From d40b0116c94bd8fc2b63aae35ce8e66bb53bba42 Mon Sep 17 00:00:00 2001 From: Daniel Borkmann Date: Thu, 16 Aug 2018 21:49:08 +0200 Subject: bpf, sockmap: fix leakage of smap_psock_map_entry While working on sockmap I noticed that we do not always kfree the struct smap_psock_map_entry list elements which track psocks attached to maps. In the case of sock_hash_ctx_update_elem(), these map entries are allocated outside of __sock_map_ctx_update_elem() with their linkage to the socket hash table filled. In the case of sock array, the map entries are allocated inside of __sock_map_ctx_update_elem() and added with their linkage to the psock->maps. Both additions are under psock->maps_lock each. Now, we drop these elements from their psock->maps list in a few occasions: i) in sock array via smap_list_map_remove() when an entry is either deleted from the map from user space, or updated via user space or BPF program where we drop the old socket at that map slot, or the sock array is freed via sock_map_free() and drops all its elements; ii) for sock hash via smap_list_hash_remove() in exactly the same occasions as just described for sock array; iii) in the bpf_tcp_close() where we remove the elements from the list via psock_map_pop() and iterate over them dropping themselves from either sock array or sock hash; and last but not least iv) once again in smap_gc_work() which is a callback for deferring the work once the psock refcount hit zero and thus the socket is being destroyed. Problem is that the only case where we kfree() the list entry is in case iv), which at that point should have an empty list in normal cases. So in cases from i) to iii) we unlink the elements without freeing where they go out of reach from us. Hence fix is to properly kfree() them as well to stop the leakage. Given these are all handled under psock->maps_lock there is no need for deferred RCU freeing. I later also ran with kmemleak detector and it confirmed the finding as well where in the state before the fix the object goes unreferenced while after the patch no kmemleak report related to BPF showed up. [...] unreferenced object 0xffff880378eadae0 (size 64): comm "test_sockmap", pid 2225, jiffies 4294720701 (age 43.504s) hex dump (first 32 bytes): 00 01 00 00 00 00 ad de 00 02 00 00 00 00 ad de ................ 50 4d 75 5d 03 88 ff ff 00 00 00 00 00 00 00 00 PMu]............ backtrace: [<000000005225ac3c>] sock_map_ctx_update_elem.isra.21+0xd8/0x210 [<0000000045dd6d3c>] bpf_sock_map_update+0x29/0x60 [<00000000877723aa>] ___bpf_prog_run+0x1e1f/0x4960 [<000000002ef89e83>] 0xffffffffffffffff unreferenced object 0xffff880378ead240 (size 64): comm "test_sockmap", pid 2225, jiffies 4294720701 (age 43.504s) hex dump (first 32 bytes): 00 01 00 00 00 00 ad de 00 02 00 00 00 00 ad de ................ 00 44 75 5d 03 88 ff ff 00 00 00 00 00 00 00 00 .Du]............ backtrace: [<000000005225ac3c>] sock_map_ctx_update_elem.isra.21+0xd8/0x210 [<0000000030e37a3a>] sock_map_update_elem+0x125/0x240 [<000000002e5ce36e>] map_update_elem+0x4eb/0x7b0 [<00000000db453cc9>] __x64_sys_bpf+0x1f9/0x360 [<0000000000763660>] do_syscall_64+0x9a/0x300 [<00000000422a2bb2>] entry_SYSCALL_64_after_hwframe+0x44/0xa9 [<000000002ef89e83>] 0xffffffffffffffff [...] Fixes: e9db4ef6bf4c ("bpf: sockhash fix omitted bucket lock in sock_close") Fixes: 54fedb42c653 ("bpf: sockmap, fix smap_list_map_remove when psock is in many maps") Fixes: 2f857d04601a ("bpf: sockmap, remove STRPARSER map_flags and add multi-map support") Signed-off-by: Daniel Borkmann Acked-by: John Fastabend Acked-by: Song Liu Signed-off-by: Alexei Starovoitov --- kernel/bpf/sockmap.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/bpf/sockmap.c b/kernel/bpf/sockmap.c index 0c1a696b041b..94a324bc6afc 100644 --- a/kernel/bpf/sockmap.c +++ b/kernel/bpf/sockmap.c @@ -370,6 +370,7 @@ static void bpf_tcp_close(struct sock *sk, long timeout) } raw_spin_unlock_bh(&b->lock); } + kfree(e); e = psock_map_pop(sk, psock); } rcu_read_unlock(); @@ -1675,8 +1676,10 @@ static void smap_list_map_remove(struct smap_psock *psock, spin_lock_bh(&psock->maps_lock); list_for_each_entry_safe(e, tmp, &psock->maps, list) { - if (e->entry == entry) + if (e->entry == entry) { list_del(&e->list); + kfree(e); + } } spin_unlock_bh(&psock->maps_lock); } @@ -1690,8 +1693,10 @@ static void smap_list_hash_remove(struct smap_psock *psock, list_for_each_entry_safe(e, tmp, &psock->maps, list) { struct htab_elem *c = rcu_dereference(e->hash_link); - if (c == hash_link) + if (c == hash_link) { list_del(&e->list); + kfree(e); + } } spin_unlock_bh(&psock->maps_lock); } -- cgit v1.2.3 From 166ab6f0a0702fdd4d865ad5090bf3094ed83428 Mon Sep 17 00:00:00 2001 From: Daniel Borkmann Date: Thu, 16 Aug 2018 21:49:09 +0200 Subject: bpf, sockmap: fix map elem deletion race with smap_stop_sock The smap_start_sock() and smap_stop_sock() are each protected under the sock->sk_callback_lock from their call-sites except in the case of sock_map_delete_elem() where we drop the old socket from the map slot. This is racy because the same sock could be part of multiple sock maps, so we run smap_stop_sock() in parallel, and given at that point psock->strp_enabled might be true on both CPUs, we might for example wrongly restore the sk->sk_data_ready / sk->sk_write_space. Therefore, hold the sock->sk_callback_lock as well on delete. Looks like 2f857d04601a ("bpf: sockmap, remove STRPARSER map_flags and add multi-map support") had this right, but later on e9db4ef6bf4c ("bpf: sockhash fix omitted bucket lock in sock_close") removed it again from delete leaving this smap_stop_sock() instance unprotected. Fixes: e9db4ef6bf4c ("bpf: sockhash fix omitted bucket lock in sock_close") Signed-off-by: Daniel Borkmann Acked-by: John Fastabend Acked-by: Song Liu Signed-off-by: Alexei Starovoitov --- kernel/bpf/sockmap.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/bpf/sockmap.c b/kernel/bpf/sockmap.c index 94a324bc6afc..921cb6b8c862 100644 --- a/kernel/bpf/sockmap.c +++ b/kernel/bpf/sockmap.c @@ -1786,8 +1786,11 @@ static int sock_map_delete_elem(struct bpf_map *map, void *key) if (!psock) goto out; - if (psock->bpf_parse) + if (psock->bpf_parse) { + write_lock_bh(&sock->sk_callback_lock); smap_stop_sock(psock, sock); + write_unlock_bh(&sock->sk_callback_lock); + } smap_list_map_remove(psock, &stab->sock_map[k]); smap_release_sock(psock, sock); out: -- cgit v1.2.3 From 585f5a6252ee43ec8feeee07387e3fcc7e8bb292 Mon Sep 17 00:00:00 2001 From: Daniel Borkmann Date: Thu, 16 Aug 2018 21:49:10 +0200 Subject: bpf, sockmap: fix sock_map_ctx_update_elem race with exist/noexist The current code in sock_map_ctx_update_elem() allows for BPF_EXIST and BPF_NOEXIST map update flags. While on array-like maps this approach is rather uncommon, e.g. bpf_fd_array_map_update_elem() and others enforce map update flags to be BPF_ANY such that xchg() can be used directly, the current implementation in sock map does not guarantee that such operation with BPF_EXIST / BPF_NOEXIST is atomic. The initial test does a READ_ONCE(stab->sock_map[i]) to fetch the socket from the slot which is then tested for NULL / non-NULL. However later after __sock_map_ctx_update_elem(), the actual update is done through osock = xchg(&stab->sock_map[i], sock). Problem is that in the meantime a different CPU could have updated / deleted a socket on that specific slot and thus flag contraints won't hold anymore. I've been thinking whether best would be to just break UAPI and do an enforcement of BPF_ANY to check if someone actually complains, however trouble is that already in BPF kselftest we use BPF_NOEXIST for the map update, and therefore it might have been copied into applications already. The fix to keep the current behavior intact would be to add a map lock similar to the sock hash bucket lock only for covering the whole map. Fixes: 174a79ff9515 ("bpf: sockmap with sk redirect support") Signed-off-by: Daniel Borkmann Acked-by: John Fastabend Acked-by: Song Liu Signed-off-by: Alexei Starovoitov --- kernel/bpf/sockmap.c | 106 +++++++++++++++++++++++++++------------------------ 1 file changed, 57 insertions(+), 49 deletions(-) (limited to 'kernel') diff --git a/kernel/bpf/sockmap.c b/kernel/bpf/sockmap.c index 921cb6b8c862..98e621a29e8e 100644 --- a/kernel/bpf/sockmap.c +++ b/kernel/bpf/sockmap.c @@ -58,6 +58,7 @@ struct bpf_stab { struct bpf_map map; struct sock **sock_map; struct bpf_sock_progs progs; + raw_spinlock_t lock; }; struct bucket { @@ -89,9 +90,9 @@ enum smap_psock_state { struct smap_psock_map_entry { struct list_head list; + struct bpf_map *map; struct sock **entry; struct htab_elem __rcu *hash_link; - struct bpf_htab __rcu *htab; }; struct smap_psock { @@ -343,13 +344,18 @@ static void bpf_tcp_close(struct sock *sk, long timeout) e = psock_map_pop(sk, psock); while (e) { if (e->entry) { - osk = cmpxchg(e->entry, sk, NULL); + struct bpf_stab *stab = container_of(e->map, struct bpf_stab, map); + + raw_spin_lock_bh(&stab->lock); + osk = *e->entry; if (osk == sk) { + *e->entry = NULL; smap_release_sock(psock, sk); } + raw_spin_unlock_bh(&stab->lock); } else { struct htab_elem *link = rcu_dereference(e->hash_link); - struct bpf_htab *htab = rcu_dereference(e->htab); + struct bpf_htab *htab = container_of(e->map, struct bpf_htab, map); struct hlist_head *head; struct htab_elem *l; struct bucket *b; @@ -1642,6 +1648,7 @@ static struct bpf_map *sock_map_alloc(union bpf_attr *attr) return ERR_PTR(-ENOMEM); bpf_map_init_from_attr(&stab->map, attr); + raw_spin_lock_init(&stab->lock); /* make sure page count doesn't overflow */ cost = (u64) stab->map.max_entries * sizeof(struct sock *); @@ -1716,14 +1723,15 @@ static void sock_map_free(struct bpf_map *map) * and a grace period expire to ensure psock is really safe to remove. */ rcu_read_lock(); + raw_spin_lock_bh(&stab->lock); for (i = 0; i < stab->map.max_entries; i++) { struct smap_psock *psock; struct sock *sock; - sock = xchg(&stab->sock_map[i], NULL); + sock = stab->sock_map[i]; if (!sock) continue; - + stab->sock_map[i] = NULL; psock = smap_psock_sk(sock); /* This check handles a racing sock event that can get the * sk_callback_lock before this case but after xchg happens @@ -1735,6 +1743,7 @@ static void sock_map_free(struct bpf_map *map) smap_release_sock(psock, sock); } } + raw_spin_unlock_bh(&stab->lock); rcu_read_unlock(); sock_map_remove_complete(stab); @@ -1778,14 +1787,16 @@ static int sock_map_delete_elem(struct bpf_map *map, void *key) if (k >= map->max_entries) return -EINVAL; - sock = xchg(&stab->sock_map[k], NULL); + raw_spin_lock_bh(&stab->lock); + sock = stab->sock_map[k]; + stab->sock_map[k] = NULL; + raw_spin_unlock_bh(&stab->lock); if (!sock) return -EINVAL; psock = smap_psock_sk(sock); if (!psock) - goto out; - + return 0; if (psock->bpf_parse) { write_lock_bh(&sock->sk_callback_lock); smap_stop_sock(psock, sock); @@ -1793,7 +1804,6 @@ static int sock_map_delete_elem(struct bpf_map *map, void *key) } smap_list_map_remove(psock, &stab->sock_map[k]); smap_release_sock(psock, sock); -out: return 0; } @@ -1829,11 +1839,9 @@ out: static int __sock_map_ctx_update_elem(struct bpf_map *map, struct bpf_sock_progs *progs, struct sock *sock, - struct sock **map_link, void *key) { struct bpf_prog *verdict, *parse, *tx_msg; - struct smap_psock_map_entry *e = NULL; struct smap_psock *psock; bool new = false; int err = 0; @@ -1906,14 +1914,6 @@ static int __sock_map_ctx_update_elem(struct bpf_map *map, new = true; } - if (map_link) { - e = kzalloc(sizeof(*e), GFP_ATOMIC | __GFP_NOWARN); - if (!e) { - err = -ENOMEM; - goto out_free; - } - } - /* 3. At this point we have a reference to a valid psock that is * running. Attach any BPF programs needed. */ @@ -1935,17 +1935,6 @@ static int __sock_map_ctx_update_elem(struct bpf_map *map, write_unlock_bh(&sock->sk_callback_lock); } - /* 4. Place psock in sockmap for use and stop any programs on - * the old sock assuming its not the same sock we are replacing - * it with. Because we can only have a single set of programs if - * old_sock has a strp we can stop it. - */ - if (map_link) { - e->entry = map_link; - spin_lock_bh(&psock->maps_lock); - list_add_tail(&e->list, &psock->maps); - spin_unlock_bh(&psock->maps_lock); - } return err; out_free: smap_release_sock(psock, sock); @@ -1956,7 +1945,6 @@ out_progs: } if (tx_msg) bpf_prog_put(tx_msg); - kfree(e); return err; } @@ -1966,36 +1954,57 @@ static int sock_map_ctx_update_elem(struct bpf_sock_ops_kern *skops, { struct bpf_stab *stab = container_of(map, struct bpf_stab, map); struct bpf_sock_progs *progs = &stab->progs; - struct sock *osock, *sock; + struct sock *osock, *sock = skops->sk; + struct smap_psock_map_entry *e; + struct smap_psock *psock; u32 i = *(u32 *)key; int err; if (unlikely(flags > BPF_EXIST)) return -EINVAL; - if (unlikely(i >= stab->map.max_entries)) return -E2BIG; - sock = READ_ONCE(stab->sock_map[i]); - if (flags == BPF_EXIST && !sock) - return -ENOENT; - else if (flags == BPF_NOEXIST && sock) - return -EEXIST; + e = kzalloc(sizeof(*e), GFP_ATOMIC | __GFP_NOWARN); + if (!e) + return -ENOMEM; - sock = skops->sk; - err = __sock_map_ctx_update_elem(map, progs, sock, &stab->sock_map[i], - key); + err = __sock_map_ctx_update_elem(map, progs, sock, key); if (err) goto out; - osock = xchg(&stab->sock_map[i], sock); - if (osock) { - struct smap_psock *opsock = smap_psock_sk(osock); + /* psock guaranteed to be present. */ + psock = smap_psock_sk(sock); + raw_spin_lock_bh(&stab->lock); + osock = stab->sock_map[i]; + if (osock && flags == BPF_NOEXIST) { + err = -EEXIST; + goto out_unlock; + } + if (!osock && flags == BPF_EXIST) { + err = -ENOENT; + goto out_unlock; + } + + e->entry = &stab->sock_map[i]; + e->map = map; + spin_lock_bh(&psock->maps_lock); + list_add_tail(&e->list, &psock->maps); + spin_unlock_bh(&psock->maps_lock); - smap_list_map_remove(opsock, &stab->sock_map[i]); - smap_release_sock(opsock, osock); + stab->sock_map[i] = sock; + if (osock) { + psock = smap_psock_sk(osock); + smap_list_map_remove(psock, &stab->sock_map[i]); + smap_release_sock(psock, osock); } + raw_spin_unlock_bh(&stab->lock); + return 0; +out_unlock: + smap_release_sock(psock, sock); + raw_spin_unlock_bh(&stab->lock); out: + kfree(e); return err; } @@ -2358,7 +2367,7 @@ static int sock_hash_ctx_update_elem(struct bpf_sock_ops_kern *skops, b = __select_bucket(htab, hash); head = &b->head; - err = __sock_map_ctx_update_elem(map, progs, sock, NULL, key); + err = __sock_map_ctx_update_elem(map, progs, sock, key); if (err) goto err; @@ -2384,8 +2393,7 @@ static int sock_hash_ctx_update_elem(struct bpf_sock_ops_kern *skops, } rcu_assign_pointer(e->hash_link, l_new); - rcu_assign_pointer(e->htab, - container_of(map, struct bpf_htab, map)); + e->map = map; spin_lock_bh(&psock->maps_lock); list_add_tail(&e->list, &psock->maps); spin_unlock_bh(&psock->maps_lock); -- cgit v1.2.3