commit 212ed75dc5fb9d1423b3942c8f872a868cda3466 upstream.

The pipapo set backend follows copy-on-update approach, maintaining one clone of the existing datastructure that is being updated. The clone and current datastructures are swapped via rcu from the commit step.

The existing integration with the commit protocol is flawed because there is no operation to clean up the clone if the transaction is aborted. Moreover, the datastructure swap happens on set element activation.

This patch adds two new operations for sets: commit and abort, these new operations are invoked from the commit and abort steps, after the transactions have been digested, and it updates the pipapo set backend to use it.

This patch adds a new ->pending_update field to sets to maintain a list of sets that require this new commit and abort operations.

Fixes: 3c4287f62044 ("nf_tables: Add set type for arbitrary concatenation of ranges") Signed-off-by: Pablo Neira Ayuso pablo@netfilter.org --- include/net/netfilter/nf_tables.h | 4 ++- net/netfilter/nf_tables_api.c | 56 +++++++++++++++++++++++++++++++ net/netfilter/nft_set_pipapo.c | 55 +++++++++++++++++++++--------- 3 files changed, 99 insertions(+), 16 deletions(-)

diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h index eec29dd6681c..a3068ed0f316 100644 --- a/include/net/netfilter/nf_tables.h +++ b/include/net/netfilter/nf_tables.h @@ -373,7 +373,8 @@ struct nft_set_ops { const struct nft_set *set, const struct nft_set_elem *elem, unsigned int flags); - + void (*commit)(const struct nft_set *set); + void (*abort)(const struct nft_set *set); u64 (*privsize)(const struct nlattr * const nla[], const struct nft_set_desc *desc); bool (*estimate)(const struct nft_set_desc *desc, @@ -454,6 +455,7 @@ struct nft_set { u16 udlen; unsigned char *udata; struct nft_expr *expr; + struct list_head pending_update; /* runtime data below here */ const struct nft_set_ops *ops ____cacheline_aligned; u16 flags:14, diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index 2669999d1bc9..430dcd0f6c3b 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -4509,6 +4509,7 @@ static int nf_tables_newset(struct net *net, struct sock *nlsk, }

set->handle = nf_tables_alloc_handle(table); + INIT_LIST_HEAD(&set->pending_update);

err = nft_trans_set_add(&ctx, NFT_MSG_NEWSET, set); if (err < 0) @@ -8141,10 +8142,25 @@ static void nf_tables_commit_audit_log(struct list_head *adl, u32 generation) } }

+static void nft_set_commit_update(struct list_head *set_update_list) +{ + struct nft_set *set, *next; + + list_for_each_entry_safe(set, next, set_update_list, pending_update) { + list_del_init(&set->pending_update); + + if (!set->ops->commit) + continue; + + set->ops->commit(set); + } +} + static int nf_tables_commit(struct net *net, struct sk_buff *skb) { struct nftables_pernet *nft_net = net_generic(net, nf_tables_net_id); struct nft_trans *trans, *next; + LIST_HEAD(set_update_list); struct nft_trans_elem *te; struct nft_chain *chain; struct nft_table *table; @@ -8310,6 +8326,11 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb) nf_tables_setelem_notify(&trans->ctx, te->set, &te->elem, NFT_MSG_NEWSETELEM, 0); + if (te->set->ops->commit && + list_empty(&te->set->pending_update)) { + list_add_tail(&te->set->pending_update, + &set_update_list); + } nft_trans_destroy(trans); break; case NFT_MSG_DELSETELEM: @@ -8321,6 +8342,11 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb) te->set->ops->remove(net, te->set, &te->elem); atomic_dec(&te->set->nelems); te->set->ndeact--; + if (te->set->ops->commit && + list_empty(&te->set->pending_update)) { + list_add_tail(&te->set->pending_update, + &set_update_list); + } break; case NFT_MSG_NEWOBJ: if (nft_trans_obj_update(trans)) { @@ -8381,6 +8407,8 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb) } }

+ nft_set_commit_update(&set_update_list); + nft_commit_notify(net, NETLINK_CB(skb).portid); nf_tables_gen_notify(net, skb, NFT_MSG_NEWGEN); nf_tables_commit_audit_log(&adl, nft_net->base_seq); @@ -8437,10 +8465,25 @@ static void nf_tables_abort_release(struct nft_trans *trans) kfree(trans); }

+static void nft_set_abort_update(struct list_head *set_update_list) +{ + struct nft_set *set, *next; + + list_for_each_entry_safe(set, next, set_update_list, pending_update) { + list_del_init(&set->pending_update); + + if (!set->ops->abort) + continue; + + set->ops->abort(set); + } +} + static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action) { struct nftables_pernet *nft_net = net_generic(net, nf_tables_net_id); struct nft_trans *trans, *next; + LIST_HEAD(set_update_list); struct nft_trans_elem *te;

if (action == NFNL_ABORT_VALIDATE && @@ -8529,6 +8572,12 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action) te = (struct nft_trans_elem *)trans->data; te->set->ops->remove(net, te->set, &te->elem); atomic_dec(&te->set->nelems); + + if (te->set->ops->abort && + list_empty(&te->set->pending_update)) { + list_add_tail(&te->set->pending_update, + &set_update_list); + } break; case NFT_MSG_DELSETELEM: te = (struct nft_trans_elem *)trans->data; @@ -8537,6 +8586,11 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action) te->set->ops->activate(net, te->set, &te->elem); te->set->ndeact--;

+ if (te->set->ops->abort && + list_empty(&te->set->pending_update)) { + list_add_tail(&te->set->pending_update, + &set_update_list); + } nft_trans_destroy(trans); break; case NFT_MSG_NEWOBJ: @@ -8577,6 +8631,8 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action) } }

+ nft_set_abort_update(&set_update_list); + synchronize_rcu();

list_for_each_entry_safe_reverse(trans, next, diff --git a/net/netfilter/nft_set_pipapo.c b/net/netfilter/nft_set_pipapo.c index 50f840e312b0..ce6c07ea7244 100644 --- a/net/netfilter/nft_set_pipapo.c +++ b/net/netfilter/nft_set_pipapo.c @@ -1603,17 +1603,10 @@ static void pipapo_free_fields(struct nft_pipapo_match *m) } }

-/** - * pipapo_reclaim_match - RCU callback to free fields from old matching data - * @rcu: RCU head - */ -static void pipapo_reclaim_match(struct rcu_head *rcu) +static void pipapo_free_match(struct nft_pipapo_match *m) { - struct nft_pipapo_match *m; int i;

- m = container_of(rcu, struct nft_pipapo_match, rcu); - for_each_possible_cpu(i) kfree(*per_cpu_ptr(m->scratch, i));

@@ -1628,7 +1621,19 @@ static void pipapo_reclaim_match(struct rcu_head *rcu) }

/** - * pipapo_commit() - Replace lookup data with current working copy + * pipapo_reclaim_match - RCU callback to free fields from old matching data + * @rcu: RCU head + */ +static void pipapo_reclaim_match(struct rcu_head *rcu) +{ + struct nft_pipapo_match *m; + + m = container_of(rcu, struct nft_pipapo_match, rcu); + pipapo_free_match(m); +} + +/** + * nft_pipapo_commit() - Replace lookup data with current working copy * @set: nftables API set representation * * While at it, check if we should perform garbage collection on the working @@ -1638,7 +1643,7 @@ static void pipapo_reclaim_match(struct rcu_head *rcu) * We also need to create a new working copy for subsequent insertions and * deletions. */ -static void pipapo_commit(const struct nft_set *set) +static void nft_pipapo_commit(const struct nft_set *set) { struct nft_pipapo *priv = nft_set_priv(set); struct nft_pipapo_match *new_clone, *old; @@ -1663,6 +1668,26 @@ static void pipapo_commit(const struct nft_set *set) priv->clone = new_clone; }

+static void nft_pipapo_abort(const struct nft_set *set) +{ + struct nft_pipapo *priv = nft_set_priv(set); + struct nft_pipapo_match *new_clone, *m; + + if (!priv->dirty) + return; + + m = rcu_dereference(priv->match); + + new_clone = pipapo_clone(m); + if (IS_ERR(new_clone)) + return; + + priv->dirty = false; + + pipapo_free_match(priv->clone); + priv->clone = new_clone; +} + /** * nft_pipapo_activate() - Mark element reference as active given key, commit * @net: Network namespace @@ -1670,8 +1695,7 @@ static void pipapo_commit(const struct nft_set *set) * @elem: nftables API element representation containing key data * * On insertion, elements are added to a copy of the matching data currently - * in use for lookups, and not directly inserted into current lookup data, so - * we'll take care of that by calling pipapo_commit() here. Both + * in use for lookups, and not directly inserted into current lookup data. Both * nft_pipapo_insert() and nft_pipapo_activate() are called once for each * element, hence we can't purpose either one as a real commit operation. */ @@ -1687,8 +1711,6 @@ static void nft_pipapo_activate(const struct net *net,

nft_set_elem_change_active(net, set, &e->ext); nft_set_elem_clear_busy(&e->ext); - - pipapo_commit(set); }

/** @@ -1938,7 +1960,6 @@ static void nft_pipapo_remove(const struct net *net, const struct nft_set *set, if (i == m->field_count) { priv->dirty = true; pipapo_drop(m, rulemap); - pipapo_commit(set); return; }

@@ -2245,6 +2266,8 @@ const struct nft_set_type nft_set_pipapo_type = { .init = nft_pipapo_init, .destroy = nft_pipapo_destroy, .gc_init = nft_pipapo_gc_init, + .commit = nft_pipapo_commit, + .abort = nft_pipapo_abort, .elemsize = offsetof(struct nft_pipapo_elem, ext), }, }; @@ -2267,6 +2290,8 @@ const struct nft_set_type nft_set_pipapo_avx2_type = { .init = nft_pipapo_init, .destroy = nft_pipapo_destroy, .gc_init = nft_pipapo_gc_init, + .commit = nft_pipapo_commit, + .abort = nft_pipapo_abort, .elemsize = offsetof(struct nft_pipapo_elem, ext), }, };

commit 5f68718b34a531a556f2f50300ead2862278da26 upstream.

The set types rhashtable and rbtree use a GC worker to reclaim memory. From system work queue, in periodic intervals, a scan of the table is done.

The major caveat here is that the nft transaction mutex is not held. This causes a race between control plane and GC when they attempt to delete the same element.

We cannot grab the netlink mutex from the work queue, because the control plane has to wait for the GC work queue in case the set is to be removed, so we get following deadlock:

cpu 1 cpu2 GC work transaction comes in , lock nft mutex `acquire nft mutex // BLOCKS transaction asks to remove the set set destruction calls cancel_work_sync()

cancel_work_sync will now block forever, because it is waiting for the mutex the caller already owns.

This patch adds a new API that deals with garbage collection in two steps:

1) Lockless GC of expired elements sets on the NFT_SET_ELEM_DEAD_BIT so they are not visible via lookup. Annotate current GC sequence in the GC transaction. Enqueue GC transaction work as soon as it is full. If ruleset is updated, then GC transaction is aborted and retried later.

2) GC work grabs the mutex. If GC sequence has changed then this GC transaction lost race with control plane, abort it as it contains stale references to objects and let GC try again later. If the ruleset is intact, then this GC transaction deactivates and removes the elements and it uses call_rcu() to destroy elements.

Note that no elements are removed from GC lockless path, the _DEAD bit is set and pointers are collected. GC catchall does not remove the elements anymore too. There is a new set->dead flag that is set on to abort the GC transaction to deal with set->ops->destroy() path which removes the remaining elements in the set from commit_release, where no mutex is held.

To deal with GC when mutex is held, which allows safe deactivate and removal, add sync GC API which releases the set element object via call_rcu(). This is used by rbtree and pipapo backends which also perform garbage collection from control plane path.

Since element removal from sets can happen from control plane and element garbage collection/timeout, it is necessary to keep the set structure alive until all elements have been deactivated and destroyed.

We cannot do a cancel_work_sync or flush_work in nft_set_destroy because its called with the transaction mutex held, but the aforementioned async work queue might be blocked on the very mutex that nft_set_destroy() callchain is sitting on.

This gives us the choice of ABBA deadlock or UaF.

To avoid both, add set->refs refcount_t member. The GC API can then increment the set refcount and release it once the elements have been free'd.

Set backends are adapted to use the GC transaction API in a follow up patch entitled:

("netfilter: nf_tables: use gc transaction API in set backends")

This is joint work with Florian Westphal.

Fixes: cfed7e1b1f8e ("netfilter: nf_tables: add set garbage collection helpers") Signed-off-by: Pablo Neira Ayuso pablo@netfilter.org --- include/net/netfilter/nf_tables.h | 61 ++++++++- net/netfilter/nf_tables_api.c | 216 ++++++++++++++++++++++++++++-- 2 files changed, 267 insertions(+), 10 deletions(-)

diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h index a3068ed0f316..39a0b37e8a1a 100644 --- a/include/net/netfilter/nf_tables.h +++ b/include/net/netfilter/nf_tables.h @@ -407,6 +407,7 @@ struct nft_set_type { * * @list: table set list node * @bindings: list of set bindings + * @refs: internal refcounting for async set destruction * @table: table this set belongs to * @net: netnamespace this set belongs to * @name: name of the set @@ -436,6 +437,7 @@ struct nft_set_type { struct nft_set { struct list_head list; struct list_head bindings; + refcount_t refs; struct nft_table *table; possible_net_t net; char *name; @@ -458,7 +460,8 @@ struct nft_set { struct list_head pending_update; /* runtime data below here */ const struct nft_set_ops *ops ____cacheline_aligned; - u16 flags:14, + u16 flags:13, + dead:1, genmask:2; u8 klen; u8 dlen; @@ -1450,6 +1453,32 @@ static inline void nft_set_elem_clear_busy(struct nft_set_ext *ext) clear_bit(NFT_SET_ELEM_BUSY_BIT, word); }

+#define NFT_SET_ELEM_DEAD_MASK (1 << 3) + +#if defined(__LITTLE_ENDIAN_BITFIELD) +#define NFT_SET_ELEM_DEAD_BIT 3 +#elif defined(__BIG_ENDIAN_BITFIELD) +#define NFT_SET_ELEM_DEAD_BIT (BITS_PER_LONG - BITS_PER_BYTE + 3) +#else +#error +#endif + +static inline void nft_set_elem_dead(struct nft_set_ext *ext) +{ + unsigned long *word = (unsigned long *)ext; + + BUILD_BUG_ON(offsetof(struct nft_set_ext, genmask) != 0); + set_bit(NFT_SET_ELEM_DEAD_BIT, word); +} + +static inline int nft_set_elem_is_dead(const struct nft_set_ext *ext) +{ + unsigned long *word = (unsigned long *)ext; + + BUILD_BUG_ON(offsetof(struct nft_set_ext, genmask) != 0); + return test_bit(NFT_SET_ELEM_DEAD_BIT, word); +} + /** * struct nft_trans - nf_tables object update in transaction * @@ -1575,6 +1604,35 @@ struct nft_trans_flowtable { #define nft_trans_flowtable_flags(trans) \ (((struct nft_trans_flowtable *)trans->data)->flags)

+#define NFT_TRANS_GC_BATCHCOUNT 256 + +struct nft_trans_gc { + struct list_head list; + struct net *net; + struct nft_set *set; + u32 seq; + u8 count; + void *priv[NFT_TRANS_GC_BATCHCOUNT]; + struct rcu_head rcu; +}; + +struct nft_trans_gc *nft_trans_gc_alloc(struct nft_set *set, + unsigned int gc_seq, gfp_t gfp); +void nft_trans_gc_destroy(struct nft_trans_gc *trans); + +struct nft_trans_gc *nft_trans_gc_queue_async(struct nft_trans_gc *gc, + unsigned int gc_seq, gfp_t gfp); +void nft_trans_gc_queue_async_done(struct nft_trans_gc *gc); + +struct nft_trans_gc *nft_trans_gc_queue_sync(struct nft_trans_gc *gc, gfp_t gfp); +void nft_trans_gc_queue_sync_done(struct nft_trans_gc *trans); + +void nft_trans_gc_elem_add(struct nft_trans_gc *gc, void *priv); + +void nft_setelem_data_deactivate(const struct net *net, + const struct nft_set *set, + struct nft_set_elem *elem); + int __init nft_chain_filter_init(void); void nft_chain_filter_fini(void);

@@ -1595,6 +1653,7 @@ struct nftables_pernet { struct mutex commit_mutex; unsigned int base_seq; u8 validate_state; + unsigned int gc_seq; };

#endif /* _NET_NF_TABLES_H */ diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index 5eef671578a2..1f06dd065d75 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -32,7 +32,9 @@ static LIST_HEAD(nf_tables_expressions); static LIST_HEAD(nf_tables_objects); static LIST_HEAD(nf_tables_flowtables); static LIST_HEAD(nf_tables_destroy_list); +static LIST_HEAD(nf_tables_gc_list); static DEFINE_SPINLOCK(nf_tables_destroy_list_lock); +static DEFINE_SPINLOCK(nf_tables_gc_list_lock); static u64 table_handle;

enum { @@ -124,6 +126,9 @@ static void nft_validate_state_update(struct net *net, u8 new_validate_state) static void nf_tables_trans_destroy_work(struct work_struct *w); static DECLARE_WORK(trans_destroy_work, nf_tables_trans_destroy_work);

+static void nft_trans_gc_work(struct work_struct *work); +static DECLARE_WORK(trans_gc_work, nft_trans_gc_work); + static void nft_ctx_init(struct nft_ctx *ctx, struct net *net, const struct sk_buff *skb, @@ -559,10 +564,6 @@ static int nft_trans_set_add(const struct nft_ctx *ctx, int msg_type, return 0; }

-static void nft_setelem_data_deactivate(const struct net *net, - const struct nft_set *set, - struct nft_set_elem *elem); - static int nft_mapelem_deactivate(const struct nft_ctx *ctx, struct nft_set *set, const struct nft_set_iter *iter, @@ -4474,6 +4475,7 @@ static int nf_tables_newset(struct net *net, struct sock *nlsk, }

INIT_LIST_HEAD(&set->bindings); + refcount_set(&set->refs, 1); set->table = table; write_pnet(&set->net, net); set->ops = ops; @@ -4534,6 +4536,14 @@ static int nf_tables_newset(struct net *net, struct sock *nlsk, return err; }

+static void nft_set_put(struct nft_set *set) +{ + if (refcount_dec_and_test(&set->refs)) { + kfree(set->name); + kvfree(set); + } +} + static void nft_set_destroy(const struct nft_ctx *ctx, struct nft_set *set) { if (WARN_ON(set->use > 0)) @@ -4543,8 +4553,7 @@ static void nft_set_destroy(const struct nft_ctx *ctx, struct nft_set *set) nft_expr_destroy(ctx, set->expr);

set->ops->destroy(ctx, set); - kfree(set->name); - kvfree(set); + nft_set_put(set); }

static int nf_tables_delset(struct net *net, struct sock *nlsk, @@ -5768,9 +5777,9 @@ static void nft_setelem_data_activate(const struct net *net, nft_use_inc_restore(&(*nft_set_ext_obj(ext))->use); }

-static void nft_setelem_data_deactivate(const struct net *net, - const struct nft_set *set, - struct nft_set_elem *elem) +void nft_setelem_data_deactivate(const struct net *net, + const struct nft_set *set, + struct nft_set_elem *elem) { const struct nft_set_ext *ext = nft_set_elem_ext(set, elem->priv);

@@ -8002,6 +8011,179 @@ void nft_chain_del(struct nft_chain *chain) list_del_rcu(&chain->list); }

+static void nft_trans_gc_setelem_remove(struct nft_ctx *ctx, + struct nft_trans_gc *trans) +{ + void **priv = trans->priv; + unsigned int i; + + for (i = 0; i < trans->count; i++) { + struct nft_set_elem elem = { + .priv = priv[i], + }; + + nft_setelem_data_deactivate(ctx->net, trans->set, &elem); + trans->set->ops->remove(trans->net, trans->set, &elem); + } +} + +void nft_trans_gc_destroy(struct nft_trans_gc *trans) +{ + nft_set_put(trans->set); + put_net(trans->net); + kfree(trans); +} + +static void nft_trans_gc_trans_free(struct rcu_head *rcu) +{ + struct nft_set_elem elem = {}; + struct nft_trans_gc *trans; + struct nft_ctx ctx = {}; + unsigned int i; + + trans = container_of(rcu, struct nft_trans_gc, rcu); + ctx.net = read_pnet(&trans->set->net); + + for (i = 0; i < trans->count; i++) { + elem.priv = trans->priv[i]; + atomic_dec(&trans->set->nelems); + + nf_tables_set_elem_destroy(&ctx, trans->set, elem.priv); + } + + nft_trans_gc_destroy(trans); +} + +static bool nft_trans_gc_work_done(struct nft_trans_gc *trans) +{ + struct nftables_pernet *nft_net; + struct nft_ctx ctx = {}; + + nft_net = net_generic(trans->net, nf_tables_net_id); + + mutex_lock(&nft_net->commit_mutex); + + /* Check for race with transaction, otherwise this batch refers to + * stale objects that might not be there anymore. Skip transaction if + * set has been destroyed from control plane transaction in case gc + * worker loses race. + */ + if (READ_ONCE(nft_net->gc_seq) != trans->seq || trans->set->dead) { + mutex_unlock(&nft_net->commit_mutex); + return false; + } + + ctx.net = trans->net; + ctx.table = trans->set->table; + + nft_trans_gc_setelem_remove(&ctx, trans); + mutex_unlock(&nft_net->commit_mutex); + + return true; +} + +static void nft_trans_gc_work(struct work_struct *work) +{ + struct nft_trans_gc *trans, *next; + LIST_HEAD(trans_gc_list); + + spin_lock(&nf_tables_destroy_list_lock); + list_splice_init(&nf_tables_gc_list, &trans_gc_list); + spin_unlock(&nf_tables_destroy_list_lock); + + list_for_each_entry_safe(trans, next, &trans_gc_list, list) { + list_del(&trans->list); + if (!nft_trans_gc_work_done(trans)) { + nft_trans_gc_destroy(trans); + continue; + } + call_rcu(&trans->rcu, nft_trans_gc_trans_free); + } +} + +struct nft_trans_gc *nft_trans_gc_alloc(struct nft_set *set, + unsigned int gc_seq, gfp_t gfp) +{ + struct net *net = read_pnet(&set->net); + struct nft_trans_gc *trans; + + trans = kzalloc(sizeof(*trans), gfp); + if (!trans) + return NULL; + + refcount_inc(&set->refs); + trans->set = set; + trans->net = get_net(net); + trans->seq = gc_seq; + + return trans; +} + +void nft_trans_gc_elem_add(struct nft_trans_gc *trans, void *priv) +{ + trans->priv[trans->count++] = priv; +} + +static void nft_trans_gc_queue_work(struct nft_trans_gc *trans) +{ + spin_lock(&nf_tables_gc_list_lock); + list_add_tail(&trans->list, &nf_tables_gc_list); + spin_unlock(&nf_tables_gc_list_lock); + + schedule_work(&trans_gc_work); +} + +static int nft_trans_gc_space(struct nft_trans_gc *trans) +{ + return NFT_TRANS_GC_BATCHCOUNT - trans->count; +} + +struct nft_trans_gc *nft_trans_gc_queue_async(struct nft_trans_gc *gc, + unsigned int gc_seq, gfp_t gfp) +{ + if (nft_trans_gc_space(gc)) + return gc; + + nft_trans_gc_queue_work(gc); + + return nft_trans_gc_alloc(gc->set, gc_seq, gfp); +} + +void nft_trans_gc_queue_async_done(struct nft_trans_gc *trans) +{ + if (trans->count == 0) { + nft_trans_gc_destroy(trans); + return; + } + + nft_trans_gc_queue_work(trans); +} + +struct nft_trans_gc *nft_trans_gc_queue_sync(struct nft_trans_gc *gc, gfp_t gfp) +{ + if (WARN_ON_ONCE(!lockdep_commit_lock_is_held(gc->net))) + return NULL; + + if (nft_trans_gc_space(gc)) + return gc; + + call_rcu(&gc->rcu, nft_trans_gc_trans_free); + + return nft_trans_gc_alloc(gc->set, 0, gfp); +} + +void nft_trans_gc_queue_sync_done(struct nft_trans_gc *trans) +{ + WARN_ON_ONCE(!lockdep_commit_lock_is_held(trans->net)); + + if (trans->count == 0) { + nft_trans_gc_destroy(trans); + return; + } + + call_rcu(&trans->rcu, nft_trans_gc_trans_free); +} + static void nf_tables_module_autoload_cleanup(struct net *net) { struct nftables_pernet *nft_net = net_generic(net, nf_tables_net_id); @@ -8168,6 +8350,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb) struct nft_trans_elem *te; struct nft_chain *chain; struct nft_table *table; + unsigned int gc_seq; LIST_HEAD(adl); int err;

@@ -8240,6 +8423,10 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb) while (++nft_net->base_seq == 0) ;

+ /* Bump gc counter, it becomes odd, this is the busy mark. */ + gc_seq = READ_ONCE(nft_net->gc_seq); + WRITE_ONCE(nft_net->gc_seq, ++gc_seq); + /* step 3. Start new generation, rules_gen_X now in use. */ net->nft.gencursor = nft_gencursor_next(net);

@@ -8319,6 +8506,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb) nft_trans_destroy(trans); break; case NFT_MSG_DELSET: + nft_trans_set(trans)->dead = 1; list_del_rcu(&nft_trans_set(trans)->list); nf_tables_set_notify(&trans->ctx, nft_trans_set(trans), NFT_MSG_DELSET, GFP_KERNEL); @@ -8416,6 +8604,8 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb) nft_commit_notify(net, NETLINK_CB(skb).portid); nf_tables_gen_notify(net, skb, NFT_MSG_NEWGEN); nf_tables_commit_audit_log(&adl, nft_net->base_seq); + + WRITE_ONCE(nft_net->gc_seq, ++gc_seq); nf_tables_commit_release(net);

return 0; @@ -9353,6 +9543,7 @@ static int __net_init nf_tables_init_net(struct net *net) mutex_init(&nft_net->commit_mutex); nft_net->base_seq = 1; nft_net->validate_state = NFT_VALIDATE_SKIP; + nft_net->gc_seq = 0;

return 0; } @@ -9380,10 +9571,16 @@ static void __net_exit nf_tables_exit_net(struct net *net) WARN_ON_ONCE(!list_empty(&nft_net->notify_list)); }

+static void nf_tables_exit_batch(struct list_head *net_exit_list) +{ + flush_work(&trans_gc_work); +} + static struct pernet_operations nf_tables_net_ops = { .init = nf_tables_init_net, .pre_exit = nf_tables_pre_exit_net, .exit = nf_tables_exit_net, + .exit_batch = nf_tables_exit_batch, .id = &nf_tables_net_id, .size = sizeof(struct nftables_pernet), }; @@ -9448,6 +9645,7 @@ static void __exit nf_tables_module_exit(void) nft_chain_filter_fini(); nft_chain_route_fini(); unregister_pernet_subsys(&nf_tables_net_ops); + cancel_work_sync(&trans_gc_work); cancel_work_sync(&trans_destroy_work); rcu_barrier(); rhltable_destroy(&nft_objname_ht);

commit c92db3030492b8ad1d0faace7a93bbcf53850d0c upstream.

Set on the NFT_SET_ELEM_DEAD_BIT flag on this element, instead of performing element removal which might race with an ongoing transaction. Enable gc when dynamic flag is set on since dynset deletion requires garbage collection after this patch.

Fixes: d0a8d877da97 ("netfilter: nft_dynset: support for element deletion") Signed-off-by: Pablo Neira Ayuso pablo@netfilter.org --- net/netfilter/nft_set_hash.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/net/netfilter/nft_set_hash.c b/net/netfilter/nft_set_hash.c index 6ae99b3107bc..9cdf348b048a 100644 --- a/net/netfilter/nft_set_hash.c +++ b/net/netfilter/nft_set_hash.c @@ -251,7 +251,9 @@ static bool nft_rhash_delete(const struct nft_set *set, if (he == NULL) return false;

- return rhashtable_remove_fast(&priv->ht, &he->node, nft_rhash_params) == 0; + nft_set_elem_dead(&he->ext); + + return true; }

static void nft_rhash_walk(const struct nft_ctx *ctx, struct nft_set *set, @@ -398,7 +400,7 @@ static int nft_rhash_init(const struct nft_set *set, return err;

INIT_DEFERRABLE_WORK(&priv->gc_work, nft_rhash_gc); - if (set->flags & NFT_SET_TIMEOUT) + if (set->flags & (NFT_SET_TIMEOUT | NFT_SET_EVAL)) nft_rhash_gc_init(set);

return 0;

From: Florian Westphal fw@strlen.de

commit 7845914f45f066497ac75b30c50dbc735e84e884 upstream.

nftables selftests fail: run-tests.sh testcases/sets/0044interval_overlap_0 Expected: 0-2 . 0-3, got: W: [FAILED] ./testcases/sets/0044interval_overlap_0: got 1

Insertion must ignore duplicate but expired entries.

Moreover, there is a strange asymmetry in nft_pipapo_activate:

It refetches the current element, whereas the other ->activate callbacks (bitmap, hash, rhash, rbtree) use elem->priv. Same for .remove: other set implementations take elem->priv, nft_pipapo_remove fetches elem->priv, then does a relookup, remove this.

I suspect this was the reason for the change that prompted the removal of the expired check in pipapo_get() in the first place, but skipping exired elements there makes no sense to me, this helper is used for normal get requests, insertions (duplicate check) and deactivate callback.

In first two cases expired elements must be skipped.

For ->deactivate(), this gets called for DELSETELEM, so it seems to me that expired elements should be skipped as well, i.e. delete request should fail with -ENOENT error.

Fixes: 24138933b97b ("netfilter: nf_tables: don't skip expired elements during walk") Signed-off-by: Florian Westphal fw@strlen.de Signed-off-by: Pablo Neira Ayuso pablo@netfilter.org --- net/netfilter/nft_set_pipapo.c | 25 +++++-------------------- 1 file changed, 5 insertions(+), 20 deletions(-)

diff --git a/net/netfilter/nft_set_pipapo.c b/net/netfilter/nft_set_pipapo.c index 63d0723950d3..80440ac5d44c 100644 --- a/net/netfilter/nft_set_pipapo.c +++ b/net/netfilter/nft_set_pipapo.c @@ -566,6 +566,8 @@ static struct nft_pipapo_elem *pipapo_get(const struct net *net, goto out;

if (last) { + if (nft_set_elem_expired(&f->mt[b].e->ext)) + goto next_match; if ((genmask && !nft_set_elem_active(&f->mt[b].e->ext, genmask))) goto next_match; @@ -600,17 +602,8 @@ static struct nft_pipapo_elem *pipapo_get(const struct net *net, static void *nft_pipapo_get(const struct net *net, const struct nft_set *set, const struct nft_set_elem *elem, unsigned int flags) { - struct nft_pipapo_elem *ret; - - ret = pipapo_get(net, set, (const u8 *)elem->key.val.data, - nft_genmask_cur(net)); - if (IS_ERR(ret)) - return ret; - - if (nft_set_elem_expired(&ret->ext)) - return ERR_PTR(-ENOENT); - - return ret; + return pipapo_get(net, set, (const u8 *)elem->key.val.data, + nft_genmask_cur(net)); }

/** @@ -1737,11 +1730,7 @@ static void nft_pipapo_activate(const struct net *net, const struct nft_set *set, const struct nft_set_elem *elem) { - struct nft_pipapo_elem *e; - - e = pipapo_get(net, set, (const u8 *)elem->key.val.data, 0); - if (IS_ERR(e)) - return; + struct nft_pipapo_elem *e = elem->priv;

nft_set_elem_change_active(net, set, &e->ext); } @@ -1955,10 +1944,6 @@ static void nft_pipapo_remove(const struct net *net, const struct nft_set *set,

data = (const u8 *)nft_set_ext_key(&e->ext);

- e = pipapo_get(net, set, data, 0); - if (IS_ERR(e)) - return; - while ((rules_f0 = pipapo_rules_same_key(m->f, first_rule))) { union nft_pipapo_map_bucket rulemap[NFT_PIPAPO_MAX_FIELDS]; const u8 *match_start, *match_end;

commit 02c6c24402bf1c1e986899c14ba22a10b510916b upstream.

Use maybe_get_net() since GC workqueue might race with netns exit path.

Fixes: 5f68718b34a5 ("netfilter: nf_tables: GC transaction API to avoid race with control plane") Signed-off-by: Pablo Neira Ayuso pablo@netfilter.org Signed-off-by: Florian Westphal fw@strlen.de --- net/netfilter/nf_tables_api.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index 43da2f0a5262..78bf82f89ecd 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -8089,9 +8089,14 @@ struct nft_trans_gc *nft_trans_gc_alloc(struct nft_set *set, if (!trans) return NULL;

+ trans->net = maybe_get_net(net); + if (!trans->net) { + kfree(trans); + return NULL; + } + refcount_inc(&set->refs); trans->set = set; - trans->net = get_net(net); trans->seq = gc_seq;

return trans;

commit 8357bc946a2abc2a10ca40e5a2105d2b4c57515e upstream.

Use nf_tables_gc_list_lock spinlock, not nf_tables_destroy_list_lock to protect the gc_list.

Fixes: 5f68718b34a5 ("netfilter: nf_tables: GC transaction API to avoid race with control plane") Signed-off-by: Pablo Neira Ayuso pablo@netfilter.org --- net/netfilter/nf_tables_api.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index 1f67931b86d8..9fc302a6836b 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -8065,9 +8065,9 @@ static void nft_trans_gc_work(struct work_struct *work) struct nft_trans_gc *trans, *next; LIST_HEAD(trans_gc_list);

- spin_lock(&nf_tables_destroy_list_lock); + spin_lock(&nf_tables_gc_list_lock); list_splice_init(&nf_tables_gc_list, &trans_gc_list); - spin_unlock(&nf_tables_destroy_list_lock); + spin_unlock(&nf_tables_gc_list_lock);

list_for_each_entry_safe(trans, next, &trans_gc_list, list) { list_del(&trans->list);

commit 2ee52ae94baabf7ee09cf2a8d854b990dac5d0e4 upstream.

New elements in this transaction might expired before such transaction ends. Skip sync GC for such elements otherwise commit path might walk over an already released object. Once transaction is finished, async GC will collect such expired element.

Fixes: f6c383b8c31a ("netfilter: nf_tables: adapt set backend to use GC transaction API") Signed-off-by: Pablo Neira Ayuso pablo@netfilter.org Signed-off-by: Florian Westphal fw@strlen.de Signed-off-by: Pablo Neira Ayuso pablo@netfilter.org --- net/netfilter/nft_set_rbtree.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/net/netfilter/nft_set_rbtree.c b/net/netfilter/nft_set_rbtree.c index 9b0bdd421615..535076b4de53 100644 --- a/net/netfilter/nft_set_rbtree.c +++ b/net/netfilter/nft_set_rbtree.c @@ -314,6 +314,7 @@ static int __nft_rbtree_insert(const struct net *net, const struct nft_set *set, struct nft_rbtree_elem *rbe, *rbe_le = NULL, *rbe_ge = NULL; struct rb_node *node, *next, *parent, **p, *first = NULL; struct nft_rbtree *priv = nft_set_priv(set); + u8 cur_genmask = nft_genmask_cur(net); u8 genmask = nft_genmask_next(net); int d, err;

@@ -359,8 +360,11 @@ static int __nft_rbtree_insert(const struct net *net, const struct nft_set *set, if (!nft_set_elem_active(&rbe->ext, genmask)) continue;

- /* perform garbage collection to avoid bogus overlap reports. */ - if (nft_set_elem_expired(&rbe->ext)) { + /* perform garbage collection to avoid bogus overlap reports + * but skip new elements in this transaction. + */ + if (nft_set_elem_expired(&rbe->ext) && + nft_set_elem_active(&rbe->ext, cur_genmask)) { err = nft_rbtree_gc_elem(set, priv, rbe, genmask); if (err < 0) return err;

commit 6d365eabce3c018a80f6e0379b17df2abb17405e upstream.

nft_trans_gc_queue_sync() enqueues the GC transaction and it allocates a new one. If this allocation fails, then stop this GC sync run and retry later.

Fixes: 5f68718b34a5 ("netfilter: nf_tables: GC transaction API to avoid race with control plane") Signed-off-by: Pablo Neira Ayuso pablo@netfilter.org --- net/netfilter/nft_set_pipapo.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/netfilter/nft_set_pipapo.c b/net/netfilter/nft_set_pipapo.c index 80440ac5d44c..fbfcc3275cad 100644 --- a/net/netfilter/nft_set_pipapo.c +++ b/net/netfilter/nft_set_pipapo.c @@ -1595,7 +1595,7 @@ static void pipapo_gc(const struct nft_set *_set, struct nft_pipapo_match *m)

gc = nft_trans_gc_queue_sync(gc, GFP_ATOMIC); if (!gc) - break; + return;

nft_pipapo_gc_deactivate(net, set, e); pipapo_drop(m, rulemap);

From: Florian Westphal fw@strlen.de

commit cf5000a7787cbc10341091d37245a42c119d26c5 upstream.

When more than 255 elements expired we're supposed to switch to a new gc container structure.

This never happens: u8 type will wrap before reaching the boundary and nft_trans_gc_space() always returns true.

This means we recycle the initial gc container structure and lose track of the elements that came before.

While at it, don't deref 'gc' after we've passed it to call_rcu.

Fixes: 5f68718b34a5 ("netfilter: nf_tables: GC transaction API to avoid race with control plane") Reported-by: Pablo Neira Ayuso pablo@netfilter.org Signed-off-by: Florian Westphal fw@strlen.de Signed-off-by: Pablo Neira Ayuso pablo@netfilter.org --- include/net/netfilter/nf_tables.h | 2 +- net/netfilter/nf_tables_api.c | 10 ++++++++-- 2 files changed, 9 insertions(+), 3 deletions(-)

diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h index bbe472c07d07..5619642b9ad4 100644 --- a/include/net/netfilter/nf_tables.h +++ b/include/net/netfilter/nf_tables.h @@ -1525,7 +1525,7 @@ struct nft_trans_gc { struct net *net; struct nft_set *set; u32 seq; - u8 count; + u16 count; void *priv[NFT_TRANS_GC_BATCHCOUNT]; struct rcu_head rcu; }; diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index 9fc302a6836b..32c97cc87ddc 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -8124,12 +8124,15 @@ static int nft_trans_gc_space(struct nft_trans_gc *trans) struct nft_trans_gc *nft_trans_gc_queue_async(struct nft_trans_gc *gc, unsigned int gc_seq, gfp_t gfp) { + struct nft_set *set; + if (nft_trans_gc_space(gc)) return gc;

+ set = gc->set; nft_trans_gc_queue_work(gc);

- return nft_trans_gc_alloc(gc->set, gc_seq, gfp); + return nft_trans_gc_alloc(set, gc_seq, gfp); }

void nft_trans_gc_queue_async_done(struct nft_trans_gc *trans) @@ -8144,15 +8147,18 @@ void nft_trans_gc_queue_async_done(struct nft_trans_gc *trans)

struct nft_trans_gc *nft_trans_gc_queue_sync(struct nft_trans_gc *gc, gfp_t gfp) { + struct nft_set *set; + if (WARN_ON_ONCE(!lockdep_commit_lock_is_held(gc->net))) return NULL;

if (nft_trans_gc_space(gc)) return gc;

+ set = gc->set; call_rcu(&gc->rcu, nft_trans_gc_trans_free);

- return nft_trans_gc_alloc(gc->set, 0, gfp); + return nft_trans_gc_alloc(set, 0, gfp); }

void nft_trans_gc_queue_sync_done(struct nft_trans_gc *trans)