts[conn->state],
> now);
> > + conn->state = reply ? ICMPS_REPLY : ICMPS_FIRST;
> > + conn_update_expiration(&conn->up, icmp_timeouts[conn->state], now);
> >
> > return CT_UPDATE_VALID;
> > }
> > @@ -79,15 +76,11 @@ icmp6_valid_new(struct dp_packet *pkt)
> > }
> >
> > static struct conn *
> > -icmp_new_conn(struct conntrack_bucket *ctb, struct dp_packet *pkt
> OVS_UNUSED,
> > - long long now)
> > +icmp_new_conn(struct dp_packet *pkt OVS_UNUSED, long long now)
> > {
> > - struct conn_icmp *conn;
> > -
> > - conn = xzalloc(sizeof *conn);
> > + struct conn_icmp *conn = xzalloc(sizeof *conn);
> > conn->state = ICMPS_FIRST;
> > -
> > - conn_init_expiration(ctb, &conn->up, icmp_timeouts[conn->state],
> now);
> > + conn_init_expiration(&conn->up, icmp_timeouts[conn->state], now);
> >
> > return &conn->up;
> > }
> > diff --git a/lib/conntrack-other.c b/lib/conntrack-other.c
> > index 2920889..813be88 100644
> > --- a/lib/conntrack-other.c
> > +++ b/lib/conntrack-other.c
> > @@ -1,5 +1,5 @@
> > /*
> > - * Copyright (c) 2015, 2016 Nicira, Inc.
> > + * Copyright (c) 2015-2018 Nicira, Inc.
> > *
> > * Licensed under the Apache License, Version 2.0 (the "License");
> > * you may not use this file except in compliance with the License.
> > @@ -43,8 +43,8 @@ conn_other_cast(const struct conn *conn)
> > }
> >
> > static enum ct_update_res
> > -other_conn_update(struct conn *conn_, struct conntrack_bucket *ctb,
> > - struct dp_packet *pkt OVS_UNUSED, bool reply, long
> long now)
> > +other_conn_update(struct conn *conn_, struct dp_packet *pkt OVS_UNUSED,
> > + bool reply, long long now)
> > {
> > struct conn_other *conn = conn_other_cast(conn_);
> >
> > @@ -54,7 +54,7 @@ other_conn_update(struct conn *conn_, struct
> conntrack_bucket *ctb,
> > conn->state = OTHERS_MULTIPLE;
> > }
> >
> > - conn_update_expiration(ctb, &conn->up, other_timeouts[conn->state],
> now);
> > + conn_update_expiration(&conn->up, other_timeouts[conn->state], now);
> >
> > return CT_UPDATE_VALID;
> > }
> > @@ -66,15 +66,14 @@ other_valid_new(struct dp_packet *pkt OVS_UNUSED)
> > }
> >
> > static struct conn *
> > -other_new_conn(struct conntrack_bucket *ctb, struct dp_packet *pkt
> OVS_UNUSED,
> > - long long now)
> > +other_new_conn(struct dp_packet *pkt OVS_UNUSED, long long now)
> > {
> > struct conn_other *conn;
> >
> > conn = xzalloc(sizeof *conn);
> > conn->state = OTHERS_FIRST;
> >
> > - conn_init_expiration(ctb, &conn->up, other_timeouts[conn->state],
> now);
> > + conn_init_expiration(&conn->up, other_timeouts[conn->state], now);
> >
> > return &conn->up;
> > }
> > diff --git a/lib/conntrack-private.h b/lib/conntrack-private.h
> > index 27ece38..3d838e4 100644
> > --- a/lib/conntrack-private.h
> > +++ b/lib/conntrack-private.h
> > @@ -21,6 +21,7 @@
> > #include <netinet/in.h>
> > #include <netinet/ip6.h>
> >
> > +#include "cmap.h"
> > #include "conntrack.h"
> > #include "ct-dpif.h"
> > #include "openvswitch/hmap.h"
> > @@ -51,18 +52,11 @@ BUILD_ASSERT_DECL(sizeof(struct ct_endpoint) ==
> sizeof(struct ct_addr) + 4);
> > struct conn_key {
> > struct ct_endpoint src;
> > struct ct_endpoint dst;
> > -
> > ovs_be16 dl_type;
> > uint16_t zone;
> > uint8_t nw_proto;
> > };
> >
> > -struct nat_conn_key_node {
> > - struct hmap_node node;
> > - struct conn_key key;
> > - struct conn_key value;
> > -};
> > -
> > /* This is used for alg expectations; an expectation is a
> > * context created in preparation for establishing a data
> > * connection. The expectation is created by the control
> > @@ -87,27 +81,43 @@ struct alg_exp_node {
> > bool nat_rpl_dst;
> > };
> >
> > +struct OVS_LOCKABLE ct_ce_lock {
> > + struct ovs_mutex lock;
> > +};
> > +
> > struct conn {
> > struct conn_key key;
> > struct conn_key rev_key;
> > /* Only used for orig_tuple support. */
> > struct conn_key master_key;
> > + struct ct_ce_lock lock;
> > long long expiration;
> > struct ovs_list exp_node;
> > - struct hmap_node node;
> > + struct cmap_node cm_node;
> > ovs_u128 label;
> > - /* XXX: consider flattening. */
> > struct nat_action_info_t *nat_info;
> > char *alg;
> > + struct conn *nat_conn;
> > int seq_skew;
> > uint32_t mark;
> > + /* See ct_conn_type. */
> > uint8_t conn_type;
> > - /* TCP sequence skew due to NATTing of FTP control messages. */
> > - uint8_t seq_skew_dir;
> > + /* Update expiry list id of which there are 'N_CT_TM' possible
> values.
> > + * This field is used to signal an update to the specified list.
> The
> > + * value 'NO_UPD_EXP_LIST' is used to indicate no update to any
> list. */
> > + uint8_t exp_list_id;
> > + /* TCP sequence skew direction due to NATTing of FTP control
> messages;
> > + * true if reply direction. */
> > + bool seq_skew_dir;
> > /* True if alg data connection. */
> > - uint8_t alg_related;
> > + bool alg_related;
> > + /* Inserted into the cmap; handle theoretical expiry list race;
> although
> > + * such a race would probably mean a system meltdown. */
> > + bool inserted;
> > };
> >
> > +#define NO_UPD_EXP_LIST 255
> > +
> > enum ct_update_res {
> > CT_UPDATE_INVALID,
> > CT_UPDATE_VALID,
> > @@ -119,68 +129,70 @@ enum ct_conn_type {
> > CT_CONN_TYPE_UN_NAT,
> > };
> >
> > -/* Locking:
> > - *
> > - * The connections are kept in different buckets, which are completely
> > - * independent. The connection bucket is determined by the hash of its
> key.
> > - *
> > - * Each bucket has two locks. Acquisition order is, from outermost to
> > - * innermost:
> > - *
> > - * cleanup_mutex
> > - * lock
> > - *
> > - * */
> > -struct conntrack_bucket {
> > - /* Protects 'connections' and 'exp_lists'. Used in the fast path */
> > - struct ct_lock lock;
> > - /* Contains the connections in the bucket, indexed by 'struct
> conn_key' */
> > - struct hmap connections OVS_GUARDED;
> > - /* For each possible timeout we have a list of connections. When the
> > - * timeout of a connection is updated, we move it to the back of
> the list.
> > - * Since the connection in a list have the same relative timeout,
> the list
> > - * will be ordered, with the oldest connections to the front. */
> > - struct ovs_list exp_lists[N_CT_TM] OVS_GUARDED;
> > -
> > - /* Protects 'next_cleanup'. Used to make sure that there's only one
> thread
> > - * performing the cleanup. */
> > - struct ovs_mutex cleanup_mutex;
> > - long long next_cleanup OVS_GUARDED;
> > -};
> > +extern struct ct_l4_proto ct_proto_tcp;
> > +extern struct ct_l4_proto ct_proto_other;
> > +extern struct ct_l4_proto ct_proto_icmp4;
> > +extern struct ct_l4_proto ct_proto_icmp6;
> >
> > struct ct_l4_proto {
> > - struct conn *(*new_conn)(struct conntrack_bucket *, struct
> dp_packet *pkt,
> > - long long now);
> > + struct conn *(*new_conn)(struct dp_packet *pkt, long long now);
> > bool (*valid_new)(struct dp_packet *pkt);
> > enum ct_update_res (*conn_update)(struct conn *conn,
> > - struct conntrack_bucket *,
> > struct dp_packet *pkt, bool reply,
> > long long now);
> > void (*conn_get_protoinfo)(const struct conn *,
> > struct ct_dpif_protoinfo *);
> > };
> >
> > -extern struct ct_l4_proto ct_proto_tcp;
> > -extern struct ct_l4_proto ct_proto_other;
> > -extern struct ct_l4_proto ct_proto_icmp4;
> > -extern struct ct_l4_proto ct_proto_icmp6;
> > +/* Timeouts: all the possible timeout states passed to
> update_expiration()
> > + * are listed here. The name will be prefix by CT_TM_ and the value is
> in
> > + * milliseconds */
> > +#define CT_TIMEOUTS \
> > + CT_TIMEOUT(TCP_FIRST_PACKET, 30 * 1000) \
> > + CT_TIMEOUT(TCP_OPENING, 30 * 1000) \
> > + CT_TIMEOUT(TCP_ESTABLISHED, 24 * 60 * 60 * 1000) \
> > + CT_TIMEOUT(TCP_CLOSING, 15 * 60 * 1000) \
> > + CT_TIMEOUT(TCP_FIN_WAIT, 45 * 1000) \
> > + CT_TIMEOUT(TCP_CLOSED, 30 * 1000) \
> > + CT_TIMEOUT(OTHER_FIRST, 60 * 1000) \
> > + CT_TIMEOUT(OTHER_MULTIPLE, 60 * 1000) \
> > + CT_TIMEOUT(OTHER_BIDIR, 30 * 1000) \
> > + CT_TIMEOUT(ICMP_FIRST, 60 * 1000) \
> > + CT_TIMEOUT(ICMP_REPLY, 30 * 1000)
> > +
> > +/* The smallest of the above values: it is used as an upper bound for
> the
> > + * interval between two rounds of cleanup of expired entries */
> > +#define CT_TM_MIN (30 * 1000)
> > +
> > +#define CT_TIMEOUT(NAME, VAL) BUILD_ASSERT_DECL(VAL >= CT_TM_MIN);
> > + CT_TIMEOUTS
> > +#undef CT_TIMEOUT
> > +
> > +enum ct_timeout {
> > +#define CT_TIMEOUT(NAME, VALUE) CT_TM_##NAME,
> > + CT_TIMEOUTS
> > +#undef CT_TIMEOUT
> > + N_CT_TM
> > +};
> >
> > extern long long ct_timeout_val[];
> > +extern struct ovs_list cm_exp_lists[N_CT_TM];
> >
> > +/* ct_lock must be held. */
> > static inline void
> > -conn_init_expiration(struct conntrack_bucket *ctb, struct conn *conn,
> > - enum ct_timeout tm, long long now)
> > +conn_init_expiration(struct conn *conn, enum ct_timeout tm, long long
> now)
> > {
> > conn->expiration = now + ct_timeout_val[tm];
> > - ovs_list_push_back(&ctb->exp_lists[tm], &conn->exp_node);
> > + conn->exp_list_id = NO_UPD_EXP_LIST;
> > + ovs_list_push_back(&cm_exp_lists[tm], &conn->exp_node);
> > }
> >
> > +/* The conn entry lock must be held. */
> > static inline void
> > -conn_update_expiration(struct conntrack_bucket *ctb, struct conn *conn,
> > - enum ct_timeout tm, long long now)
> > +conn_update_expiration(struct conn *conn, enum ct_timeout tm, long long
> now)
> > {
> > - ovs_list_remove(&conn->exp_node);
> > - conn_init_expiration(ctb, conn, tm, now);
> > + conn->expiration = now + ct_timeout_val[tm];
> > + conn->exp_list_id = tm;
> > }
> >
> > static inline uint32_t
> > diff --git a/lib/conntrack-tcp.c b/lib/conntrack-tcp.c
> > index 86d313d..19fdf1d 100644
> > --- a/lib/conntrack-tcp.c
> > +++ b/lib/conntrack-tcp.c
> > @@ -145,8 +145,8 @@ tcp_get_wscale(const struct tcp_header *tcp)
> > }
> >
> > static enum ct_update_res
> > -tcp_conn_update(struct conn *conn_, struct conntrack_bucket *ctb,
> > - struct dp_packet *pkt, bool reply, long long now)
> > +tcp_conn_update(struct conn *conn_, struct dp_packet *pkt, bool reply,
> > + long long now)
> > {
> > struct conn_tcp *conn = conn_tcp_cast(conn_);
> > struct tcp_header *tcp = dp_packet_l4(pkt);
> > @@ -317,18 +317,18 @@ tcp_conn_update(struct conn *conn_, struct
> conntrack_bucket *ctb,
> >
> > if (src->state >= CT_DPIF_TCPS_FIN_WAIT_2
> > && dst->state >= CT_DPIF_TCPS_FIN_WAIT_2) {
> > - conn_update_expiration(ctb, &conn->up, CT_TM_TCP_CLOSED,
> now);
> > + conn_update_expiration(&conn->up, CT_TM_TCP_CLOSED, now);
> > } else if (src->state >= CT_DPIF_TCPS_CLOSING
> > && dst->state >= CT_DPIF_TCPS_CLOSING) {
> > - conn_update_expiration(ctb, &conn->up, CT_TM_TCP_FIN_WAIT,
> now);
> > + conn_update_expiration(&conn->up, CT_TM_TCP_FIN_WAIT, now);
> > } else if (src->state < CT_DPIF_TCPS_ESTABLISHED
> > || dst->state < CT_DPIF_TCPS_ESTABLISHED) {
> > - conn_update_expiration(ctb, &conn->up, CT_TM_TCP_OPENING,
> now);
> > + conn_update_expiration(&conn->up, CT_TM_TCP_OPENING, now);
> > } else if (src->state >= CT_DPIF_TCPS_CLOSING
> > || dst->state >= CT_DPIF_TCPS_CLOSING) {
> > - conn_update_expiration(ctb, &conn->up, CT_TM_TCP_CLOSING,
> now);
> > + conn_update_expiration(&conn->up, CT_TM_TCP_CLOSING, now);
> > } else {
> > - conn_update_expiration(ctb, &conn->up,
> CT_TM_TCP_ESTABLISHED, now);
> > + conn_update_expiration(&conn->up, CT_TM_TCP_ESTABLISHED,
> now);
> > }
> > } else if ((dst->state < CT_DPIF_TCPS_SYN_SENT
> > || dst->state >= CT_DPIF_TCPS_FIN_WAIT_2
> > @@ -412,8 +412,7 @@ tcp_valid_new(struct dp_packet *pkt)
> > }
> >
> > static struct conn *
> > -tcp_new_conn(struct conntrack_bucket *ctb, struct dp_packet *pkt,
> > - long long now)
> > +tcp_new_conn(struct dp_packet *pkt, long long now)
> > {
> > struct conn_tcp* newconn = NULL;
> > struct tcp_header *tcp = dp_packet_l4(pkt);
> > @@ -448,9 +447,7 @@ tcp_new_conn(struct conntrack_bucket *ctb, struct
> dp_packet *pkt,
> > dst->max_win = 1;
> > src->state = CT_DPIF_TCPS_SYN_SENT;
> > dst->state = CT_DPIF_TCPS_CLOSED;
> > -
> > - conn_init_expiration(ctb, &newconn->up, CT_TM_TCP_FIRST_PACKET,
> > - now);
> > + conn_init_expiration(&newconn->up, CT_TM_TCP_FIRST_PACKET, now);
> >
> > return &newconn->up;
> > }
> > diff --git a/lib/conntrack.c b/lib/conntrack.c
> > index 07ab0d0..8eb73a9 100644
> > --- a/lib/conntrack.c
> > +++ b/lib/conntrack.c
> > @@ -76,79 +76,67 @@ enum ct_alg_ctl_type {
> > CT_ALG_CTL_SIP,
> > };
> >
> > -#define CONNTRACK_BUCKETS_SHIFT 8
> > -#define CONNTRACK_BUCKETS (1 << CONNTRACK_BUCKETS_SHIFT)
> > -/* Independent buckets containing the connections */
> > -struct conntrack_bucket buckets[CONNTRACK_BUCKETS];
> > +struct OVS_LOCKABLE ct_rwlock {
> > + struct ovs_rwlock lock;
> > +};
> > +
> > +/* This lock is used to guard alg_expectations and
> > + * alg_expectation_refs. */
> > +static struct ct_rwlock resources_lock;
> > +
> > +/* Hash table for alg expectations. Expectations are created
> > + * by control connections to help create data connections. */
> > +static struct hmap alg_expectations OVS_GUARDED_BY(resources_lock);
> > +/* Only needed to be able to cleanup expectations from non-control
> > + * connection context; otherwise a pointer to the expectation from
> > + * the control connection would suffice. */
> > +static struct hindex alg_expectation_refs
> OVS_GUARDED_BY(resources_lock);
> > +
> > +struct OVS_LOCKABLE ct_lock {
> > + struct ovs_mutex lock;
> > +};
> > +
> > +static struct ct_lock ct_lock;
> > +static struct cmap cm_conns OVS_GUARDED_BY(ct_lock);
> > +struct ovs_list cm_exp_lists[N_CT_TM] OVS_GUARDED_BY(ct_lock);
> > /* Salt for hashing a connection key. */
> > -uint32_t hash_basis;
> > +static uint32_t hash_basis;
> > /* The thread performing periodic cleanup of the connection
> > * tracker */
> > -pthread_t clean_thread;
> > +static pthread_t clean_thread;
> > /* Latch to destroy the 'clean_thread' */
> > -struct latch clean_thread_exit;
> > +static struct latch clean_thread_exit;
> > +
> > /* Number of connections currently in the connection tracker. */
> > -atomic_count n_conn;
> > +static atomic_count n_conn;
> > /* Connections limit. When this limit is reached, no new connection
> > * will be accepted. */
> > -atomic_uint n_conn_limit;
> > -/* The following resources are referenced during nat connection
> > - * creation and deletion. */
> > -struct hmap nat_conn_keys OVS_GUARDED;
> > -/* Hash table for alg expectations. Expectations are created
> > - * by control connections to help create data connections. */
> > -struct hmap alg_expectations OVS_GUARDED;
> > -/* Used to lookup alg expectations from the control context. */
> > -struct hindex alg_expectation_refs OVS_GUARDED;
> > -/* Expiry list for alg expectations. */
> > -struct ovs_list alg_exp_list OVS_GUARDED;
> > -/* This lock is used during NAT connection creation and deletion;
> > - * it is taken after a bucket lock and given back before that
> > - * bucket unlock.
> > - * This lock is similarly used to guard alg_expectations and
> > - * alg_expectation_refs. If a bucket lock is also held during
> > - * the normal code flow, then is must be taken first and released
> > - * last.
> > - */
> > -struct ct_rwlock resources_lock;
> > +static atomic_uint n_conn_limit;
> > +
> > +/* Lock acquisition order: If multiple locks are taken, then the order
> is
> > + * 'ct_lock', then conn entry lock and then 'resources_lock' and release
> > + * happens in the reverse order. */
>
> A bunch of the variables above are added in patch 1/ as nonstatic, not
> used outside this translation unit, and then here made static. Can you
> make them static in patch 1/ so that there's less churn in the series?
>
I spliced the original big patch poorly...
The intention was that Patch 1 should incorporate all the static qualifiers
not this Patch 2.
Thanks for pointing it out; I did another audit and found other splicing
problems.
>
> > static bool conn_key_extract(struct dp_packet *, ovs_be16 dl_type,
> > struct conn_lookup_ctx *, uint16_t zone);
> > static uint32_t conn_key_hash(const struct conn_key *, uint32_t basis);
> > static void conn_key_reverse(struct conn_key *);
> > -static void conn_key_lookup(struct conntrack_bucket *ctb,
> > - struct conn_lookup_ctx *ctx,
> > - long long now);
> > static bool valid_new(struct dp_packet *pkt, struct conn_key *);
> > -static struct conn *new_conn(struct conntrack_bucket *, struct
> dp_packet *pkt,
> > - struct conn_key *, long long now);
> > -static void delete_conn(struct conn *);
> > -static enum ct_update_res conn_update(struct conn *,
> > - struct conntrack_bucket *ctb,
> > - struct dp_packet *, bool reply,
> > +static struct conn *new_conn(struct dp_packet *pkt, struct conn_key *,
> > + long long now);
> > +static enum ct_update_res conn_update(struct dp_packet *pkt,
> > + struct conn *conn,
> > + struct conn_lookup_ctx *ctx,
> > long long now);
> > +static void delete_conn(struct conn *);
> > +static void delete_conn_one(struct conn *conn);
> > static bool conn_expired(struct conn *, long long now);
> > static void set_mark(struct dp_packet *, struct conn *,
> > uint32_t val, uint32_t mask);
> > static void set_label(struct dp_packet *, struct conn *,
> > const struct ovs_key_ct_labels *val,
> > const struct ovs_key_ct_labels *mask);
> > -static void *clean_thread_main(void *f_);
> > -
> > -static struct nat_conn_key_node *
> > -nat_conn_keys_lookup(struct hmap *nat_conn_keys_,
> > - const struct conn_key *key,
> > - uint32_t basis);
> > -
> > -static bool
> > -nat_conn_keys_insert(struct hmap *nat_conn_keys_,
> > - const struct conn *nat_conn,
> > - uint32_t hash_basis);
> > -
> > -static void
> > -nat_conn_keys_remove(struct hmap *nat_conn_keys_,
> > - const struct conn_key *key,
> > - uint32_t basis);
> > +static void *clean_thread_main(void *);
>
> For this 'clean' thread, isn't it possible to just delete it entirely
> and use the ovsrcu_postpone() call to your garbage collection function?
>
No, the postpone thread should not be doing the kind of activities the
clean thread is doing.
Furthermore, they should run in parallel.
>
> > static bool
> > nat_select_range_tuple(const struct conn *conn, struct conn *nat_conn);
> > @@ -184,7 +172,7 @@ detect_ftp_ctl_type(const struct conn_lookup_ctx
> *ctx,
> > struct dp_packet *pkt);
> >
> > static void
> > -expectation_clean(const struct conn_key *master_key, uint32_t basis);
> > +expectation_clean(const struct conn_key *master_key);
> >
> > static struct ct_l4_proto *l4_protos[] = {
> > [IPPROTO_TCP] = &ct_proto_tcp,
> > @@ -279,6 +267,50 @@ conn_key_cmp(const struct conn_key *key1, const
> struct conn_key *key2)
> > }
> >
> > static void
> > +conn_key_lookup(const struct conn_key *key, uint32_t hash, long long
> now,
> > + struct conn **conn_out, bool *reply)
> > +{
> > + struct conn *conn;
> > + *conn_out = NULL;
> > +
> > + CMAP_FOR_EACH_WITH_HASH (conn, cm_node, hash, &cm_conns) {
> > + if (!conn_key_cmp(&conn->key, key) && !conn_expired(conn, now))
> {
> > + *conn_out = conn;
> > + *reply = false;
> > + break;
> > + }
> > + if (!conn_key_cmp(&conn->rev_key, key) && !conn_expired(conn,
> now)) {
> > + *conn_out = conn;
> > + *reply = true;
> > + break;
> > + }
> > + }
> > +}
> > +
> > +static bool
> > +conn_available(const struct conn_key *key, uint32_t hash, long long now)
> > +{
> > + struct conn *conn;
> > + bool found = false;
> > +
> > + CMAP_FOR_EACH_WITH_HASH (conn, cm_node, hash, &cm_conns) {
> > + if (!conn_key_cmp(&conn->key, key)
> > + && !conn_expired(conn, now)) {
> > + found = true;
> > + break;
> > + }
> > +
> > + if (!conn_key_cmp(&conn->rev_key, key)
> > + && !conn_expired(conn, now)) {
> > + found = true;
> > + break;
> > + }
> > + }
> > +
>
> Can't this function be expressed as:
>
> static bool
> conn_available(const struct conn_key *key, uint32_t hash, long long now)
> {
> struct conn *c;
> bool b;
>
> conn_key_lookup(key, hash, now, &c, &b);
> return !c;
> }
>
Your function is fine and thanks for drawing attention to it.
However, conn_available() was only left around for debugging a problem and
I forgot to
to remove it, since it is semantically very similar to conn_key_lookup().
The api is now removed and conn_key_lookup() was modified.
*static* bool
conn_key_lookup(*const* *struct* conn_key *key, uint32_t hash, *long* *long*
now,
*struct* conn **conn_out, bool *reply)
{
*struct* conn *conn;
bool found = false;
CMAP_FOR_EACH_WITH_HASH (conn, cm_node, hash, &cm_conns) {
*if* (!conn_key_cmp(&conn->key, key) && !conn_expired(conn, now)) {
found = true;
*if* (reply) {
*reply = false;
}
*break*;
}
*if* (!conn_key_cmp(&conn->rev_key, key) && !conn_expired(conn,
now)) {
found = true;
*if* (reply) {
*reply = true;
}
*break*;
}
}
*if* (found && conn_out) {
*conn_out = conn;
} *else* *if* (conn_out) {
*conn_out = NULL;
}
*return* found;
}
and the caller to conn_available()
>
> > + return !found;
> > +}
> > +
> > +static void
> > ct_print_conn_info(const struct conn *c, const char *log_msg,
> > enum vlog_level vll, bool force, bool rl_on)
> > {
> > @@ -338,31 +370,20 @@ ct_print_conn_info(const struct conn *c, const
> char *log_msg,
> > void
> > conntrack_init(void)
> > {
> > - long long now = time_msec();
> > -
> > - ct_rwlock_init(&resources_lock);
> > - ct_rwlock_wrlock(&resources_lock);
> > - hmap_init(&nat_conn_keys);
> > + ovs_rwlock_init(&resources_lock.lock);
> > + ovs_rwlock_wrlock(&resources_lock.lock);
> > hmap_init(&alg_expectations);
> > hindex_init(&alg_expectation_refs);
> > - ovs_list_init(&alg_exp_list);
> > - ct_rwlock_unlock(&resources_lock);
> > + ovs_rwlock_unlock(&resources_lock.lock);
> >
> > - for (unsigned i = 0; i < CONNTRACK_BUCKETS; i++) {
> > - struct conntrack_bucket *ctb = &buckets[i];
> > -
> > - ct_lock_init(&ctb->lock);
> > - ct_lock_lock(&ctb->lock);
> > - hmap_init(&ctb->connections);
> > - for (unsigned j = 0; j < ARRAY_SIZE(ctb->exp_lists); j++) {
> > - ovs_list_init(&ctb->exp_lists[j]);
> > - }
> > - ct_lock_unlock(&ctb->lock);
> > - ovs_mutex_init(&ctb->cleanup_mutex);
> > - ovs_mutex_lock(&ctb->cleanup_mutex);
> > - ctb->next_cleanup = now + CT_TM_MIN;
> > - ovs_mutex_unlock(&ctb->cleanup_mutex);
> > + ovs_mutex_init_adaptive(&ct_lock.lock);
> > + ovs_mutex_lock(&ct_lock.lock);
> > + cmap_init(&cm_conns);
> > + for (unsigned i = 0; i < ARRAY_SIZE(cm_exp_lists); i++) {
> > + ovs_list_init(&cm_exp_lists[i]);
> > }
> > + ovs_mutex_unlock(&ct_lock.lock);
> > +
> > hash_basis = random_uint32();
> > atomic_count_init(&n_conn, 0);
> > atomic_init(&n_conn_limit, DEFAULT_N_CONN_LIMIT);
> > @@ -370,56 +391,76 @@ conntrack_init(void)
> > clean_thread = ovs_thread_create("ct_clean", clean_thread_main,
> NULL);
> > }
> >
> > +/* Must be called with 'conn' of 'conn_type' CT_CONN_TYPE_DEFAULT. Also
> > + * removes the associated nat 'conn' from the lookup datastructures. */
> > +static void
> > +conn_clean(struct conn *conn)
> > + OVS_NO_THREAD_SAFETY_ANALYSIS
> > +{
> > + ovs_assert(conn->conn_type == CT_CONN_TYPE_DEFAULT);
> > +
> > + if (conn->alg) {
> > + expectation_clean(&conn->key);
> > + }
> > +
> > + uint32_t hash = conn_key_hash(&conn->key, hash_basis);
> > + cmap_remove(&cm_conns, &conn->cm_node, hash);
> > + ovs_list_remove(&conn->exp_node);
> > + if (conn->nat_conn) {
> > + hash = conn_key_hash(&conn->nat_conn->key, hash_basis);
> > + cmap_remove(&cm_conns, &conn->nat_conn->cm_node, hash);
> > + }
> > + ovsrcu_postpone(delete_conn, conn);
> > + atomic_count_dec(&n_conn);
> > +}
> > +
> > +static void
> > +conn_clean_one(struct conn *conn)
> > + OVS_NO_THREAD_SAFETY_ANALYSIS
> > +{
> > + if (conn->alg) {
> > + expectation_clean(&conn->key);
> > + }
> > +
> > + uint32_t hash = conn_key_hash(&conn->key, hash_basis);
> > + cmap_remove(&cm_conns, &conn->cm_node, hash);
> > + ovs_list_remove(&conn->exp_node);
> > + if (conn->conn_type == CT_CONN_TYPE_DEFAULT) {
> > + atomic_count_dec(&n_conn);
> > + }
> > + ovsrcu_postpone(delete_conn_one, conn);
> > +}
> > +
>
> Why have two different clean up routines here? Wouldn't it make more
> sense to always call conn_clean_one() and change the implementation
> like:
>
conn_clean_one() and associated delete_conn_one() only exist because
of the usages in CMAP_FOR_EACH and I added a comment to make that
clear. Essentially, I don't want to remove the current node and another
node somewhere
else in the cmap in one iteration of the loop.
Although, I want to keep both apis, I spliced out the common code for
conn_clean* in this patch
to conn_clean_cmn(). I also moved the delete_conn_cmn() from Patch 4 to
Patch 2.
static void
conn_clean_cmn(struct conn *conn)
OVS_NO_THREAD_SAFETY_ANALYSIS
{
if (conn->alg) {
expectation_clean(&conn->key);
}
uint32_t hash = conn_key_hash(&conn->key, hash_basis);
cmap_remove(&cm_conns, &conn->cm_node, hash);
ovs_list_remove(&conn->exp_node);
}
/* Must be called with 'conn' of 'conn_type' CT_CONN_TYPE_DEFAULT. Also
* removes the associated nat 'conn' from the lookup datastructures. */
static void
conn_clean(struct conn *conn)
OVS_NO_THREAD_SAFETY_ANALYSIS
{
ovs_assert(conn->conn_type == CT_CONN_TYPE_DEFAULT);
conn_clean_cmn(conn);
if (conn->nat_conn) {
uint32_t hash = conn_key_hash(&conn->nat_conn->key, hash_basis);
cmap_remove(&cm_conns, &conn->nat_conn->cm_node, hash);
}
ovsrcu_postpone(delete_conn, conn);
atomic_count_dec(&n_conn);
}
/* Needed because of usage in CMAP_FOR_EACH. */
static void
conn_clean_one(struct conn *conn)
OVS_NO_THREAD_SAFETY_ANALYSIS
{
conn_clean_cmn(conn);
if (conn->conn_type == CT_CONN_TYPE_DEFAULT) {
atomic_count_dec(&n_conn);
}
ovsrcu_postpone(delete_conn_one, conn);
}
>
> static void conn_clean_one(struct conn *conn)
> OVS_NO_THREAD_SAFETY_ANALYSIS
> {
> if (conn->alg) {
> expectation_clean(&conn->key);
> }
>
> uint32_t hash = conn_key_hash(&conn->key, hash_basis);
> cmap_remove(&cm_conns, &conn->cm_node, hash);
> ovs_list_remove(&conn->exp_node);
> if (conn->conn_type == CT_CONN_TYPE_DEFAULT) {
> hash = conn_key_hash(&conn->nat_conn->key, hash_basis);
> cmap_remove(&cm_conns, &conn->nat_conn->cm_node, hash);
> atomic_count_dec(&n_conn);
> }
> ovsrcu_postpone(delete_conn_one, conn);
> }
>
>
> You can also change around the deletes that appear later.
>
> > /* Destroys the connection tracker 'ct' and frees all the allocated
> memory. */
> > void
> > conntrack_destroy(void)
> > + OVS_NO_THREAD_SAFETY_ANALYSIS
> > {
> > + struct conn *conn;
> > latch_set(&clean_thread_exit);
> > pthread_join(clean_thread, NULL);
> > latch_destroy(&clean_thread_exit);
> > - for (unsigned i = 0; i < CONNTRACK_BUCKETS; i++) {
> > - struct conntrack_bucket *ctb = &buckets[i];
> > - struct conn *conn;
> >
> > - ovs_mutex_destroy(&ctb->cleanup_mutex);
> > - ct_lock_lock(&ctb->lock);
> > - HMAP_FOR_EACH_POP (conn, node, &ctb->connections) {
> > - if (conn->conn_type == CT_CONN_TYPE_DEFAULT) {
> > - atomic_count_dec(&n_conn);
> > - }
> > - delete_conn(conn);
> > - }
> > - hmap_destroy(&ctb->connections);
> > - ct_lock_unlock(&ctb->lock);
> > - ct_lock_destroy(&ctb->lock);
> > + ovs_mutex_lock(&ct_lock.lock);
> > + CMAP_FOR_EACH (conn, cm_node, &cm_conns) {
> > + conn_clean_one(conn);
> > }
> > - ct_rwlock_wrlock(&resources_lock);
> > - struct nat_conn_key_node *nat_conn_key_node;
> > - HMAP_FOR_EACH_POP (nat_conn_key_node, node, &nat_conn_keys) {
> > - free(nat_conn_key_node);
> > - }
> > - hmap_destroy(&nat_conn_keys);
> > + cmap_destroy(&cm_conns);
> > + ovs_mutex_unlock(&ct_lock.lock);
> > + ovs_mutex_destroy(&ct_lock.lock);
> >
> > + ovs_rwlock_wrlock(&resources_lock.lock);
> > struct alg_exp_node *alg_exp_node;
> > HMAP_FOR_EACH_POP (alg_exp_node, node, &alg_expectations) {
> > free(alg_exp_node);
> > }
> >
> > - ovs_list_poison(&alg_exp_list);
> > hmap_destroy(&alg_expectations);
> > hindex_destroy(&alg_expectation_refs);
> > - ct_rwlock_unlock(&resources_lock);
> > - ct_rwlock_destroy(&resources_lock);
> > + ovs_rwlock_unlock(&resources_lock.lock);
> > + ovs_rwlock_destroy(&resources_lock.lock);
> > }
> >
> > -static unsigned hash_to_bucket(uint32_t hash)
> > -{
> > - /* Extracts the most significant bits in hash. The least
> significant bits
> > - * are already used internally by the hmap implementation. */
> > - BUILD_ASSERT(CONNTRACK_BUCKETS_SHIFT < 32 &&
> CONNTRACK_BUCKETS_SHIFT >= 1);
> > -
> > - return (hash >> (32 - CONNTRACK_BUCKETS_SHIFT)) % CONNTRACK_BUCKETS;
> > -}
> >
> > static void
> > write_ct_md(struct dp_packet *pkt, uint16_t zone, const struct conn
> *conn,
> > @@ -544,13 +585,14 @@ alg_src_ip_wc(enum ct_alg_ctl_type alg_ctl_type)
> > static void
> > handle_alg_ctl(const struct conn_lookup_ctx *ctx, struct dp_packet *pkt,
> > enum ct_alg_ctl_type ct_alg_ctl, const struct conn *conn,
> > - long long now, bool nat,
> > - const struct conn *conn_for_expectation)
> > + long long now, bool nat)
> > {
> > /* ALG control packet handling with expectation creation. */
> > if (OVS_UNLIKELY(alg_helpers[ct_alg_ctl] && conn && conn->alg)) {
> > - alg_helpers[ct_alg_ctl](ctx, pkt, conn_for_expectation, now,
> > - CT_FTP_CTL_INTEREST, nat);
> > + ovs_mutex_lock(&conn->lock.lock);
> > + alg_helpers[ct_alg_ctl](ctx, pkt, conn, now,
> CT_FTP_CTL_INTEREST,
> > + nat);
> > + ovs_mutex_unlock(&conn->lock.lock);
> > }
> > }
> >
> > @@ -767,86 +809,19 @@ un_nat_packet(struct dp_packet *pkt, const struct
> conn *conn,
> > }
> > }
> >
> > -/* Typical usage of this helper is in non per-packet code;
> > - * this is because the bucket lock needs to be held for lookup
> > - * and a hash would have already been needed. Hence, this function
> > - * is just intended for code clarity. */
> > -static struct conn *
> > -conn_lookup(const struct conn_key *key, long long now)
> > -{
> > - struct conn_lookup_ctx ctx;
> > - ctx.conn = NULL;
> > - ctx.key = *key;
> > - ctx.hash = conn_key_hash(key, hash_basis);
> > - unsigned bucket = hash_to_bucket(ctx.hash);
> > - conn_key_lookup(&buckets[bucket], &ctx, now);
> > - return ctx.conn;
> > -}
> > -
> > static void
> > conn_seq_skew_set(const struct conn_key *key, long long now, int
> seq_skew,
> > bool seq_skew_dir)
> > {
> > - unsigned bucket = hash_to_bucket(conn_key_hash(key, hash_basis));
> > - ct_lock_lock(&buckets[bucket].lock);
> > - struct conn *conn = conn_lookup(key, now);
> > + struct conn *conn;
> > + bool reply;
> > + uint32_t hash = conn_key_hash(key, hash_basis);
> > + conn_key_lookup(key, hash, now, &conn, &reply);
> > +
> > if (conn && seq_skew) {
> > conn->seq_skew = seq_skew;
> > conn->seq_skew_dir = seq_skew_dir;
> > }
> > - ct_lock_unlock(&buckets[bucket].lock);
> > -}
> > -
> > -static void
> > -nat_clean(struct conn *conn, struct conntrack_bucket *ctb)
> > - OVS_REQUIRES(ctb->lock)
> > -{
> > - ct_rwlock_wrlock(&resources_lock);
> > - nat_conn_keys_remove(&nat_conn_keys, &conn->rev_key, hash_basis);
> > - ct_rwlock_unlock(&resources_lock);
> > - ct_lock_unlock(&ctb->lock);
> > - unsigned bucket_rev_conn =
> > - hash_to_bucket(conn_key_hash(&conn->rev_key, hash_basis));
> > - ct_lock_lock(&buckets[bucket_rev_conn].lock);
> > - ct_rwlock_wrlock(&resources_lock);
> > - long long now = time_msec();
> > - struct conn *rev_conn = conn_lookup(&conn->rev_key, now);
> > - struct nat_conn_key_node *nat_conn_key_node =
> > - nat_conn_keys_lookup(&nat_conn_keys, &conn->rev_key,
> hash_basis);
> > -
> > - /* In the unlikely event, rev conn was recreated, then skip
> > - * rev_conn cleanup. */
> > - if (rev_conn && (!nat_conn_key_node ||
> > - conn_key_cmp(&nat_conn_key_node->value,
> > - &rev_conn->rev_key))) {
> > - hmap_remove(&buckets[bucket_rev_conn].connections,
> &rev_conn->node);
> > - free(rev_conn);
> > - }
> > -
> > - delete_conn(conn);
> > - ct_rwlock_unlock(&resources_lock);
> > - ct_lock_unlock(&buckets[bucket_rev_conn].lock);
> > - ct_lock_lock(&ctb->lock);
> > -}
> > -
> > -/* Must be called with 'CT_CONN_TYPE_DEFAULT' 'conn_type'. */
> > -static void
> > -conn_clean(struct conn *conn, struct conntrack_bucket *ctb)
> > - OVS_REQUIRES(ctb->lock)
> > -{
> > - ovs_assert(conn->conn_type == CT_CONN_TYPE_DEFAULT);
> > -
> > - if (conn->alg) {
> > - expectation_clean(&conn->key, hash_basis);
> > - }
> > - ovs_list_remove(&conn->exp_node);
> > - hmap_remove(&ctb->connections, &conn->node);
> > - atomic_count_dec(&n_conn);
> > - if (conn->nat_info) {
> > - nat_clean(conn, ctb);
> > - } else {
> > - delete_conn(conn);
> > - }
> > }
> >
> > static bool
> > @@ -869,17 +844,15 @@ ct_verify_helper(const char *helper, enum
> ct_alg_ctl_type ct_alg_ctl)
> > }
> > }
> >
> > -/* This function is called with the bucket lock held. */
> > static struct conn *
> > conn_not_found(struct dp_packet *pkt, struct conn_lookup_ctx *ctx,
> > bool commit, long long now,
> > const struct nat_action_info_t *nat_action_info,
> > - struct conn *conn_for_un_nat_copy,
> > - const char *helper,
> > - const struct alg_exp_node *alg_exp,
> > + const char *helper, const struct alg_exp_node *alg_exp,
> > enum ct_alg_ctl_type ct_alg_ctl)
> > {
> > struct conn *nc = NULL;
> > + struct conn *nat_conn = NULL;
> >
> > if (!valid_new(pkt, &ctx->key)) {
> > pkt->md.ct_state = CS_INVALID;
> > @@ -901,8 +874,7 @@ conn_not_found(struct dp_packet *pkt, struct
> conn_lookup_ctx *ctx,
> > return nc;
> > }
> >
> > - unsigned bucket = hash_to_bucket(ctx->hash);
> > - nc = new_conn(&buckets[bucket], pkt, &ctx->key, now);
> > + nc = new_conn(pkt, &ctx->key, now);
> > ctx->conn = nc;
> > nc->rev_key = nc->key;
> > conn_key_reverse(&nc->rev_key);
> > @@ -921,6 +893,8 @@ conn_not_found(struct dp_packet *pkt, struct
> conn_lookup_ctx *ctx,
> > if (nat_action_info) {
> > nc->nat_info = xmemdup(nat_action_info, sizeof
> *nc->nat_info);
> >
> > + nat_conn = xzalloc(sizeof *nat_conn);
> > +
> > if (alg_exp) {
> > if (alg_exp->nat_rpl_dst) {
> > nc->rev_key.dst.addr = alg_exp->alg_nat_repl_addr;
> > @@ -929,59 +903,50 @@ conn_not_found(struct dp_packet *pkt, struct
> conn_lookup_ctx *ctx,
> > nc->rev_key.src.addr = alg_exp->alg_nat_repl_addr;
> > nc->nat_info->nat_action = NAT_ACTION_DST;
> > }
> > - *conn_for_un_nat_copy = *nc;
> > - ct_rwlock_wrlock(&resources_lock);
> > - bool new_insert = nat_conn_keys_insert(&nat_conn_keys,
> > -
> conn_for_un_nat_copy,
> > - hash_basis);
> > - ct_rwlock_unlock(&resources_lock);
> > - if (!new_insert) {
> > - char *log_msg = xasprintf("Pre-existing alg "
> > - "nat_conn_key");
> > - ct_print_conn_info(conn_for_un_nat_copy, log_msg,
> VLL_INFO,
> > - true, false);
> > - free(log_msg);
> > - }
> > + *nat_conn = *nc;
> > } else {
> > - *conn_for_un_nat_copy = *nc;
> > - ct_rwlock_wrlock(&resources_lock);
> > - bool nat_res = nat_select_range_tuple(nc,
> > -
> conn_for_un_nat_copy);
> > + *nat_conn = *nc;
> > + bool nat_res = nat_select_range_tuple(nc, nat_conn);
> >
> > if (!nat_res) {
> > goto nat_res_exhaustion;
> > }
> >
> > - /* Update nc with nat adjustments made to
> > - * conn_for_un_nat_copy by nat_select_range_tuple(). */
> > - *nc = *conn_for_un_nat_copy;
> > - ct_rwlock_unlock(&resources_lock);
> > + /* Update nc with nat adjustments. */
> > + *nc = *nat_conn;
> > }
> > - conn_for_un_nat_copy->conn_type = CT_CONN_TYPE_UN_NAT;
> > - conn_for_un_nat_copy->nat_info = NULL;
> > - conn_for_un_nat_copy->alg = NULL;
> > nat_packet(pkt, nc, ctx->icmp_related);
> > - }
> > - hmap_insert(&buckets[bucket].connections, &nc->node, ctx->hash);
> > +
> > + nat_conn->key = nc->rev_key;
> > + nat_conn->rev_key = nc->key;
> > + nat_conn->conn_type = CT_CONN_TYPE_UN_NAT;
> > + nat_conn->nat_info = NULL;
> > + nat_conn->alg = NULL;
> > + nat_conn->nat_conn = NULL;
> > + uint32_t nat_hash = conn_key_hash(&nat_conn->key,
> > + hash_basis);
> > + cmap_insert(&cm_conns, &nat_conn->cm_node, nat_hash);
> > + }
> > +
> > + nc->nat_conn = nat_conn;
> > + ovs_mutex_init_adaptive(&nc->lock.lock);
> > + nc->conn_type = CT_CONN_TYPE_DEFAULT;
> > + cmap_insert(&cm_conns, &nc->cm_node, ctx->hash);
> > + nc->inserted = true;
> > atomic_count_inc(&n_conn);
> > }
> >
> > return nc;
> >
> > - /* This would be a user error or a DOS attack.
> > - * A user error is prevented by allocating enough
> > - * combinations of NAT addresses when combined with
> > - * ephemeral ports. A DOS attack should be protected
> > - * against with firewall rules or a separate firewall.
> > - * Also using zone partitioning can limit DoS impact. */
> > + /* This would be a user error or a DOS attack. A user error is
> prevented
> > + * by allocating enough combinations of NAT addresses when combined
> with
> > + * ephemeral ports. A DOS attack should be protected against with
> > + * firewall rules or a separate firewall. Also using zone
> partitioning
> > + * can limit DoS impact. */
> > nat_res_exhaustion:
> > + free(nat_conn);
> > ovs_list_remove(&nc->exp_node);
> > delete_conn(nc);
> > - /* conn_for_un_nat_copy is a local variable in process_one; this
> > - * memset() serves to document that conn_for_un_nat_copy is from
> > - * this point on unused. */
> > - memset(conn_for_un_nat_copy, 0, sizeof *conn_for_un_nat_copy);
> > - ct_rwlock_unlock(&resources_lock);
> > static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(5, 5);
> > VLOG_WARN_RL(&rl, "Unable to NAT due to tuple space exhaustion - "
> > "if DoS attack, use firewalling and/or zone
> partitioning.");
> > @@ -990,9 +955,10 @@ nat_res_exhaustion:
> >
> > static bool
> > conn_update_state(struct dp_packet *pkt, struct conn_lookup_ctx *ctx,
> > - struct conn **conn, long long now, unsigned bucket)
> > - OVS_REQUIRES(buckets[bucket].lock)
> > + struct conn *conn, long long now)
> > {
> > + ovs_assert(conn->conn_type == CT_CONN_TYPE_DEFAULT);
> > +
> > bool create_new_conn = false;
> >
> > if (ctx->icmp_related) {
> > @@ -1001,12 +967,11 @@ conn_update_state(struct dp_packet *pkt, struct
> conn_lookup_ctx *ctx,
> > pkt->md.ct_state |= CS_REPLY_DIR;
> > }
> > } else {
> > - if ((*conn)->alg_related) {
> > + if (conn->alg_related) {
> > pkt->md.ct_state |= CS_RELATED;
> > }
> >
> > - enum ct_update_res res = conn_update(*conn, &buckets[bucket],
> > - pkt, ctx->reply, now);
> > + enum ct_update_res res = conn_update(pkt, conn, ctx, now);
> >
> > switch (res) {
> > case CT_UPDATE_VALID:
> > @@ -1020,7 +985,9 @@ conn_update_state(struct dp_packet *pkt, struct
> conn_lookup_ctx *ctx,
> > pkt->md.ct_state = CS_INVALID;
> > break;
> > case CT_UPDATE_NEW:
> > - conn_clean(*conn, &buckets[bucket]);
> > + ovs_mutex_lock(&ct_lock.lock);
> > + conn_clean(conn);
> > + ovs_mutex_unlock(&ct_lock.lock);
> > create_new_conn = true;
> > break;
> > default:
> > @@ -1031,51 +998,6 @@ conn_update_state(struct dp_packet *pkt, struct
> conn_lookup_ctx *ctx,
> > }
> >
> > static void
> > -create_un_nat_conn(struct conn *conn_for_un_nat_copy, long long now,
> > - bool alg_un_nat)
> > -{
> > - struct conn *nc = xmemdup(conn_for_un_nat_copy, sizeof *nc);
> > - nc->key = conn_for_un_nat_copy->rev_key;
> > - nc->rev_key = conn_for_un_nat_copy->key;
> > - uint32_t un_nat_hash = conn_key_hash(&nc->key, hash_basis);
> > - unsigned un_nat_conn_bucket = hash_to_bucket(un_nat_hash);
> > - ct_lock_lock(&buckets[un_nat_conn_bucket].lock);
> > - struct conn *rev_conn = conn_lookup(&nc->key, now);
> > -
> > - if (alg_un_nat) {
> > - if (!rev_conn) {
> > - hmap_insert(&buckets[un_nat_conn_bucket].connections,
> > - &nc->node, un_nat_hash);
> > - } else {
> > - char *log_msg = xasprintf("Unusual condition for un_nat
> conn "
> > - "create for alg: rev_conn %p",
> rev_conn);
> > - ct_print_conn_info(nc, log_msg, VLL_INFO, true, false);
> > - free(log_msg);
> > - free(nc);
> > - }
> > - } else {
> > - ct_rwlock_rdlock(&resources_lock);
> > -
> > - struct nat_conn_key_node *nat_conn_key_node =
> > - nat_conn_keys_lookup(&nat_conn_keys, &nc->key, hash_basis);
> > - if (nat_conn_key_node &&
> !conn_key_cmp(&nat_conn_key_node->value,
> > - &nc->rev_key) && !rev_conn) {
> > - hmap_insert(&buckets[un_nat_conn_bucket].connections,
> &nc->node,
> > - un_nat_hash);
> > - } else {
> > - char *log_msg = xasprintf("Unusual condition for un_nat
> conn "
> > - "create:
> nat_conn_key_node/rev_conn "
> > - "%p/%p", nat_conn_key_node,
> rev_conn);
> > - ct_print_conn_info(nc, log_msg, VLL_INFO, true, false);
> > - free(log_msg);
> > - free(nc);
> > - }
> > - ct_rwlock_unlock(&resources_lock);
> > - }
> > - ct_lock_unlock(&buckets[un_nat_conn_bucket].lock);
> > -}
> > -
> > -static void
> > handle_nat(struct dp_packet *pkt, struct conn *conn,
> > uint16_t zone, bool reply, bool related)
> > {
> > @@ -1097,9 +1019,8 @@ handle_nat(struct dp_packet *pkt, struct conn
> *conn,
> >
> > static bool
> > check_orig_tuple(struct dp_packet *pkt, struct conn_lookup_ctx *ctx_in,
> > - long long now, unsigned *bucket, struct conn **conn,
> > + long long now, struct conn **conn,
> > const struct nat_action_info_t *nat_action_info)
> > - OVS_REQUIRES(buckets[(*bucket)].lock)
> > {
> > if ((ctx_in->key.dl_type == htons(ETH_TYPE_IP) &&
> > !pkt->md.ct_orig_tuple.ipv4.ipv4_proto) ||
> > @@ -1110,57 +1031,48 @@ check_orig_tuple(struct dp_packet *pkt, struct
> conn_lookup_ctx *ctx_in,
> > return false;
> > }
> >
> > - ct_lock_unlock(&buckets[(*bucket)].lock);
> > - struct conn_lookup_ctx ctx;
> > - memset(&ctx, 0 , sizeof ctx);
> > - ctx.conn = NULL;
> > + struct conn_key key;
> > + memset(&key, 0 , sizeof key);
> >
> > if (ctx_in->key.dl_type == htons(ETH_TYPE_IP)) {
> > - ctx.key.src.addr.ipv4_aligned =
> pkt->md.ct_orig_tuple.ipv4.ipv4_src;
> > - ctx.key.dst.addr.ipv4_aligned =
> pkt->md.ct_orig_tuple.ipv4.ipv4_dst;
> > + key.src.addr.ipv4_aligned = pkt->md.ct_orig_tuple.ipv4.ipv4_src;
> > + key.dst.addr.ipv4_aligned = pkt->md.ct_orig_tuple.ipv4.ipv4_dst;
> >
> > if (ctx_in->key.nw_proto == IPPROTO_ICMP) {
> > - ctx.key.src.icmp_id = ctx_in->key.src.icmp_id;
> > - ctx.key.dst.icmp_id = ctx_in->key.dst.icmp_id;
> > + key.src.icmp_id = ctx_in->key.src.icmp_id;
> > + key.dst.icmp_id = ctx_in->key.dst.icmp_id;
> > uint16_t src_port =
> ntohs(pkt->md.ct_orig_tuple.ipv4.src_port);
> > - ctx.key.src.icmp_type = (uint8_t) src_port;
> > - ctx.key.dst.icmp_type =
> reverse_icmp_type(ctx.key.src.icmp_type);
> > + key.src.icmp_type = (uint8_t) src_port;
> > + key.dst.icmp_type = reverse_icmp_type(key.src.icmp_type);
> > } else {
> > - ctx.key.src.port = pkt->md.ct_orig_tuple.ipv4.src_port;
> > - ctx.key.dst.port = pkt->md.ct_orig_tuple.ipv4.dst_port;
> > + key.src.port = pkt->md.ct_orig_tuple.ipv4.src_port;
> > + key.dst.port = pkt->md.ct_orig_tuple.ipv4.dst_port;
> > }
> > - ctx.key.nw_proto = pkt->md.ct_orig_tuple.ipv4.ipv4_proto;
> > + key.nw_proto = pkt->md.ct_orig_tuple.ipv4.ipv4_proto;
> > } else {
> > - ctx.key.src.addr.ipv6_aligned =
> pkt->md.ct_orig_tuple.ipv6.ipv6_src;
> > - ctx.key.dst.addr.ipv6_aligned =
> pkt->md.ct_orig_tuple.ipv6.ipv6_dst;
> > + key.src.addr.ipv6_aligned = pkt->md.ct_orig_tuple.ipv6.ipv6_src;
> > + key.dst.addr.ipv6_aligned = pkt->md.ct_orig_tuple.ipv6.ipv6_dst;
> >
> > if (ctx_in->key.nw_proto == IPPROTO_ICMPV6) {
> > - ctx.key.src.icmp_id = ctx_in->key.src.icmp_id;
> > - ctx.key.dst.icmp_id = ctx_in->key.dst.icmp_id;
> > + key.src.icmp_id = ctx_in->key.src.icmp_id;
> > + key.dst.icmp_id = ctx_in->key.dst.icmp_id;
> > uint16_t src_port =
> ntohs(pkt->md.ct_orig_tuple.ipv6.src_port);
> > - ctx.key.src.icmp_type = (uint8_t) src_port;
> > - ctx.key.dst.icmp_type =
> reverse_icmp6_type(ctx.key.src.icmp_type);
> > + key.src.icmp_type = (uint8_t) src_port;
> > + key.dst.icmp_type = reverse_icmp6_type(key.src.icmp_type);
> > } else {
> > - ctx.key.src.port = pkt->md.ct_orig_tuple.ipv6.src_port;
> > - ctx.key.dst.port = pkt->md.ct_orig_tuple.ipv6.dst_port;
> > + key.src.port = pkt->md.ct_orig_tuple.ipv6.src_port;
> > + key.dst.port = pkt->md.ct_orig_tuple.ipv6.dst_port;
> > }
> > - ctx.key.nw_proto = pkt->md.ct_orig_tuple.ipv6.ipv6_proto;
> > + key.nw_proto = pkt->md.ct_orig_tuple.ipv6.ipv6_proto;
> > }
> >
> > - ctx.key.dl_type = ctx_in->key.dl_type;
> > - ctx.key.zone = pkt->md.ct_zone;
> > - ctx.hash = conn_key_hash(&ctx.key, hash_basis);
> > - *bucket = hash_to_bucket(ctx.hash);
> > - ct_lock_lock(&buckets[(*bucket)].lock);
> > - conn_key_lookup(&buckets[(*bucket)], &ctx, now);
> > - *conn = ctx.conn;
> > - return *conn ? true : false;
> > -}
> > + key.dl_type = ctx_in->key.dl_type;
> > + key.zone = pkt->md.ct_zone;
> > + uint32_t hash = conn_key_hash(&key, hash_basis);
> > + bool reply;
> > + conn_key_lookup(&key, hash, now, conn, &reply);
> >
> > -static bool
> > -is_un_nat_conn_valid(const struct conn *un_nat_conn)
> > -{
> > - return un_nat_conn->conn_type == CT_CONN_TYPE_UN_NAT;
> > + return *conn ? true : false;
> > }
> >
> > static bool
> > @@ -1168,25 +1080,28 @@ conn_update_state_alg(struct dp_packet *pkt,
> struct conn_lookup_ctx *ctx,
> > struct conn *conn,
> > const struct nat_action_info_t *nat_action_info,
> > enum ct_alg_ctl_type ct_alg_ctl, long long now,
> > - unsigned bucket, bool *create_new_conn)
> > - OVS_REQUIRES(buckets[bucket].lock)
> > + bool *create_new_conn)
> > {
> > if (is_ftp_ctl(ct_alg_ctl)) {
> > /* Keep sequence tracking in sync with the source of the
> > * sequence skew. */
> > + ovs_mutex_lock(&conn->lock.lock);
> > if (ctx->reply != conn->seq_skew_dir) {
> > handle_ftp_ctl(ctx, pkt, conn, now, CT_FTP_CTL_OTHER,
> > !!nat_action_info);
> > - *create_new_conn = conn_update_state(pkt, ctx, &conn, now,
> > - bucket);
> > + /* conn_update_state locks for unrelated fields, so unlock.
> */
> > + ovs_mutex_unlock(&conn->lock.lock);
> > + *create_new_conn = conn_update_state(pkt, ctx, conn, now);
> > } else {
> > - *create_new_conn = conn_update_state(pkt, ctx, &conn, now,
> > - bucket);
> > -
> > + /* conn_update_state locks for unrelated fields, so unlock.
> */
> > + ovs_mutex_unlock(&conn->lock.lock);
> > + *create_new_conn = conn_update_state(pkt, ctx, conn, now);
> > + ovs_mutex_lock(&conn->lock.lock);
> > if (*create_new_conn == false) {
> > handle_ftp_ctl(ctx, pkt, conn, now, CT_FTP_CTL_OTHER,
> > !!nat_action_info);
> > }
> > + ovs_mutex_unlock(&conn->lock.lock);
> > }
> > return true;
> > }
> > @@ -1195,74 +1110,57 @@ conn_update_state_alg(struct dp_packet *pkt,
> struct conn_lookup_ctx *ctx,
> >
> > static void
> > process_one(struct dp_packet *pkt, struct conn_lookup_ctx *ctx,
> uint16_t zone,
> > - bool force, bool commit, long long now, const uint32_t
> *setmark,
> > + bool force, bool commit, long long now,
> > + const uint32_t *setmark,
> > const struct ovs_key_ct_labels *setlabel,
> > const struct nat_action_info_t *nat_action_info,
> > ovs_be16 tp_src, ovs_be16 tp_dst, const char *helper)
> > {
> > - struct conn *conn;
> > - unsigned bucket = hash_to_bucket(ctx->hash);
> > - ct_lock_lock(&buckets[bucket].lock);
> > - conn_key_lookup(&buckets[bucket], ctx, now);
> > - conn = ctx->conn;
> > + bool create_new_conn = false;
> > + conn_key_lookup(&ctx->key, ctx->hash, now, &ctx->conn, &ctx->reply);
> > + struct conn *conn = ctx->conn;
> >
> > /* Delete found entry if in wrong direction. 'force' implies
> commit. */
> > if (conn && force && ctx->reply) {
> > - conn_clean(conn, &buckets[bucket]);
> > + ovs_mutex_lock(&ct_lock.lock);
> > + conn_clean(conn);
> > + ovs_mutex_unlock(&ct_lock.lock);
> > conn = NULL;
> > }
> >
> > if (OVS_LIKELY(conn)) {
> > if (conn->conn_type == CT_CONN_TYPE_UN_NAT) {
> > -
> > ctx->reply = true;
> > + struct conn *rev_conn = conn; /* Save for debugging. */
> > + uint32_t hash = conn_key_hash(&conn->rev_key, hash_basis);
> > + conn_key_lookup(&ctx->key, hash, now, &conn, &ctx->reply);
> >
> > - struct conn_lookup_ctx ctx2;
> > - ctx2.conn = NULL;
> > - ctx2.key = conn->rev_key;
> > - ctx2.hash = conn_key_hash(&conn->rev_key, hash_basis);
> > -
> > - ct_lock_unlock(&buckets[bucket].lock);
> > - bucket = hash_to_bucket(ctx2.hash);
> > -
> > - ct_lock_lock(&buckets[bucket].lock);
> > - conn_key_lookup(&buckets[bucket], &ctx2, now);
> > -
> > - if (ctx2.conn) {
> > - conn = ctx2.conn;
> > - } else {
> > - /* It is a race condition where conn has timed out and
> removed
> > - * between unlock of the rev_conn and lock of the
> forward conn;
> > - * nothing to do. */
> > + if (!conn) {
> > pkt->md.ct_state |= CS_TRACKED | CS_INVALID;
> > - ct_lock_unlock(&buckets[bucket].lock);
> > + char *log_msg = xasprintf("Missing master conn %p",
> rev_conn);
> > + ct_print_conn_info(conn, log_msg, VLL_INFO, true, true);
> > + free(log_msg);
> > return;
> > }
> > }
> > }
> >
> > - bool create_new_conn = false;
> > - struct conn conn_for_un_nat_copy;
> > - conn_for_un_nat_copy.conn_type = CT_CONN_TYPE_DEFAULT;
> > -
> > enum ct_alg_ctl_type ct_alg_ctl = get_alg_ctl_type(pkt, tp_src,
> tp_dst,
> > helper);
> >
> > if (OVS_LIKELY(conn)) {
> > if (OVS_LIKELY(!conn_update_state_alg(pkt, ctx, conn,
> > nat_action_info,
> > - ct_alg_ctl, now, bucket,
> > + ct_alg_ctl, now,
> > &create_new_conn))) {
> > - create_new_conn = conn_update_state(pkt, ctx, &conn, now,
> > - bucket);
> > +
> > + create_new_conn = conn_update_state(pkt, ctx, conn, now);
> > }
> > if (nat_action_info && !create_new_conn) {
> > handle_nat(pkt, conn, zone, ctx->reply, ctx->icmp_related);
> > }
> > -
> > - } else if (check_orig_tuple(pkt, ctx, now, &bucket, &conn,
> > - nat_action_info)) {
> > - create_new_conn = conn_update_state(pkt, ctx, &conn, now,
> bucket);
> > + } else if (check_orig_tuple(pkt, ctx, now, &conn, nat_action_info))
> {
> > + create_new_conn = conn_update_state(pkt, ctx, conn, now);
> > } else {
> > if (ctx->icmp_related) {
> > /* An icmp related conn should always be found; no new
> > @@ -1277,19 +1175,20 @@ process_one(struct dp_packet *pkt, struct
> conn_lookup_ctx *ctx, uint16_t zone,
> > struct alg_exp_node alg_exp_entry;
> >
> > if (OVS_UNLIKELY(create_new_conn)) {
> > -
> > - ct_rwlock_rdlock(&resources_lock);
> > - alg_exp = expectation_lookup(&alg_expectations, &ctx->key,
> hash_basis,
> > + ovs_rwlock_rdlock(&resources_lock.lock);
> > + alg_exp = expectation_lookup(&alg_expectations, &ctx->key,
> > + hash_basis,
> > alg_src_ip_wc(ct_alg_ctl));
> > if (alg_exp) {
> > alg_exp_entry = *alg_exp;
> > alg_exp = &alg_exp_entry;
> > }
> > - ct_rwlock_unlock(&resources_lock);
> > + ovs_rwlock_unlock(&resources_lock.lock);
> >
> > + ovs_mutex_lock(&ct_lock.lock);
> > conn = conn_not_found(pkt, ctx, commit, now, nat_action_info,
> > - &conn_for_un_nat_copy, helper, alg_exp,
> > - ct_alg_ctl);
> > + helper, alg_exp, ct_alg_ctl);
> > + ovs_mutex_unlock(&ct_lock.lock);
> > }
> >
> > write_ct_md(pkt, zone, conn, &ctx->key, alg_exp);
> > @@ -1302,23 +1201,11 @@ process_one(struct dp_packet *pkt, struct
> conn_lookup_ctx *ctx, uint16_t zone,
> > set_label(pkt, conn, &setlabel[0], &setlabel[1]);
> > }
> >
> > - struct conn conn_for_expectation;
> > - if (OVS_UNLIKELY((ct_alg_ctl != CT_ALG_CTL_NONE) && conn)) {
> > - conn_for_expectation = *conn;
> > - }
> > -
> > - ct_lock_unlock(&buckets[bucket].lock);
> > -
> > - if (is_un_nat_conn_valid(&conn_for_un_nat_copy)) {
> > - create_un_nat_conn(&conn_for_un_nat_copy, now, !!alg_exp);
> > - }
> > -
> > - handle_alg_ctl(ctx, pkt, ct_alg_ctl, conn, now, !!nat_action_info,
> > - &conn_for_expectation);
> > + handle_alg_ctl(ctx, pkt, ct_alg_ctl, conn, now, !!nat_action_info);
> > }
> >
> > /* Sends the packets in '*pkt_batch' through the connection tracker
> 'ct'. All
> > - * the packets should have the same 'dl_type' (IPv4 or IPv6) and should
> have
> > + * the packets must have the same 'dl_type' (IPv4 or IPv6) and should
> have
> > * the l3 and and l4 offset properly set.
> > *
> > * If 'commit' is true, the packets are allowed to create new entries
> in the
> > @@ -1334,12 +1221,12 @@ conntrack_execute(struct dp_packet_batch
> *pkt_batch, ovs_be16 dl_type,
> > const struct nat_action_info_t *nat_action_info,
> > long long now)
> > {
> > -
> > struct dp_packet *packet;
> > struct conn_lookup_ctx ctx;
> >
> > DP_PACKET_BATCH_FOR_EACH (i, packet, pkt_batch) {
> > - if (!conn_key_extract(packet, dl_type, &ctx, zone)) {
> > + if (packet->md.ct_state == CS_INVALID
> > + || !conn_key_extract(packet, dl_type, &ctx, zone)) {
> > packet->md.ct_state = CS_INVALID;
> > write_ct_md(packet, zone, NULL, NULL, NULL);
> > continue;
> > @@ -1392,35 +1279,57 @@ set_label(struct dp_packet *pkt, struct conn
> *conn,
> > }
> >
> >
> > -/* Delete the expired connections from 'ctb', up to 'limit'. Returns the
> > - * earliest expiration time among the remaining connections in 'ctb'.
> Returns
> > - * LLONG_MAX if 'ctb' is empty. The return value might be smaller than
> 'now',
> > - * if 'limit' is reached */
> > +/* Delete the expired connections, up to 'limit'. Returns the earliest
> > + * expiration time among the remaining connections in all expiration
> lists.
> > + * Returns LLONG_MAX if all expiration lists are empty. The return
> value
> > + * might be smaller than 'now',if 'limit' is reached */
> > static long long
> > -sweep_bucket(struct conntrack_bucket *ctb, long long now, size_t limit)
> > - OVS_REQUIRES(ctb->lock)
> > +ct_sweep(long long now, size_t limit)
> > {
> > struct conn *conn, *next;
> > long long min_expiration = LLONG_MAX;
> > size_t count = 0;
> >
> > + ovs_mutex_lock(&ct_lock.lock);
> > +
> > for (unsigned i = 0; i < N_CT_TM; i++) {
> > - LIST_FOR_EACH_SAFE (conn, next, exp_node, &ctb->exp_lists[i]) {
> > + LIST_FOR_EACH_SAFE (conn, next, exp_node, &cm_exp_lists[i]) {
> > if (conn->conn_type == CT_CONN_TYPE_DEFAULT) {
> > - if (!conn_expired(conn, now) || count >= limit) {
> > + ovs_mutex_lock(&conn->lock.lock);
> > + if (conn->exp_list_id != NO_UPD_EXP_LIST) {
> > + ovs_list_remove(&conn->exp_node);
> > + ovs_list_push_back(&cm_exp_lists[conn->exp_list_id],
> > + &conn->exp_node);
> > + conn->exp_list_id = NO_UPD_EXP_LIST;
> > + ovs_mutex_unlock(&conn->lock.lock);
> > + } else if (!conn_expired(conn, now) || count >= limit) {
> > + /* Not looking at conn changable fields. */
> > + ovs_mutex_unlock(&conn->lock.lock);
> > min_expiration = MIN(min_expiration,
> conn->expiration);
> > if (count >= limit) {
> > /* Do not check other lists. */
> > COVERAGE_INC(conntrack_long_cleanup);
> > - return min_expiration;
> > + goto out;
> > }
> > break;
> > + } else {
> > + /* Not looking at conn changable fields. */
> > + ovs_mutex_unlock(&conn->lock.lock);
> > + if (conn->inserted) {
> > + conn_clean(conn);
> > + } else {
> > + break;
> > + }
> > }
> > - conn_clean(conn, ctb);
> > count++;
> > }
> > }
> > }
> > +
> > +out:
> > + VLOG_DBG("conntrack cleanup %"PRIuSIZE" entries in %lld msec",
> count,
> > + time_msec() - now);
> > + ovs_mutex_unlock(&ct_lock.lock);
> > return min_expiration;
> > }
> >
> > @@ -1431,50 +1340,11 @@ sweep_bucket(struct conntrack_bucket *ctb, long
> long now, size_t limit)
> > static long long
> > conntrack_clean(long long now)
> > {
> > - long long next_wakeup = now + CT_TM_MIN;
> > unsigned int n_conn_limit_;
> > - size_t clean_count = 0;
> > -
> > atomic_read_relaxed(&n_conn_limit, &n_conn_limit_);
> >
> > - for (unsigned i = 0; i < CONNTRACK_BUCKETS; i++) {
> > - struct conntrack_bucket *ctb = &buckets[i];
> > - size_t prev_count;
> > - long long min_exp;
> > -
> > - ovs_mutex_lock(&ctb->cleanup_mutex);
> > - if (ctb->next_cleanup > now) {
> > - goto next_bucket;
> > - }
> > -
> > - ct_lock_lock(&ctb->lock);
> > - prev_count = hmap_count(&ctb->connections);
> > - /* If the connections are well distributed among buckets, we
> want to
> > - * limit to 10% of the global limit equally split among
> buckets. If
> > - * the bucket is busier than the others, we limit to 10% of its
> > - * current size. */
> > - min_exp = sweep_bucket(ctb, now,
> > - MAX(prev_count / 10, n_conn_limit_ / (CONNTRACK_BUCKETS *
> 10)));
> > - clean_count += prev_count - hmap_count(&ctb->connections);
> > -
> > - if (min_exp > now) {
> > - /* We call hmap_shrink() only if sweep_bucket() managed to
> delete
> > - * every expired connection. */
> > - hmap_shrink(&ctb->connections);
> > - }
> > -
> > - ct_lock_unlock(&ctb->lock);
> > -
> > - ctb->next_cleanup = MIN(min_exp, now + CT_TM_MIN);
> > -
> > -next_bucket:
> > - next_wakeup = MIN(next_wakeup, ctb->next_cleanup);
> > - ovs_mutex_unlock(&ctb->cleanup_mutex);
> > - }
> > -
> > - VLOG_DBG("conntrack cleanup %"PRIuSIZE" entries in %lld msec",
> > - clean_count, time_msec() - now);
> > -
> > + long long min_exp = ct_sweep(now, n_conn_limit_ / 50);
> > + long long next_wakeup = MIN(min_exp, now + CT_TM_MIN);
> > return next_wakeup;
> > }
> >
> > @@ -1492,16 +1362,16 @@ next_bucket:
> > * are coping with the current cleanup tasks, then we wait at least
> > * 5 seconds to do further cleanup.
> > *
> > - * - We don't want to keep the buckets locked too long, as we might
> prevent
> > + * - We don't want to keep the map locked too long, as we might prevent
> > * traffic from flowing. CT_CLEAN_MIN_INTERVAL ensures that if
> cleanup is
> > - * behind, there is at least some 200ms blocks of time when buckets
> will be
> > + * behind, there is at least some 200ms blocks of time when the map
> will be
> > * left alone, so the datapath can operate unhindered.
> > */
> > #define CT_CLEAN_INTERVAL 5000 /* 5 seconds */
> > #define CT_CLEAN_MIN_INTERVAL 200 /* 0.2 seconds */
> >
> > static void *
> > -clean_thread_main(void *f_ OVS_UNUSED)
> > +clean_thread_main(void *f OVS_UNUSED)
> > {
> > while (!latch_is_set(&clean_thread_exit)) {
> > long long next_wake;
> > @@ -2192,7 +2062,9 @@ nat_select_range_tuple(const struct conn *conn,
> struct conn *nat_conn)
> >
> > uint16_t port = first_port;
> > bool all_ports_tried = false;
> > - bool original_ports_tried = false;
> > + /* For DNAT, we don't use ephemeral ports. */
> > + bool ephemeral_ports_tried = conn->nat_info->nat_action &
> NAT_ACTION_DST
> > + ? true : false;
> > struct ct_addr first_addr = ct_addr;
> >
> > while (true) {
> > @@ -2211,8 +2083,10 @@ nat_select_range_tuple(const struct conn *conn,
> struct conn *nat_conn)
> > nat_conn->rev_key.src.port = htons(port);
> > }
> >
> > - bool new_insert = nat_conn_keys_insert(&nat_conn_keys, nat_conn,
> > - hash_basis);
> > + uint32_t conn_hash = conn_key_hash(&nat_conn->rev_key,
> hash_basis);
> > + bool new_insert = conn_available(&nat_conn->rev_key, conn_hash,
> > + time_msec());
> > +
> > if (new_insert) {
> > return true;
> > } else if (!all_ports_tried) {
> > @@ -2238,13 +2112,14 @@ nat_select_range_tuple(const struct conn *conn,
> struct conn *nat_conn)
> > ct_addr = conn->nat_info->min_addr;
> > }
> > if (!memcmp(&ct_addr, &first_addr, sizeof ct_addr)) {
> > - if (!original_ports_tried) {
> > - original_ports_tried = true;
> > + if (ephemeral_ports_tried) {
> > + break;
> > + } else {
> > + ephemeral_ports_tried = true;
> > ct_addr = conn->nat_info->min_addr;
> > + first_addr = ct_addr;
> > min_port = MIN_NAT_EPHEMERAL_PORT;
> > max_port = MAX_NAT_EPHEMERAL_PORT;
> > - } else {
> > - break;
> > }
> > }
> > first_port = min_port;
> > @@ -2255,95 +2130,6 @@ nat_select_range_tuple(const struct conn *conn,
> struct conn *nat_conn)
> > return false;
> > }
> >
> > -/* This function must be called with the resources lock taken. */
> > -static struct nat_conn_key_node *
> > -nat_conn_keys_lookup(struct hmap *nat_conn_keys_,
> > - const struct conn_key *key,
> > - uint32_t basis)
> > -{
> > - struct nat_conn_key_node *nat_conn_key_node;
> > -
> > - HMAP_FOR_EACH_WITH_HASH (nat_conn_key_node, node,
> > - conn_key_hash(key, basis), nat_conn_keys_)
> {
> > - if (!conn_key_cmp(&nat_conn_key_node->key, key)) {
> > - return nat_conn_key_node;
> > - }
> > - }
> > - return NULL;
> > -}
> > -
> > -/* This function must be called with the resources lock taken. */
> > -static bool
> > -nat_conn_keys_insert(struct hmap *nat_conn_keys_, const struct conn
> *nat_conn,
> > - uint32_t basis)
> > -{
> > - struct nat_conn_key_node *nat_conn_key_node =
> > - nat_conn_keys_lookup(nat_conn_keys_, &nat_conn->rev_key, basis);
> > -
> > - if (!nat_conn_key_node) {
> > - struct nat_conn_key_node *nat_conn_key =
> > - xzalloc(sizeof *nat_conn_key);
> > - nat_conn_key->key = nat_conn->rev_key;
> > - nat_conn_key->value = nat_conn->key;
> > - hmap_insert(nat_conn_keys_, &nat_conn_key->node,
> > - conn_key_hash(&nat_conn_key->key, basis));
> > - return true;
> > - }
> > - return false;
> > -}
> > -
> > -/* This function must be called with the resources write lock taken. */
> > -static void
> > -nat_conn_keys_remove(struct hmap *nat_conn_keys_,
> > - const struct conn_key *key,
> > - uint32_t basis)
> > -{
> > - struct nat_conn_key_node *nat_conn_key_node;
> > -
> > - HMAP_FOR_EACH_WITH_HASH (nat_conn_key_node, node,
> > - conn_key_hash(key, basis), nat_conn_keys_)
> {
> > - if (!conn_key_cmp(&nat_conn_key_node->key, key)) {
> > - hmap_remove(nat_conn_keys_, &nat_conn_key_node->node);
> > - free(nat_conn_key_node);
> > - return;
> > - }
> > - }
> > -}
> > -
> > -static void
> > -conn_key_lookup(struct conntrack_bucket *ctb, struct conn_lookup_ctx
> *ctx,
> > - long long now)
> > - OVS_REQUIRES(ctb->lock)
> > -{
> > - uint32_t hash = ctx->hash;
> > - struct conn *conn;
> > -
> > - ctx->conn = NULL;
> > -
> > - HMAP_FOR_EACH_WITH_HASH (conn, node, hash, &ctb->connections) {
> > - if (!conn_key_cmp(&conn->key, &ctx->key)
> > - && !conn_expired(conn, now)) {
> > - ctx->conn = conn;
> > - ctx->reply = false;
> > - break;
> > - }
> > - if (!conn_key_cmp(&conn->rev_key, &ctx->key)
> > - && !conn_expired(conn, now)) {
> > - ctx->conn = conn;
> > - ctx->reply = true;
> > - break;
> > - }
> > - }
> > -}
> > -
> > -static enum ct_update_res
> > -conn_update(struct conn *conn, struct conntrack_bucket *ctb,
> > - struct dp_packet *pkt, bool reply, long long now)
> > -{
> > - return l4_protos[conn->key.nw_proto]->conn_update(conn, ctb, pkt,
> > - reply, now);
> > -}
> > -
> > static bool
> > conn_expired(struct conn *conn, long long now)
> > {
> > @@ -2360,10 +2146,9 @@ valid_new(struct dp_packet *pkt, struct conn_key
> *key)
> > }
> >
> > static struct conn *
> > -new_conn(struct conntrack_bucket *ctb, struct dp_packet *pkt,
> > - struct conn_key *key, long long now)
> > +new_conn(struct dp_packet *pkt, struct conn_key *key, long long now)
> > {
> > - struct conn *newconn = l4_protos[key->nw_proto]->new_conn(ctb, pkt,
> now);
> > + struct conn *newconn = l4_protos[key->nw_proto]->new_conn(pkt, now);
> > if (newconn) {
> > newconn->key = *key;
> > }
> > @@ -2371,11 +2156,38 @@ new_conn(struct conntrack_bucket *ctb, struct
> dp_packet *pkt,
> > return newconn;
> > }
> >
> > +static enum ct_update_res
> > +conn_update(struct dp_packet *pkt, struct conn *conn,
> > + struct conn_lookup_ctx *ctx, long long now)
> > +{
> > + enum ct_update_res update_res =
> > + l4_protos[conn->key.nw_proto]->conn_update(conn, pkt,
> ctx->reply,
> > + now);
> > + return update_res;
> > +}
> > +
> > static void
> > delete_conn(struct conn *conn)
> > {
> > - free(conn->nat_info);
> > - free(conn->alg);
> > + if (conn->conn_type == CT_CONN_TYPE_DEFAULT) {
> > + free(conn->nat_info);
> > + free(conn->alg);
> > + ovs_mutex_destroy(&conn->lock.lock);
> > + if (conn->nat_conn) {
>
> IIRC, free() can take a null pointer, so this check is redundant.
>
Maybe your memory was jogged by the code a couple lines above
> + free(conn->nat_info);
> + free(conn->alg);
or all the free() calls in the file, which don't have NULL checks :-)
BTW, I had removed this spurious NULL check in Patch 4, but now removed in
this Patch 2.
Thanks !
>
> > + free(conn->nat_conn);
> > + }
> > + free(conn);
>
> No matter what, I think this must be wrong.
>
Can you elaborate on why it is wrong ?
Sorry, I am too lazy to make more coffee :-)
> > + }
> > +}
> > +
> > +static void
> > +delete_conn_one(struct conn *conn)
> > +{
> > + if (conn->conn_type == CT_CONN_TYPE_DEFAULT) {
> > + free(conn->nat_info);
> > + free(conn->alg);
> > + ovs_mutex_destroy(&conn->lock.lock);
> > + }
> > free(conn);
> > }
>
> I think this can be expressed as a single routine:
>
See the comments I added above about conn_clean()/conn_clean_one()
which also refers to delete_conn()/delete_conn_one()
>
> static void delete_conn_one(struct conn *conn)
> {
> if (conn->conn_type == CT_CONN_TYPE_DEFAULT) {
> free(conn->nat_info);
> free(conn->alg);
> free(conn->nat_conn);
> }
> free(conn);
> }
>
> Then you can drop delete_conn().
>
I moved delete_conn_cmn() from Patch 4 to this Patch 2.
I ended up with something like this.
static void
delete_conn_cmn(struct conn *conn)
{
free(conn->nat_info);
free(conn->alg);
ovs_mutex_destroy(&conn->lock.lock);
}
static void
delete_conn(struct conn *conn)
{
ovs_assert(conn->conn_type == CT_CONN_TYPE_DEFAULT);
delete_conn_cmn(conn);
if (conn->nat_conn) {
free(conn->nat_conn);
}
free(conn);
}
/* Only used by conn_clean_one(). */
static void
delete_conn_one(struct conn *conn)
{
if (conn->conn_type == CT_CONN_TYPE_DEFAULT) {
delete_conn_cmn(conn);
}
free(conn);
}
> >
> > @@ -2507,7 +2319,7 @@ conntrack_dump_start(struct conntrack_dump *dump,
> const uint16_t *pzone,
> > dump->filter_zone = true;
> > }
> >
> > - *ptot_bkts = CONNTRACK_BUCKETS;
> > + *ptot_bkts = 1; /* Need to clean up the callers. */
> > return 0;
> > }
> >
> > @@ -2516,36 +2328,21 @@ conntrack_dump_next(struct conntrack_dump *dump,
> struct ct_dpif_entry *entry)
> > {
> > long long now = time_msec();
> >
> > - while (dump->bucket < CONNTRACK_BUCKETS) {
> > - struct hmap_node *node;
> > -
> > - ct_lock_lock(&buckets[dump->bucket].lock);
> > - for (;;) {
> > - struct conn *conn;
> > -
> > - node = hmap_at_position(&buckets[dump->bucket].connections,
> > - &dump->bucket_pos);
> > - if (!node) {
> > - break;
> > - }
> > - INIT_CONTAINER(conn, node, node);
> > - if ((!dump->filter_zone || conn->key.zone == dump->zone) &&
> > - (conn->conn_type != CT_CONN_TYPE_UN_NAT)) {
> > - conn_to_ct_dpif_entry(conn, entry, now, dump->bucket);
> > - break;
> > - }
> > - /* Else continue, until we find an entry in the appropriate
> zone
> > - * or the bucket has been scanned completely. */
> > + for (;;) {
> > + struct cmap_node *cm_node = cmap_next_position(&cm_conns,
> > + &dump->cm_pos);
> > + if (!cm_node) {
> > + break;
> > }
> > - ct_lock_unlock(&buckets[dump->bucket].lock);
> > -
> > - if (!node) {
> > - memset(&dump->bucket_pos, 0, sizeof dump->bucket_pos);
> > - dump->bucket++;
> > - } else {
> > + struct conn *conn;
> > + INIT_CONTAINER(conn, cm_node, cm_node);
> > + if ((!dump->filter_zone || conn->key.zone == dump->zone) &&
> > + (conn->conn_type != CT_CONN_TYPE_UN_NAT)) {
> > + conn_to_ct_dpif_entry(conn, entry, now, 0);
> > return 0;
> > }
> > }
> > +
> > return EOF;
> > }
> >
> > @@ -2558,42 +2355,41 @@ conntrack_dump_done(struct conntrack_dump *dump
> OVS_UNUSED)
> > int
> > conntrack_flush(const uint16_t *zone)
> > {
> > - for (unsigned i = 0; i < CONNTRACK_BUCKETS; i++) {
> > - struct conn *conn, *next;
> > + struct conn *conn;
> >
> > - ct_lock_lock(&buckets[i].lock);
> > - HMAP_FOR_EACH_SAFE (conn, next, node, &buckets[i].connections) {
> > - if ((!zone || *zone == conn->key.zone) &&
> > - (conn->conn_type == CT_CONN_TYPE_DEFAULT)) {
> > - conn_clean(conn, &buckets[i]);
> > - }
> > + ovs_mutex_lock(&ct_lock.lock);
> > +
> > + CMAP_FOR_EACH (conn, cm_node, &cm_conns) {
> > + if (!zone || *zone == conn->key.zone) {
> > + conn_clean_one(conn);
> > }
> > - ct_lock_unlock(&buckets[i].lock);
> > }
> >
> > + ovs_mutex_unlock(&ct_lock.lock);
> > +
> > return 0;
> > }
> >
> > int
> > conntrack_flush_tuple(const struct ct_dpif_tuple *tuple, uint16_t zone)
> > {
> > - struct conn_lookup_ctx ctx;
> > int error = 0;
> > + struct conn_lookup_ctx ctx;
> >
> > memset(&ctx, 0, sizeof(ctx));
> > tuple_to_conn_key(tuple, zone, &ctx.key);
> > ctx.hash = conn_key_hash(&ctx.key, hash_basis);
> > - unsigned bucket = hash_to_bucket(ctx.hash);
> >
> > - ct_lock_lock(&buckets[bucket].lock);
> > - conn_key_lookup(&buckets[bucket], &ctx, time_msec());
> > + ovs_mutex_lock(&ct_lock.lock);
> > + conn_key_lookup(&ctx.key, ctx.hash, time_msec(), &ctx.conn,
> &ctx.reply);
> > +
> > if (ctx.conn && ctx.conn->conn_type == CT_CONN_TYPE_DEFAULT) {
> > - conn_clean(ctx.conn, &buckets[bucket]);
> > + conn_clean(ctx.conn);
> > } else {
> > VLOG_WARN("Must flush tuple using the original pre-NATed
> tuple");
> > error = ENOENT;
> > }
> > - ct_lock_unlock(&buckets[bucket].lock);
> > + ovs_mutex_unlock(&ct_lock.lock);
> > return error;
> > }
> >
> > @@ -2693,22 +2489,22 @@ expectation_ref_create(struct hindex
> *alg_expectation_refs_,
> > }
> >
> > static void
> > -expectation_clean(const struct conn_key *master_key, uint32_t basis)
> > +expectation_clean(const struct conn_key *master_key)
> > {
> > - ct_rwlock_wrlock(&resources_lock);
> > + ovs_rwlock_wrlock(&resources_lock.lock);
> >
> > struct alg_exp_node *node, *next;
> > HINDEX_FOR_EACH_WITH_HASH_SAFE (node, next, node_ref,
> > - conn_key_hash(master_key, basis),
> > + conn_key_hash(master_key,
> hash_basis),
> > &alg_expectation_refs) {
> > if (!conn_key_cmp(&node->master_key, master_key)) {
> > - expectation_remove(&alg_expectations, &node->key, basis);
> > + expectation_remove(&alg_expectations, &node->key,
> hash_basis);
> > hindex_remove(&alg_expectation_refs, &node->node_ref);
> > free(node);
> > }
> > }
> >
> > - ct_rwlock_unlock(&resources_lock);
> > + ovs_rwlock_unlock(&resources_lock.lock);
> > }
> >
> > static void
> > @@ -2756,12 +2552,12 @@ expectation_create(ovs_be16 dst_port, const
> struct conn *master_conn,
> > /* Take the write lock here because it is almost 100%
> > * likely that the lookup will fail and
> > * expectation_create() will be called below. */
> > - ct_rwlock_wrlock(&resources_lock);
> > + ovs_rwlock_wrlock(&resources_lock.lock);
> > struct alg_exp_node *alg_exp = expectation_lookup(
> > &alg_expectations, &alg_exp_node->key, hash_basis, src_ip_wc);
> > if (alg_exp) {
> > free(alg_exp_node);
> > - ct_rwlock_unlock(&resources_lock);
> > + ovs_rwlock_unlock(&resources_lock.lock);
> > return;
> > }
> >
> > @@ -2770,7 +2566,7 @@ expectation_create(ovs_be16 dst_port, const struct
> conn *master_conn,
> > conn_key_hash(&alg_exp_node->key, hash_basis));
> > expectation_ref_create(&alg_expectation_refs, alg_exp_node,
> > hash_basis);
> > - ct_rwlock_unlock(&resources_lock);
> > + ovs_rwlock_unlock(&resources_lock.lock);
> > }
> >
> > static uint8_t
> > diff --git a/lib/conntrack.h b/lib/conntrack.h
> > index 80ba80e..58981bd 100644
> > --- a/lib/conntrack.h
> > +++ b/lib/conntrack.h
> > @@ -19,6 +19,7 @@
> >
> > #include <stdbool.h>
> >
> > +#include "cmap.h"
> > #include "latch.h"
> > #include "odp-netlink.h"
> > #include "openvswitch/hmap.h"
> > @@ -42,10 +43,6 @@
> > *
> > * conntrack_init();
> > *
> > - * It is necessary to periodically issue a call to
> > - *
> > - * to allow the module to clean up expired connections.
> > - *
> > * To send a group of packets through the connection tracker:
> > *
> > * conntrack_execute(pkt_batch, ...);
> > @@ -94,9 +91,8 @@ int conntrack_execute(struct dp_packet_batch
> *pkt_batch, ovs_be16 dl_type,
> > void conntrack_clear(struct dp_packet *packet);
> >
> > struct conntrack_dump {
> > - struct conntrack *ct;
> > unsigned bucket;
> > - struct hmap_position bucket_pos;
> > + struct cmap_position cm_pos;
> > bool filter_zone;
> > uint16_t zone;
> > };
> > @@ -114,103 +110,5 @@ int conntrack_set_maxconns(uint32_t maxconns);
> > int conntrack_get_maxconns(uint32_t *maxconns);
> > int conntrack_get_nconns(uint32_t *nconns);
> >
> > -/* 'struct ct_lock' is a wrapper for an adaptive mutex. It's useful to
> try
> > - * different types of locks (e.g. spinlocks) */
> > -
> > -struct OVS_LOCKABLE ct_lock {
> > - struct ovs_mutex lock;
> > -};
> > -
> > -struct OVS_LOCKABLE ct_rwlock {
> > - struct ovs_rwlock lock;
> > -};
> > -
> > -static inline void ct_lock_init(struct ct_lock *lock)
> > -{
> > - ovs_mutex_init_adaptive(&lock->lock);
> > -}
> > -
> > -static inline void ct_lock_lock(struct ct_lock *lock)
> > - OVS_ACQUIRES(lock)
> > - OVS_NO_THREAD_SAFETY_ANALYSIS
> > -{
> > - ovs_mutex_lock(&lock->lock);
> > -}
> > -
> > -static inline void ct_lock_unlock(struct ct_lock *lock)
> > - OVS_RELEASES(lock)
> > - OVS_NO_THREAD_SAFETY_ANALYSIS
> > -{
> > - ovs_mutex_unlock(&lock->lock);
> > -}
> > -
> > -static inline void ct_lock_destroy(struct ct_lock *lock)
> > -{
> > - ovs_mutex_destroy(&lock->lock);
> > -}
> > -
> > -static inline void ct_rwlock_init(struct ct_rwlock *lock)
> > -{
> > - ovs_rwlock_init(&lock->lock);
> > -}
> > -
> > -
> > -static inline void ct_rwlock_wrlock(struct ct_rwlock *lock)
> > - OVS_ACQ_WRLOCK(lock)
> > - OVS_NO_THREAD_SAFETY_ANALYSIS
> > -{
> > - ovs_rwlock_wrlock(&lock->lock);
> > -}
> > -
> > -static inline void ct_rwlock_rdlock(struct ct_rwlock *lock)
> > - OVS_ACQ_RDLOCK(lock)
> > - OVS_NO_THREAD_SAFETY_ANALYSIS
> > -{
> > - ovs_rwlock_rdlock(&lock->lock);
> > -}
> > -
> > -static inline void ct_rwlock_unlock(struct ct_rwlock *lock)
> > - OVS_RELEASES(lock)
> > - OVS_NO_THREAD_SAFETY_ANALYSIS
> > -{
> > - ovs_rwlock_unlock(&lock->lock);
> > -}
> > -
> > -static inline void ct_rwlock_destroy(struct ct_rwlock *lock)
> > -{
> > - ovs_rwlock_destroy(&lock->lock);
> > -}
> > -
> > -
> > -/* Timeouts: all the possible timeout states passed to
> update_expiration()
> > - * are listed here. The name will be prefix by CT_TM_ and the value is
> in
> > - * milliseconds */
> > -#define CT_TIMEOUTS \
> > - CT_TIMEOUT(TCP_FIRST_PACKET, 30 * 1000) \
> > - CT_TIMEOUT(TCP_OPENING, 30 * 1000) \
> > - CT_TIMEOUT(TCP_ESTABLISHED, 24 * 60 * 60 * 1000) \
> > - CT_TIMEOUT(TCP_CLOSING, 15 * 60 * 1000) \
> > - CT_TIMEOUT(TCP_FIN_WAIT, 45 * 1000) \
> > - CT_TIMEOUT(TCP_CLOSED, 30 * 1000) \
> > - CT_TIMEOUT(OTHER_FIRST, 60 * 1000) \
> > - CT_TIMEOUT(OTHER_MULTIPLE, 60 * 1000) \
> > - CT_TIMEOUT(OTHER_BIDIR, 30 * 1000) \
> > - CT_TIMEOUT(ICMP_FIRST, 60 * 1000) \
> > - CT_TIMEOUT(ICMP_REPLY, 30 * 1000)
> > -
> > -/* The smallest of the above values: it is used as an upper bound for
> the
> > - * interval between two rounds of cleanup of expired entries */
> > -#define CT_TM_MIN (30 * 1000)
> > -
> > -#define CT_TIMEOUT(NAME, VAL) BUILD_ASSERT_DECL(VAL >= CT_TM_MIN);
> > - CT_TIMEOUTS
> > -#undef CT_TIMEOUT
> > -
> > -enum ct_timeout {
> > -#define CT_TIMEOUT(NAME, VALUE) CT_TM_##NAME,
> > - CT_TIMEOUTS
> > -#undef CT_TIMEOUT
> > - N_CT_TM
> > -};
> >
> > #endif /* conntrack.h */
>
More information about the dev
mailing list