5.10-stable review patch. If anyone has any objections, please let me know.
------------------
From: Pablo Neira Ayuso pablo@netfilter.org
commit 7395dfacfff65e9938ac0889dafa1ab01e987d15 upstream
Add a timestamp field at the beginning of the transaction, store it in the nftables per-netns area.
Update set backend .insert, .deactivate and sync gc path to use the timestamp, this avoids that an element expires while control plane transaction is still unfinished.
.lookup and .update, which are used from packet path, still use the current time to check if the element has expired. And .get path and dump also since this runs lockless under rcu read size lock. Then, there is async gc which also needs to check the current time since it runs asynchronously from a workqueue.
[ NB: rbtree GC updates has been excluded because GC is asynchronous. ]
Fixes: c3e1b005ed1c ("netfilter: nf_tables: add set element timeout support") Signed-off-by: Pablo Neira Ayuso pablo@netfilter.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- include/net/netfilter/nf_tables.h | 21 +++++++++++++++++++-- net/netfilter/nf_tables_api.c | 1 + net/netfilter/nft_set_hash.c | 8 +++++++- net/netfilter/nft_set_pipapo.c | 18 +++++++++++------- net/netfilter/nft_set_rbtree.c | 6 ++++-- 5 files changed, 42 insertions(+), 12 deletions(-)
--- a/include/net/netfilter/nf_tables.h +++ b/include/net/netfilter/nf_tables.h @@ -13,6 +13,7 @@ #include <net/netfilter/nf_flow_table.h> #include <net/netlink.h> #include <net/flow_offload.h> +#include <net/netns/generic.h>
#define NFT_MAX_HOOKS (NF_INET_INGRESS + 1)
@@ -686,10 +687,16 @@ static inline struct nft_expr *nft_set_e return nft_set_ext(ext, NFT_SET_EXT_EXPR); }
-static inline bool nft_set_elem_expired(const struct nft_set_ext *ext) +static inline bool __nft_set_elem_expired(const struct nft_set_ext *ext, + u64 tstamp) { return nft_set_ext_exists(ext, NFT_SET_EXT_EXPIRATION) && - time_is_before_eq_jiffies64(*nft_set_ext_expiration(ext)); + time_after_eq64(tstamp, *nft_set_ext_expiration(ext)); +} + +static inline bool nft_set_elem_expired(const struct nft_set_ext *ext) +{ + return __nft_set_elem_expired(ext, get_jiffies_64()); }
static inline struct nft_set_ext *nft_set_elem_ext(const struct nft_set *set, @@ -1580,9 +1587,19 @@ struct nftables_pernet { struct list_head module_list; struct list_head notify_list; struct mutex commit_mutex; + u64 tstamp; unsigned int base_seq; u8 validate_state; unsigned int gc_seq; };
+extern unsigned int nf_tables_net_id; + +static inline u64 nft_net_tstamp(const struct net *net) +{ + struct nftables_pernet *nft_net = net_generic(net, nf_tables_net_id); + + return nft_net->tstamp; +} + #endif /* _NET_NF_TABLES_H */ --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -9034,6 +9034,7 @@ static bool nf_tables_valid_genid(struct bool genid_ok;
mutex_lock(&nft_net->commit_mutex); + nft_net->tstamp = get_jiffies_64();
genid_ok = genid == 0 || nft_net->base_seq == genid; if (!genid_ok) --- a/net/netfilter/nft_set_hash.c +++ b/net/netfilter/nft_set_hash.c @@ -38,6 +38,7 @@ struct nft_rhash_cmp_arg { const struct nft_set *set; const u32 *key; u8 genmask; + u64 tstamp; };
static inline u32 nft_rhash_key(const void *data, u32 len, u32 seed) @@ -64,7 +65,7 @@ static inline int nft_rhash_cmp(struct r return 1; if (nft_set_elem_is_dead(&he->ext)) return 1; - if (nft_set_elem_expired(&he->ext)) + if (__nft_set_elem_expired(&he->ext, x->tstamp)) return 1; if (!nft_set_elem_active(&he->ext, x->genmask)) return 1; @@ -88,6 +89,7 @@ static bool nft_rhash_lookup(const struc .genmask = nft_genmask_cur(net), .set = set, .key = key, + .tstamp = get_jiffies_64(), };
he = rhashtable_lookup(&priv->ht, &arg, nft_rhash_params); @@ -106,6 +108,7 @@ static void *nft_rhash_get(const struct .genmask = nft_genmask_cur(net), .set = set, .key = elem->key.val.data, + .tstamp = get_jiffies_64(), };
he = rhashtable_lookup(&priv->ht, &arg, nft_rhash_params); @@ -129,6 +132,7 @@ static bool nft_rhash_update(struct nft_ .genmask = NFT_GENMASK_ANY, .set = set, .key = key, + .tstamp = get_jiffies_64(), };
he = rhashtable_lookup(&priv->ht, &arg, nft_rhash_params); @@ -172,6 +176,7 @@ static int nft_rhash_insert(const struct .genmask = nft_genmask_next(net), .set = set, .key = elem->key.val.data, + .tstamp = nft_net_tstamp(net), }; struct nft_rhash_elem *prev;
@@ -214,6 +219,7 @@ static void *nft_rhash_deactivate(const .genmask = nft_genmask_next(net), .set = set, .key = elem->key.val.data, + .tstamp = nft_net_tstamp(net), };
rcu_read_lock(); --- a/net/netfilter/nft_set_pipapo.c +++ b/net/netfilter/nft_set_pipapo.c @@ -504,6 +504,7 @@ out: * @set: nftables API set representation * @data: Key data to be matched against existing elements * @genmask: If set, check that element is active in given genmask + * @tstamp: timestamp to check for expired elements * * This is essentially the same as the lookup function, except that it matches * key data against the uncommitted copy and doesn't use preallocated maps for @@ -513,7 +514,8 @@ out: */ static struct nft_pipapo_elem *pipapo_get(const struct net *net, const struct nft_set *set, - const u8 *data, u8 genmask) + const u8 *data, u8 genmask, + u64 tstamp) { struct nft_pipapo_elem *ret = ERR_PTR(-ENOENT); struct nft_pipapo *priv = nft_set_priv(set); @@ -566,7 +568,7 @@ next_match: goto out;
if (last) { - if (nft_set_elem_expired(&f->mt[b].e->ext)) + if (__nft_set_elem_expired(&f->mt[b].e->ext, tstamp)) goto next_match; if ((genmask && !nft_set_elem_active(&f->mt[b].e->ext, genmask))) @@ -603,7 +605,7 @@ static void *nft_pipapo_get(const struct const struct nft_set_elem *elem, unsigned int flags) { return pipapo_get(net, set, (const u8 *)elem->key.val.data, - nft_genmask_cur(net)); + nft_genmask_cur(net), get_jiffies_64()); }
/** @@ -1197,6 +1199,7 @@ static int nft_pipapo_insert(const struc struct nft_pipapo *priv = nft_set_priv(set); struct nft_pipapo_match *m = priv->clone; u8 genmask = nft_genmask_next(net); + u64 tstamp = nft_net_tstamp(net); struct nft_pipapo_field *f; const u8 *start_p, *end_p; int i, bsize_max, err = 0; @@ -1206,7 +1209,7 @@ static int nft_pipapo_insert(const struc else end = start;
- dup = pipapo_get(net, set, start, genmask); + dup = pipapo_get(net, set, start, genmask, tstamp); if (!IS_ERR(dup)) { /* Check if we already have the same exact entry */ const struct nft_data *dup_key, *dup_end; @@ -1228,7 +1231,7 @@ static int nft_pipapo_insert(const struc
if (PTR_ERR(dup) == -ENOENT) { /* Look for partially overlapping entries */ - dup = pipapo_get(net, set, end, nft_genmask_next(net)); + dup = pipapo_get(net, set, end, nft_genmask_next(net), tstamp); }
if (PTR_ERR(dup) != -ENOENT) { @@ -1580,6 +1583,7 @@ static void pipapo_gc(const struct nft_s struct nft_set *set = (struct nft_set *) _set; struct nft_pipapo *priv = nft_set_priv(set); struct net *net = read_pnet(&set->net); + u64 tstamp = nft_net_tstamp(net); int rules_f0, first_rule = 0; struct nft_trans_gc *gc;
@@ -1613,7 +1617,7 @@ static void pipapo_gc(const struct nft_s /* synchronous gc never fails, there is no need to set on * NFT_SET_ELEM_DEAD_BIT. */ - if (nft_set_elem_expired(&e->ext)) { + if (__nft_set_elem_expired(&e->ext, tstamp)) { priv->dirty = true;
gc = nft_trans_gc_queue_sync(gc, GFP_ATOMIC); @@ -1772,7 +1776,7 @@ static void *pipapo_deactivate(const str { struct nft_pipapo_elem *e;
- e = pipapo_get(net, set, data, nft_genmask_next(net)); + e = pipapo_get(net, set, data, nft_genmask_next(net), nft_net_tstamp(net)); if (IS_ERR(e)) return NULL;
--- a/net/netfilter/nft_set_rbtree.c +++ b/net/netfilter/nft_set_rbtree.c @@ -316,6 +316,7 @@ static int __nft_rbtree_insert(const str struct nft_rbtree *priv = nft_set_priv(set); u8 cur_genmask = nft_genmask_cur(net); u8 genmask = nft_genmask_next(net); + u64 tstamp = nft_net_tstamp(net); int d;
/* Descend the tree to search for an existing element greater than the @@ -363,7 +364,7 @@ static int __nft_rbtree_insert(const str /* perform garbage collection to avoid bogus overlap reports * but skip new elements in this transaction. */ - if (nft_set_elem_expired(&rbe->ext) && + if (__nft_set_elem_expired(&rbe->ext, tstamp) && nft_set_elem_active(&rbe->ext, cur_genmask)) { const struct nft_rbtree_elem *removed_end;
@@ -550,6 +551,7 @@ static void *nft_rbtree_deactivate(const const struct rb_node *parent = priv->root.rb_node; struct nft_rbtree_elem *rbe, *this = elem->priv; u8 genmask = nft_genmask_next(net); + u64 tstamp = nft_net_tstamp(net); int d;
while (parent != NULL) { @@ -570,7 +572,7 @@ static void *nft_rbtree_deactivate(const nft_rbtree_interval_end(this)) { parent = parent->rb_right; continue; - } else if (nft_set_elem_expired(&rbe->ext)) { + } else if (__nft_set_elem_expired(&rbe->ext, tstamp)) { break; } else if (!nft_set_elem_active(&rbe->ext, genmask)) { parent = parent->rb_left;