From: Florian Westphal <fw@xxxxxxxxx> nf_tables_chain_destroy can sleep, it can't be used from call_rcu callbacks. Moreover, nf_tables_rule_release() is only safe for error unwinding, while transaction mutex is held and the to-be-desroyed rule was not exposed to either dataplane or dumps, as it deactives+frees without the required synchronize_rcu() in-between. nft_rule_expr_deactivate() callbacks will change ->use counters of other chains/sets, see e.g. nft_lookup .deactivate callback, these must be serialized via transaction mutex. Also add a few lockdep asserts to make this more explicit. Calling synchronize_rcu() isn't ideal, but fixing this without is hard and way more intrusive. As-is, we can get: WARNING: .. net/netfilter/nf_tables_api.c:5515 nft_set_destroy+0x.. Workqueue: events nf_tables_trans_destroy_work RIP: 0010:nft_set_destroy+0x3fe/0x5c0 Call Trace: <TASK> nf_tables_trans_destroy_work+0x6b7/0xad0 process_one_work+0x64a/0xce0 worker_thread+0x613/0x10d0 In case the synchronize_rcu becomes an issue, we can explore alternatives. One way would be to allocate nft_trans_rule objects + one nft_trans_chain object, deactivate the rules + the chain and then defer the freeing to the nft destroy workqueue. We'd still need to keep the synchronize_rcu path as a fallback to handle -ENOMEM corner cases though. Reported-by: syzbot+b26935466701e56cfdc2@xxxxxxxxxxxxxxxxxxxxxxxxx Closes: https://lore.kernel.org/all/67478d92.050a0220.253251.0062.GAE@xxxxxxxxxx/T/ Fixes: c03d278fdf35 ("netfilter: nf_tables: wait for rcu grace period on net_device removal") Signed-off-by: Florian Westphal <fw@xxxxxxxxx> Signed-off-by: Pablo Neira Ayuso <pablo@xxxxxxxxxxxxx> --- include/net/netfilter/nf_tables.h | 4 ---- net/netfilter/nf_tables_api.c | 32 +++++++++++++++---------------- 2 files changed, 15 insertions(+), 21 deletions(-) diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h index 80a537ac26cd..4afa64c81304 100644 --- a/include/net/netfilter/nf_tables.h +++ b/include/net/netfilter/nf_tables.h @@ -1103,7 +1103,6 @@ struct nft_rule_blob { * @name: name of the chain * @udlen: user data length * @udata: user data in the chain - * @rcu_head: rcu head for deferred release * @blob_next: rule blob pointer to the next in the chain */ struct nft_chain { @@ -1121,7 +1120,6 @@ struct nft_chain { char *name; u16 udlen; u8 *udata; - struct rcu_head rcu_head; /* Only used during control plane commit phase: */ struct nft_rule_blob *blob_next; @@ -1265,7 +1263,6 @@ static inline void nft_use_inc_restore(u32 *use) * @sets: sets in the table * @objects: stateful objects in the table * @flowtables: flow tables in the table - * @net: netnamespace this table belongs to * @hgenerator: handle generator state * @handle: table handle * @use: number of chain references to this table @@ -1285,7 +1282,6 @@ struct nft_table { struct list_head sets; struct list_head objects; struct list_head flowtables; - possible_net_t net; u64 hgenerator; u64 handle; u32 use; diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index 21b6f7410a1f..0b9f1e8dfe49 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -1596,7 +1596,6 @@ static int nf_tables_newtable(struct sk_buff *skb, const struct nfnl_info *info, INIT_LIST_HEAD(&table->sets); INIT_LIST_HEAD(&table->objects); INIT_LIST_HEAD(&table->flowtables); - write_pnet(&table->net, net); table->family = family; table->flags = flags; table->handle = ++nft_net->table_handle; @@ -3987,8 +3986,11 @@ void nf_tables_rule_destroy(const struct nft_ctx *ctx, struct nft_rule *rule) kfree(rule); } +/* can only be used if rule is no longer visible to dumps */ static void nf_tables_rule_release(const struct nft_ctx *ctx, struct nft_rule *rule) { + lockdep_commit_lock_is_held(ctx->net); + nft_rule_expr_deactivate(ctx, rule, NFT_TRANS_RELEASE); nf_tables_rule_destroy(ctx, rule); } @@ -5757,6 +5759,8 @@ void nf_tables_deactivate_set(const struct nft_ctx *ctx, struct nft_set *set, struct nft_set_binding *binding, enum nft_trans_phase phase) { + lockdep_commit_lock_is_held(ctx->net); + switch (phase) { case NFT_TRANS_PREPARE_ERROR: nft_set_trans_unbind(ctx, set); @@ -11695,19 +11699,6 @@ static void __nft_release_basechain_now(struct nft_ctx *ctx) nf_tables_chain_destroy(ctx->chain); } -static void nft_release_basechain_rcu(struct rcu_head *head) -{ - struct nft_chain *chain = container_of(head, struct nft_chain, rcu_head); - struct nft_ctx ctx = { - .family = chain->table->family, - .chain = chain, - .net = read_pnet(&chain->table->net), - }; - - __nft_release_basechain_now(&ctx); - put_net(ctx.net); -} - int __nft_release_basechain(struct nft_ctx *ctx) { struct nft_rule *rule; @@ -11722,11 +11713,18 @@ int __nft_release_basechain(struct nft_ctx *ctx) nft_chain_del(ctx->chain); nft_use_dec(&ctx->table->use); - if (maybe_get_net(ctx->net)) - call_rcu(&ctx->chain->rcu_head, nft_release_basechain_rcu); - else + if (!maybe_get_net(ctx->net)) { __nft_release_basechain_now(ctx); + return 0; + } + + /* wait for ruleset dumps to complete. Owning chain is no longer in + * lists, so new dumps can't find any of these rules anymore. + */ + synchronize_rcu(); + __nft_release_basechain_now(ctx); + put_net(ctx->net); return 0; } EXPORT_SYMBOL_GPL(__nft_release_basechain); -- 2.30.2
