diff options
author | David S. Miller <davem@davemloft.net> | 2018-05-16 22:47:11 -0400 |
---|---|---|
committer | David S. Miller <davem@davemloft.net> | 2018-05-16 22:47:11 -0400 |
commit | b9f672af148bf7a08a6031743156faffd58dbc7e (patch) | |
tree | 4e3a384636147f0fd31ec01cc267a51bdab7cbb5 /kernel | |
parent | 8e725f7caafb8e820e05707fe9853023375438cf (diff) | |
parent | e23afe5e7cba89cd0744c5218eda1b3553455c17 (diff) | |
download | linux-b9f672af148bf7a08a6031743156faffd58dbc7e.tar.bz2 |
Merge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next
Daniel Borkmann says:
====================
pull-request: bpf-next 2018-05-17
The following pull-request contains BPF updates for your *net-next* tree.
The main changes are:
1) Provide a new BPF helper for doing a FIB and neighbor lookup
in the kernel tables from an XDP or tc BPF program. The helper
provides a fast-path for forwarding packets. The API supports
IPv4, IPv6 and MPLS protocols, but currently IPv4 and IPv6 are
implemented in this initial work, from David (Ahern).
2) Just a tiny diff but huge feature enabled for nfp driver by
extending the BPF offload beyond a pure host processing offload.
Offloaded XDP programs are allowed to set the RX queue index and
thus opening the door for defining a fully programmable RSS/n-tuple
filter replacement. Once BPF decided on a queue already, the device
data-path will skip the conventional RSS processing completely,
from Jakub.
3) The original sockmap implementation was array based similar to
devmap. However unlike devmap where an ifindex has a 1:1 mapping
into the map there are use cases with sockets that need to be
referenced using longer keys. Hence, sockhash map is added reusing
as much of the sockmap code as possible, from John.
4) Introduce BTF ID. The ID is allocatd through an IDR similar as
with BPF maps and progs. It also makes BTF accessible to user
space via BPF_BTF_GET_FD_BY_ID and adds exposure of the BTF data
through BPF_OBJ_GET_INFO_BY_FD, from Martin.
5) Enable BPF stackmap with build_id also in NMI context. Due to the
up_read() of current->mm->mmap_sem build_id cannot be parsed.
This work defers the up_read() via a per-cpu irq_work so that
at least limited support can be enabled, from Song.
6) Various BPF JIT follow-up cleanups and fixups after the LD_ABS/LD_IND
JIT conversion as well as implementation of an optimized 32/64 bit
immediate load in the arm64 JIT that allows to reduce the number of
emitted instructions; in case of tested real-world programs they
were shrinking by three percent, from Daniel.
7) Add ifindex parameter to the libbpf loader in order to enable
BPF offload support. Right now only iproute2 can load offloaded
BPF and this will also enable libbpf for direct integration into
other applications, from David (Beckett).
8) Convert the plain text documentation under Documentation/bpf/ into
RST format since this is the appropriate standard the kernel is
moving to for all documentation. Also add an overview README.rst,
from Jesper.
9) Add __printf verification attribute to the bpf_verifier_vlog()
helper. Though it uses va_list we can still allow gcc to check
the format string, from Mathieu.
10) Fix a bash reference in the BPF selftest's Makefile. The '|& ...'
is a bash 4.0+ feature which is not guaranteed to be available
when calling out to shell, therefore use a more portable variant,
from Joe.
11) Fix a 64 bit division in xdp_umem_reg() by using div_u64()
instead of relying on the gcc built-in, from Björn.
12) Fix a sock hashmap kmalloc warning reported by syzbot when an
overly large key size is used in hashmap then causing overflows
in htab->elem_size. Reject bogus attr->key_size early in the
sock_hash_alloc(), from Yonghong.
13) Ensure in BPF selftests when urandom_read is being linked that
--build-id is always enabled so that test_stacktrace_build_id[_nmi]
won't be failing, from Alexei.
14) Add bitsperlong.h as well as errno.h uapi headers into the tools
header infrastructure which point to one of the arch specific
uapi headers. This was needed in order to fix a build error on
some systems for the BPF selftests, from Sirio.
15) Allow for short options to be used in the xdp_monitor BPF sample
code. And also a bpf.h tools uapi header sync in order to fix a
selftest build failure. Both from Prashant.
16) More formally clarify the meaning of ID in the direct packet access
section of the BPF documentation, from Wang.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'kernel')
-rw-r--r-- | kernel/bpf/btf.c | 136 | ||||
-rw-r--r-- | kernel/bpf/core.c | 1 | ||||
-rw-r--r-- | kernel/bpf/sockmap.c | 644 | ||||
-rw-r--r-- | kernel/bpf/stackmap.c | 59 | ||||
-rw-r--r-- | kernel/bpf/syscall.c | 41 | ||||
-rw-r--r-- | kernel/bpf/verifier.c | 16 |
6 files changed, 797 insertions, 100 deletions
diff --git a/kernel/bpf/btf.c b/kernel/bpf/btf.c index 22e1046a1a86..ded10ab47b8a 100644 --- a/kernel/bpf/btf.c +++ b/kernel/bpf/btf.c @@ -11,6 +11,7 @@ #include <linux/file.h> #include <linux/uaccess.h> #include <linux/kernel.h> +#include <linux/idr.h> #include <linux/bpf_verifier.h> #include <linux/btf.h> @@ -179,6 +180,9 @@ i < btf_type_vlen(struct_type); \ i++, member++) +static DEFINE_IDR(btf_idr); +static DEFINE_SPINLOCK(btf_idr_lock); + struct btf { union { struct btf_header *hdr; @@ -193,6 +197,8 @@ struct btf { u32 types_size; u32 data_size; refcount_t refcnt; + u32 id; + struct rcu_head rcu; }; enum verifier_phase { @@ -598,6 +604,42 @@ static int btf_add_type(struct btf_verifier_env *env, struct btf_type *t) return 0; } +static int btf_alloc_id(struct btf *btf) +{ + int id; + + idr_preload(GFP_KERNEL); + spin_lock_bh(&btf_idr_lock); + id = idr_alloc_cyclic(&btf_idr, btf, 1, INT_MAX, GFP_ATOMIC); + if (id > 0) + btf->id = id; + spin_unlock_bh(&btf_idr_lock); + idr_preload_end(); + + if (WARN_ON_ONCE(!id)) + return -ENOSPC; + + return id > 0 ? 0 : id; +} + +static void btf_free_id(struct btf *btf) +{ + unsigned long flags; + + /* + * In map-in-map, calling map_delete_elem() on outer + * map will call bpf_map_put on the inner map. + * It will then eventually call btf_free_id() + * on the inner map. Some of the map_delete_elem() + * implementation may have irq disabled, so + * we need to use the _irqsave() version instead + * of the _bh() version. + */ + spin_lock_irqsave(&btf_idr_lock, flags); + idr_remove(&btf_idr, btf->id); + spin_unlock_irqrestore(&btf_idr_lock, flags); +} + static void btf_free(struct btf *btf) { kvfree(btf->types); @@ -607,15 +649,19 @@ static void btf_free(struct btf *btf) kfree(btf); } -static void btf_get(struct btf *btf) +static void btf_free_rcu(struct rcu_head *rcu) { - refcount_inc(&btf->refcnt); + struct btf *btf = container_of(rcu, struct btf, rcu); + + btf_free(btf); } void btf_put(struct btf *btf) { - if (btf && refcount_dec_and_test(&btf->refcnt)) - btf_free(btf); + if (btf && refcount_dec_and_test(&btf->refcnt)) { + btf_free_id(btf); + call_rcu(&btf->rcu, btf_free_rcu); + } } static int env_resolve_init(struct btf_verifier_env *env) @@ -1977,7 +2023,7 @@ static struct btf *btf_parse(void __user *btf_data, u32 btf_data_size, if (!err) { btf_verifier_env_free(env); - btf_get(btf); + refcount_set(&btf->refcnt, 1); return btf; } @@ -2006,10 +2052,15 @@ const struct file_operations btf_fops = { .release = btf_release, }; +static int __btf_new_fd(struct btf *btf) +{ + return anon_inode_getfd("btf", &btf_fops, btf, O_RDONLY | O_CLOEXEC); +} + int btf_new_fd(const union bpf_attr *attr) { struct btf *btf; - int fd; + int ret; btf = btf_parse(u64_to_user_ptr(attr->btf), attr->btf_size, attr->btf_log_level, @@ -2018,12 +2069,23 @@ int btf_new_fd(const union bpf_attr *attr) if (IS_ERR(btf)) return PTR_ERR(btf); - fd = anon_inode_getfd("btf", &btf_fops, btf, - O_RDONLY | O_CLOEXEC); - if (fd < 0) + ret = btf_alloc_id(btf); + if (ret) { + btf_free(btf); + return ret; + } + + /* + * The BTF ID is published to the userspace. + * All BTF free must go through call_rcu() from + * now on (i.e. free by calling btf_put()). + */ + + ret = __btf_new_fd(btf); + if (ret < 0) btf_put(btf); - return fd; + return ret; } struct btf *btf_get_by_fd(int fd) @@ -2042,7 +2104,7 @@ struct btf *btf_get_by_fd(int fd) } btf = f.file->private_data; - btf_get(btf); + refcount_inc(&btf->refcnt); fdput(f); return btf; @@ -2052,13 +2114,55 @@ int btf_get_info_by_fd(const struct btf *btf, const union bpf_attr *attr, union bpf_attr __user *uattr) { - void __user *udata = u64_to_user_ptr(attr->info.info); - u32 copy_len = min_t(u32, btf->data_size, - attr->info.info_len); + struct bpf_btf_info __user *uinfo; + struct bpf_btf_info info = {}; + u32 info_copy, btf_copy; + void __user *ubtf; + u32 uinfo_len; + + uinfo = u64_to_user_ptr(attr->info.info); + uinfo_len = attr->info.info_len; - if (copy_to_user(udata, btf->data, copy_len) || - put_user(btf->data_size, &uattr->info.info_len)) + info_copy = min_t(u32, uinfo_len, sizeof(info)); + if (copy_from_user(&info, uinfo, info_copy)) + return -EFAULT; + + info.id = btf->id; + ubtf = u64_to_user_ptr(info.btf); + btf_copy = min_t(u32, btf->data_size, info.btf_size); + if (copy_to_user(ubtf, btf->data, btf_copy)) + return -EFAULT; + info.btf_size = btf->data_size; + + if (copy_to_user(uinfo, &info, info_copy) || + put_user(info_copy, &uattr->info.info_len)) return -EFAULT; return 0; } + +int btf_get_fd_by_id(u32 id) +{ + struct btf *btf; + int fd; + + rcu_read_lock(); + btf = idr_find(&btf_idr, id); + if (!btf || !refcount_inc_not_zero(&btf->refcnt)) + btf = ERR_PTR(-ENOENT); + rcu_read_unlock(); + + if (IS_ERR(btf)) + return PTR_ERR(btf); + + fd = __btf_new_fd(btf); + if (fd < 0) + btf_put(btf); + + return fd; +} + +u32 btf_id(const struct btf *btf) +{ + return btf->id; +} diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c index d0d7d9462368..2194c6a9df42 100644 --- a/kernel/bpf/core.c +++ b/kernel/bpf/core.c @@ -1707,6 +1707,7 @@ const struct bpf_func_proto bpf_get_current_pid_tgid_proto __weak; const struct bpf_func_proto bpf_get_current_uid_gid_proto __weak; const struct bpf_func_proto bpf_get_current_comm_proto __weak; const struct bpf_func_proto bpf_sock_map_update_proto __weak; +const struct bpf_func_proto bpf_sock_hash_update_proto __weak; const struct bpf_func_proto * __weak bpf_get_trace_printk_proto(void) { diff --git a/kernel/bpf/sockmap.c b/kernel/bpf/sockmap.c index 098eca568c2b..c6de1393df63 100644 --- a/kernel/bpf/sockmap.c +++ b/kernel/bpf/sockmap.c @@ -48,14 +48,40 @@ #define SOCK_CREATE_FLAG_MASK \ (BPF_F_NUMA_NODE | BPF_F_RDONLY | BPF_F_WRONLY) -struct bpf_stab { - struct bpf_map map; - struct sock **sock_map; +struct bpf_sock_progs { struct bpf_prog *bpf_tx_msg; struct bpf_prog *bpf_parse; struct bpf_prog *bpf_verdict; }; +struct bpf_stab { + struct bpf_map map; + struct sock **sock_map; + struct bpf_sock_progs progs; +}; + +struct bucket { + struct hlist_head head; + raw_spinlock_t lock; +}; + +struct bpf_htab { + struct bpf_map map; + struct bucket *buckets; + atomic_t count; + u32 n_buckets; + u32 elem_size; + struct bpf_sock_progs progs; +}; + +struct htab_elem { + struct rcu_head rcu; + struct hlist_node hash_node; + u32 hash; + struct sock *sk; + char key[0]; +}; + enum smap_psock_state { SMAP_TX_RUNNING, }; @@ -63,6 +89,8 @@ enum smap_psock_state { struct smap_psock_map_entry { struct list_head list; struct sock **entry; + struct htab_elem *hash_link; + struct bpf_htab *htab; }; struct smap_psock { @@ -191,6 +219,12 @@ out: rcu_read_unlock(); } +static void free_htab_elem(struct bpf_htab *htab, struct htab_elem *l) +{ + atomic_dec(&htab->count); + kfree_rcu(l, rcu); +} + static void bpf_tcp_close(struct sock *sk, long timeout) { void (*close_fun)(struct sock *sk, long timeout); @@ -227,10 +261,16 @@ static void bpf_tcp_close(struct sock *sk, long timeout) } list_for_each_entry_safe(e, tmp, &psock->maps, list) { - osk = cmpxchg(e->entry, sk, NULL); - if (osk == sk) { - list_del(&e->list); - smap_release_sock(psock, sk); + if (e->entry) { + osk = cmpxchg(e->entry, sk, NULL); + if (osk == sk) { + list_del(&e->list); + smap_release_sock(psock, sk); + } + } else { + hlist_del_rcu(&e->hash_link->hash_node); + smap_release_sock(psock, e->hash_link->sk); + free_htab_elem(e->htab, e->hash_link); } } write_unlock_bh(&sk->sk_callback_lock); @@ -461,7 +501,7 @@ static int free_curr_sg(struct sock *sk, struct sk_msg_buff *md) static int bpf_map_msg_verdict(int _rc, struct sk_msg_buff *md) { return ((_rc == SK_PASS) ? - (md->map ? __SK_REDIRECT : __SK_PASS) : + (md->sk_redir ? __SK_REDIRECT : __SK_PASS) : __SK_DROP); } @@ -1092,7 +1132,7 @@ static int smap_verdict_func(struct smap_psock *psock, struct sk_buff *skb) * when we orphan the skb so that we don't have the possibility * to reference a stale map. */ - TCP_SKB_CB(skb)->bpf.map = NULL; + TCP_SKB_CB(skb)->bpf.sk_redir = NULL; skb->sk = psock->sock; bpf_compute_data_pointers(skb); preempt_disable(); @@ -1102,7 +1142,7 @@ static int smap_verdict_func(struct smap_psock *psock, struct sk_buff *skb) /* Moving return codes from UAPI namespace into internal namespace */ return rc == SK_PASS ? - (TCP_SKB_CB(skb)->bpf.map ? __SK_REDIRECT : __SK_PASS) : + (TCP_SKB_CB(skb)->bpf.sk_redir ? __SK_REDIRECT : __SK_PASS) : __SK_DROP; } @@ -1372,7 +1412,6 @@ static int smap_init_sock(struct smap_psock *psock, } static void smap_init_progs(struct smap_psock *psock, - struct bpf_stab *stab, struct bpf_prog *verdict, struct bpf_prog *parse) { @@ -1450,14 +1489,13 @@ static void smap_gc_work(struct work_struct *w) kfree(psock); } -static struct smap_psock *smap_init_psock(struct sock *sock, - struct bpf_stab *stab) +static struct smap_psock *smap_init_psock(struct sock *sock, int node) { struct smap_psock *psock; psock = kzalloc_node(sizeof(struct smap_psock), GFP_ATOMIC | __GFP_NOWARN, - stab->map.numa_node); + node); if (!psock) return ERR_PTR(-ENOMEM); @@ -1525,12 +1563,14 @@ free_stab: return ERR_PTR(err); } -static void smap_list_remove(struct smap_psock *psock, struct sock **entry) +static void smap_list_remove(struct smap_psock *psock, + struct sock **entry, + struct htab_elem *hash_link) { struct smap_psock_map_entry *e, *tmp; list_for_each_entry_safe(e, tmp, &psock->maps, list) { - if (e->entry == entry) { + if (e->entry == entry || e->hash_link == hash_link) { list_del(&e->list); break; } @@ -1568,7 +1608,7 @@ static void sock_map_free(struct bpf_map *map) * to be null and queued for garbage collection. */ if (likely(psock)) { - smap_list_remove(psock, &stab->sock_map[i]); + smap_list_remove(psock, &stab->sock_map[i], NULL); smap_release_sock(psock, sock); } write_unlock_bh(&sock->sk_callback_lock); @@ -1627,7 +1667,7 @@ static int sock_map_delete_elem(struct bpf_map *map, void *key) if (psock->bpf_parse) smap_stop_sock(psock, sock); - smap_list_remove(psock, &stab->sock_map[k]); + smap_list_remove(psock, &stab->sock_map[k], NULL); smap_release_sock(psock, sock); out: write_unlock_bh(&sock->sk_callback_lock); @@ -1662,40 +1702,26 @@ out: * - sock_map must use READ_ONCE and (cmp)xchg operations * - BPF verdict/parse programs must use READ_ONCE and xchg operations */ -static int sock_map_ctx_update_elem(struct bpf_sock_ops_kern *skops, - struct bpf_map *map, - void *key, u64 flags) + +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_stab *stab = container_of(map, struct bpf_stab, map); - struct smap_psock_map_entry *e = NULL; struct bpf_prog *verdict, *parse, *tx_msg; - struct sock *osock, *sock; + struct smap_psock_map_entry *e = NULL; struct smap_psock *psock; - u32 i = *(u32 *)key; bool new = false; 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; - - sock = skops->sk; - /* 1. If sock map has BPF programs those will be inherited by the * sock being added. If the sock is already attached to BPF programs * this results in an error. */ - verdict = READ_ONCE(stab->bpf_verdict); - parse = READ_ONCE(stab->bpf_parse); - tx_msg = READ_ONCE(stab->bpf_tx_msg); + verdict = READ_ONCE(progs->bpf_verdict); + parse = READ_ONCE(progs->bpf_parse); + tx_msg = READ_ONCE(progs->bpf_tx_msg); if (parse && verdict) { /* bpf prog refcnt may be zero if a concurrent attach operation @@ -1703,11 +1729,11 @@ static int sock_map_ctx_update_elem(struct bpf_sock_ops_kern *skops, * we increment the refcnt. If this is the case abort with an * error. */ - verdict = bpf_prog_inc_not_zero(stab->bpf_verdict); + verdict = bpf_prog_inc_not_zero(progs->bpf_verdict); if (IS_ERR(verdict)) return PTR_ERR(verdict); - parse = bpf_prog_inc_not_zero(stab->bpf_parse); + parse = bpf_prog_inc_not_zero(progs->bpf_parse); if (IS_ERR(parse)) { bpf_prog_put(verdict); return PTR_ERR(parse); @@ -1715,7 +1741,7 @@ static int sock_map_ctx_update_elem(struct bpf_sock_ops_kern *skops, } if (tx_msg) { - tx_msg = bpf_prog_inc_not_zero(stab->bpf_tx_msg); + tx_msg = bpf_prog_inc_not_zero(progs->bpf_tx_msg); if (IS_ERR(tx_msg)) { if (verdict) bpf_prog_put(verdict); @@ -1748,7 +1774,7 @@ static int sock_map_ctx_update_elem(struct bpf_sock_ops_kern *skops, goto out_progs; } } else { - psock = smap_init_psock(sock, stab); + psock = smap_init_psock(sock, map->numa_node); if (IS_ERR(psock)) { err = PTR_ERR(psock); goto out_progs; @@ -1758,12 +1784,13 @@ static int sock_map_ctx_update_elem(struct bpf_sock_ops_kern *skops, new = true; } - e = kzalloc(sizeof(*e), GFP_ATOMIC | __GFP_NOWARN); - if (!e) { - err = -ENOMEM; - goto out_progs; + if (map_link) { + e = kzalloc(sizeof(*e), GFP_ATOMIC | __GFP_NOWARN); + if (!e) { + err = -ENOMEM; + goto out_progs; + } } - e->entry = &stab->sock_map[i]; /* 3. At this point we have a reference to a valid psock that is * running. Attach any BPF programs needed. @@ -1780,7 +1807,7 @@ static int sock_map_ctx_update_elem(struct bpf_sock_ops_kern *skops, err = smap_init_sock(psock, sock); if (err) goto out_free; - smap_init_progs(psock, stab, verdict, parse); + smap_init_progs(psock, verdict, parse); smap_start_sock(psock, sock); } @@ -1789,20 +1816,14 @@ static int sock_map_ctx_update_elem(struct bpf_sock_ops_kern *skops, * it with. Because we can only have a single set of programs if * old_sock has a strp we can stop it. */ - list_add_tail(&e->list, &psock->maps); - write_unlock_bh(&sock->sk_callback_lock); - - osock = xchg(&stab->sock_map[i], sock); - if (osock) { - struct smap_psock *opsock = smap_psock_sk(osock); - - write_lock_bh(&osock->sk_callback_lock); - smap_list_remove(opsock, &stab->sock_map[i]); - smap_release_sock(opsock, osock); - write_unlock_bh(&osock->sk_callback_lock); + if (map_link) { + e->entry = map_link; + list_add_tail(&e->list, &psock->maps); } - return 0; + write_unlock_bh(&sock->sk_callback_lock); + return err; out_free: + kfree(e); smap_release_sock(psock, sock); out_progs: if (verdict) @@ -1816,23 +1837,73 @@ out_progs: return err; } -int sock_map_prog(struct bpf_map *map, struct bpf_prog *prog, u32 type) +static int sock_map_ctx_update_elem(struct bpf_sock_ops_kern *skops, + struct bpf_map *map, + void *key, u64 flags) { struct bpf_stab *stab = container_of(map, struct bpf_stab, map); + struct bpf_sock_progs *progs = &stab->progs; + struct sock *osock, *sock; + 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; + + sock = skops->sk; + err = __sock_map_ctx_update_elem(map, progs, sock, &stab->sock_map[i], + key); + if (err) + goto out; + + osock = xchg(&stab->sock_map[i], sock); + if (osock) { + struct smap_psock *opsock = smap_psock_sk(osock); + + write_lock_bh(&osock->sk_callback_lock); + smap_list_remove(opsock, &stab->sock_map[i], NULL); + smap_release_sock(opsock, osock); + write_unlock_bh(&osock->sk_callback_lock); + } +out: + return err; +} + +int sock_map_prog(struct bpf_map *map, struct bpf_prog *prog, u32 type) +{ + struct bpf_sock_progs *progs; struct bpf_prog *orig; - if (unlikely(map->map_type != BPF_MAP_TYPE_SOCKMAP)) + if (map->map_type == BPF_MAP_TYPE_SOCKMAP) { + struct bpf_stab *stab = container_of(map, struct bpf_stab, map); + + progs = &stab->progs; + } else if (map->map_type == BPF_MAP_TYPE_SOCKHASH) { + struct bpf_htab *htab = container_of(map, struct bpf_htab, map); + + progs = &htab->progs; + } else { return -EINVAL; + } switch (type) { case BPF_SK_MSG_VERDICT: - orig = xchg(&stab->bpf_tx_msg, prog); + orig = xchg(&progs->bpf_tx_msg, prog); break; case BPF_SK_SKB_STREAM_PARSER: - orig = xchg(&stab->bpf_parse, prog); + orig = xchg(&progs->bpf_parse, prog); break; case BPF_SK_SKB_STREAM_VERDICT: - orig = xchg(&stab->bpf_verdict, prog); + orig = xchg(&progs->bpf_verdict, prog); break; default: return -EOPNOTSUPP; @@ -1880,21 +1951,421 @@ static int sock_map_update_elem(struct bpf_map *map, static void sock_map_release(struct bpf_map *map) { - struct bpf_stab *stab = container_of(map, struct bpf_stab, map); + struct bpf_sock_progs *progs; struct bpf_prog *orig; - orig = xchg(&stab->bpf_parse, NULL); + if (map->map_type == BPF_MAP_TYPE_SOCKMAP) { + struct bpf_stab *stab = container_of(map, struct bpf_stab, map); + + progs = &stab->progs; + } else { + struct bpf_htab *htab = container_of(map, struct bpf_htab, map); + + progs = &htab->progs; + } + + orig = xchg(&progs->bpf_parse, NULL); if (orig) bpf_prog_put(orig); - orig = xchg(&stab->bpf_verdict, NULL); + orig = xchg(&progs->bpf_verdict, NULL); if (orig) bpf_prog_put(orig); - orig = xchg(&stab->bpf_tx_msg, NULL); + orig = xchg(&progs->bpf_tx_msg, NULL); if (orig) bpf_prog_put(orig); } +static struct bpf_map *sock_hash_alloc(union bpf_attr *attr) +{ + struct bpf_htab *htab; + int i, err; + u64 cost; + + if (!capable(CAP_NET_ADMIN)) + return ERR_PTR(-EPERM); + + /* check sanity of attributes */ + if (attr->max_entries == 0 || attr->value_size != 4 || + attr->map_flags & ~SOCK_CREATE_FLAG_MASK) + return ERR_PTR(-EINVAL); + + if (attr->key_size > MAX_BPF_STACK) + /* eBPF programs initialize keys on stack, so they cannot be + * larger than max stack size + */ + return ERR_PTR(-E2BIG); + + err = bpf_tcp_ulp_register(); + if (err && err != -EEXIST) + return ERR_PTR(err); + + htab = kzalloc(sizeof(*htab), GFP_USER); + if (!htab) + return ERR_PTR(-ENOMEM); + + bpf_map_init_from_attr(&htab->map, attr); + + htab->n_buckets = roundup_pow_of_two(htab->map.max_entries); + htab->elem_size = sizeof(struct htab_elem) + + round_up(htab->map.key_size, 8); + err = -EINVAL; + if (htab->n_buckets == 0 || + htab->n_buckets > U32_MAX / sizeof(struct bucket)) + goto free_htab; + + cost = (u64) htab->n_buckets * sizeof(struct bucket) + + (u64) htab->elem_size * htab->map.max_entries; + + if (cost >= U32_MAX - PAGE_SIZE) + goto free_htab; + + htab->map.pages = round_up(cost, PAGE_SIZE) >> PAGE_SHIFT; + err = bpf_map_precharge_memlock(htab->map.pages); + if (err) + goto free_htab; + + err = -ENOMEM; + htab->buckets = bpf_map_area_alloc( + htab->n_buckets * sizeof(struct bucket), + htab->map.numa_node); + if (!htab->buckets) + goto free_htab; + + for (i = 0; i < htab->n_buckets; i++) { + INIT_HLIST_HEAD(&htab->buckets[i].head); + raw_spin_lock_init(&htab->buckets[i].lock); + } + + return &htab->map; +free_htab: + kfree(htab); + return ERR_PTR(err); +} + +static inline struct bucket *__select_bucket(struct bpf_htab *htab, u32 hash) +{ + return &htab->buckets[hash & (htab->n_buckets - 1)]; +} + +static inline struct hlist_head *select_bucket(struct bpf_htab *htab, u32 hash) +{ + return &__select_bucket(htab, hash)->head; +} + +static void sock_hash_free(struct bpf_map *map) +{ + struct bpf_htab *htab = container_of(map, struct bpf_htab, map); + int i; + + synchronize_rcu(); + + /* At this point no update, lookup or delete operations can happen. + * However, be aware we can still get a socket state event updates, + * and data ready callabacks that reference the psock from sk_user_data + * Also psock worker threads are still in-flight. So smap_release_sock + * will only free the psock after cancel_sync on the worker threads + * and a grace period expire to ensure psock is really safe to remove. + */ + rcu_read_lock(); + for (i = 0; i < htab->n_buckets; i++) { + struct hlist_head *head = select_bucket(htab, i); + struct hlist_node *n; + struct htab_elem *l; + + hlist_for_each_entry_safe(l, n, head, hash_node) { + struct sock *sock = l->sk; + struct smap_psock *psock; + + hlist_del_rcu(&l->hash_node); + write_lock_bh(&sock->sk_callback_lock); + 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 + * causing the refcnt to hit zero and sock user data + * (psock) to be null and queued for garbage collection. + */ + if (likely(psock)) { + smap_list_remove(psock, NULL, l); + smap_release_sock(psock, sock); + } + write_unlock_bh(&sock->sk_callback_lock); + kfree(l); + } + } + rcu_read_unlock(); + bpf_map_area_free(htab->buckets); + kfree(htab); +} + +static struct htab_elem *alloc_sock_hash_elem(struct bpf_htab *htab, + void *key, u32 key_size, u32 hash, + struct sock *sk, + struct htab_elem *old_elem) +{ + struct htab_elem *l_new; + + if (atomic_inc_return(&htab->count) > htab->map.max_entries) { + if (!old_elem) { + atomic_dec(&htab->count); + return ERR_PTR(-E2BIG); + } + } + l_new = kmalloc_node(htab->elem_size, GFP_ATOMIC | __GFP_NOWARN, + htab->map.numa_node); + if (!l_new) + return ERR_PTR(-ENOMEM); + + memcpy(l_new->key, key, key_size); + l_new->sk = sk; + l_new->hash = hash; + return l_new; +} + +static struct htab_elem *lookup_elem_raw(struct hlist_head *head, + u32 hash, void *key, u32 key_size) +{ + struct htab_elem *l; + + hlist_for_each_entry_rcu(l, head, hash_node) { + if (l->hash == hash && !memcmp(&l->key, key, key_size)) + return l; + } + + return NULL; +} + +static inline u32 htab_map_hash(const void *key, u32 key_len) +{ + return jhash(key, key_len, 0); +} + +static int sock_hash_get_next_key(struct bpf_map *map, + void *key, void *next_key) +{ + struct bpf_htab *htab = container_of(map, struct bpf_htab, map); + struct htab_elem *l, *next_l; + struct hlist_head *h; + u32 hash, key_size; + int i = 0; + + WARN_ON_ONCE(!rcu_read_lock_held()); + + key_size = map->key_size; + if (!key) + goto find_first_elem; + hash = htab_map_hash(key, key_size); + h = select_bucket(htab, hash); + + l = lookup_elem_raw(h, hash, key, key_size); + if (!l) + goto find_first_elem; + next_l = hlist_entry_safe( + rcu_dereference_raw(hlist_next_rcu(&l->hash_node)), + struct htab_elem, hash_node); + if (next_l) { + memcpy(next_key, next_l->key, key_size); + return 0; + } + + /* no more elements in this hash list, go to the next bucket */ + i = hash & (htab->n_buckets - 1); + i++; + +find_first_elem: + /* iterate over buckets */ + for (; i < htab->n_buckets; i++) { + h = select_bucket(htab, i); + + /* pick first element in the bucket */ + next_l = hlist_entry_safe( + rcu_dereference_raw(hlist_first_rcu(h)), + struct htab_elem, hash_node); + if (next_l) { + /* if it's not empty, just return it */ + memcpy(next_key, next_l->key, key_size); + return 0; + } + } + + /* iterated over all buckets and all elements */ + return -ENOENT; +} + +static int sock_hash_ctx_update_elem(struct bpf_sock_ops_kern *skops, + struct bpf_map *map, + void *key, u64 map_flags) +{ + struct bpf_htab *htab = container_of(map, struct bpf_htab, map); + struct bpf_sock_progs *progs = &htab->progs; + struct htab_elem *l_new = NULL, *l_old; + struct smap_psock_map_entry *e = NULL; + struct hlist_head *head; + struct smap_psock *psock; + u32 key_size, hash; + struct sock *sock; + struct bucket *b; + int err; + + sock = skops->sk; + + if (sock->sk_type != SOCK_STREAM || + sock->sk_protocol != IPPROTO_TCP) + return -EOPNOTSUPP; + + if (unlikely(map_flags > BPF_EXIST)) + return -EINVAL; + + e = kzalloc(sizeof(*e), GFP_ATOMIC | __GFP_NOWARN); + if (!e) + return -ENOMEM; + + WARN_ON_ONCE(!rcu_read_lock_held()); + key_size = map->key_size; + hash = htab_map_hash(key, key_size); + b = __select_bucket(htab, hash); + head = &b->head; + + err = __sock_map_ctx_update_elem(map, progs, sock, NULL, key); + if (err) + goto err; + + /* bpf_map_update_elem() can be called in_irq() */ + raw_spin_lock_bh(&b->lock); + l_old = lookup_elem_raw(head, hash, key, key_size); + if (l_old && map_flags == BPF_NOEXIST) { + err = -EEXIST; + goto bucket_err; + } + if (!l_old && map_flags == BPF_EXIST) { + err = -ENOENT; + goto bucket_err; + } + + l_new = alloc_sock_hash_elem(htab, key, key_size, hash, sock, l_old); + if (IS_ERR(l_new)) { + err = PTR_ERR(l_new); + goto bucket_err; + } + + psock = smap_psock_sk(sock); + if (unlikely(!psock)) { + err = -EINVAL; + goto bucket_err; + } + + e->hash_link = l_new; + e->htab = container_of(map, struct bpf_htab, map); + list_add_tail(&e->list, &psock->maps); + + /* add new element to the head of the list, so that + * concurrent search will find it before old elem + */ + hlist_add_head_rcu(&l_new->hash_node, head); + if (l_old) { + psock = smap_psock_sk(l_old->sk); + + hlist_del_rcu(&l_old->hash_node); + smap_list_remove(psock, NULL, l_old); + smap_release_sock(psock, l_old->sk); + free_htab_elem(htab, l_old); + } + raw_spin_unlock_bh(&b->lock); + return 0; +bucket_err: + raw_spin_unlock_bh(&b->lock); +err: + kfree(e); + psock = smap_psock_sk(sock); + if (psock) + smap_release_sock(psock, sock); + return err; +} + +static int sock_hash_update_elem(struct bpf_map *map, + void *key, void *value, u64 flags) +{ + struct bpf_sock_ops_kern skops; + u32 fd = *(u32 *)value; + struct socket *socket; + int err; + + socket = sockfd_lookup(fd, &err); + if (!socket) + return err; + + skops.sk = socket->sk; + if (!skops.sk) { + fput(socket->file); + return -EINVAL; + } + + err = sock_hash_ctx_update_elem(&skops, map, key, flags); + fput(socket->file); + return err; +} + +static int sock_hash_delete_elem(struct bpf_map *map, void *key) +{ + struct bpf_htab *htab = container_of(map, struct bpf_htab, map); + struct hlist_head *head; + struct bucket *b; + struct htab_elem *l; + u32 hash, key_size; + int ret = -ENOENT; + + key_size = map->key_size; + hash = htab_map_hash(key, key_size); + b = __select_bucket(htab, hash); + head = &b->head; + + raw_spin_lock_bh(&b->lock); + l = lookup_elem_raw(head, hash, key, key_size); + if (l) { + struct sock *sock = l->sk; + struct smap_psock *psock; + + hlist_del_rcu(&l->hash_node); + write_lock_bh(&sock->sk_callback_lock); + 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 + * causing the refcnt to hit zero and sock user data (psock) + * to be null and queued for garbage collection. + */ + if (likely(psock)) { + smap_list_remove(psock, NULL, l); + smap_release_sock(psock, sock); + } + write_unlock_bh(&sock->sk_callback_lock); + free_htab_elem(htab, l); + ret = 0; + } + raw_spin_unlock_bh(&b->lock); + return ret; +} + +struct sock *__sock_hash_lookup_elem(struct bpf_map *map, void *key) +{ + struct bpf_htab *htab = container_of(map, struct bpf_htab, map); + struct hlist_head *head; + struct htab_elem *l; + u32 key_size, hash; + struct bucket *b; + struct sock *sk; + + key_size = map->key_size; + hash = htab_map_hash(key, key_size); + b = __select_bucket(htab, hash); + head = &b->head; + + raw_spin_lock_bh(&b->lock); + l = lookup_elem_raw(head, hash, key, key_size); + sk = l ? l->sk : NULL; + raw_spin_unlock_bh(&b->lock); + return sk; +} + const struct bpf_map_ops sock_map_ops = { .map_alloc = sock_map_alloc, .map_free = sock_map_free, @@ -1905,6 +2376,15 @@ const struct bpf_map_ops sock_map_ops = { .map_release_uref = sock_map_release, }; +const struct bpf_map_ops sock_hash_ops = { + .map_alloc = sock_hash_alloc, + .map_free = sock_hash_free, + .map_lookup_elem = sock_map_lookup, + .map_get_next_key = sock_hash_get_next_key, + .map_update_elem = sock_hash_update_elem, + .map_delete_elem = sock_hash_delete_elem, +}; + BPF_CALL_4(bpf_sock_map_update, struct bpf_sock_ops_kern *, bpf_sock, struct bpf_map *, map, void *, key, u64, flags) { @@ -1922,3 +2402,21 @@ const struct bpf_func_proto bpf_sock_map_update_proto = { .arg3_type = ARG_PTR_TO_MAP_KEY, .arg4_type = ARG_ANYTHING, }; + +BPF_CALL_4(bpf_sock_hash_update, struct bpf_sock_ops_kern *, bpf_sock, + struct bpf_map *, map, void *, key, u64, flags) +{ + WARN_ON_ONCE(!rcu_read_lock_held()); + return sock_hash_ctx_update_elem(bpf_sock, map, key, flags); +} + +const struct bpf_func_proto bpf_sock_hash_update_proto = { + .func = bpf_sock_hash_update, + .gpl_only = false, + .pkt_access = true, + .ret_type = RET_INTEGER, + .arg1_type = ARG_PTR_TO_CTX, + .arg2_type = ARG_CONST_MAP_PTR, + .arg3_type = ARG_PTR_TO_MAP_KEY, + .arg4_type = ARG_ANYTHING, +}; diff --git a/kernel/bpf/stackmap.c b/kernel/bpf/stackmap.c index 3ba102b41512..b59ace0f0f09 100644 --- a/kernel/bpf/stackmap.c +++ b/kernel/bpf/stackmap.c @@ -11,6 +11,7 @@ #include <linux/perf_event.h> #include <linux/elf.h> #include <linux/pagemap.h> +#include <linux/irq_work.h> #include "percpu_freelist.h" #define STACK_CREATE_FLAG_MASK \ @@ -32,6 +33,23 @@ struct bpf_stack_map { struct stack_map_bucket *buckets[]; }; +/* irq_work to run up_read() for build_id lookup in nmi context */ +struct stack_map_irq_work { + struct irq_work irq_work; + struct rw_semaphore *sem; +}; + +static void do_up_read(struct irq_work *entry) +{ + struct stack_map_irq_work *work; + + work = container_of(entry, struct stack_map_irq_work, irq_work); + up_read(work->sem); + work->sem = NULL; +} + +static DEFINE_PER_CPU(struct stack_map_irq_work, up_read_work); + static inline bool stack_map_use_build_id(struct bpf_map *map) { return (map->map_flags & BPF_F_STACK_BUILD_ID); @@ -267,17 +285,27 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs, { int i; struct vm_area_struct *vma; + bool in_nmi_ctx = in_nmi(); + bool irq_work_busy = false; + struct stack_map_irq_work *work; + + if (in_nmi_ctx) { + work = this_cpu_ptr(&up_read_work); + if (work->irq_work.flags & IRQ_WORK_BUSY) + /* cannot queue more up_read, fallback */ + irq_work_busy = true; + } /* - * We cannot do up_read() in nmi context, so build_id lookup is - * only supported for non-nmi events. If at some point, it is - * possible to run find_vma() without taking the semaphore, we - * would like to allow build_id lookup in nmi context. + * We cannot do up_read() in nmi context. To do build_id lookup + * in nmi context, we need to run up_read() in irq_work. We use + * a percpu variable to do the irq_work. If the irq_work is + * already used by another lookup, we fall back to report ips. * * Same fallback is used for kernel stack (!user) on a stackmap * with build_id. */ - if (!user || !current || !current->mm || in_nmi() || + if (!user || !current || !current->mm || irq_work_busy || down_read_trylock(¤t->mm->mmap_sem) == 0) { /* cannot access current->mm, fall back to ips */ for (i = 0; i < trace_nr; i++) { @@ -299,7 +327,13 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs, - vma->vm_start; id_offs[i].status = BPF_STACK_BUILD_ID_VALID; } - up_read(¤t->mm->mmap_sem); + + if (!in_nmi_ctx) { + up_read(¤t->mm->mmap_sem); + } else { + work->sem = ¤t->mm->mmap_sem; + irq_work_queue(&work->irq_work); + } } BPF_CALL_3(bpf_get_stackid, struct pt_regs *, regs, struct bpf_map *, map, @@ -575,3 +609,16 @@ const struct bpf_map_ops stack_map_ops = { .map_update_elem = stack_map_update_elem, .map_delete_elem = stack_map_delete_elem, }; + +static int __init stack_map_init(void) +{ + int cpu; + struct stack_map_irq_work *work; + + for_each_possible_cpu(cpu) { + work = per_cpu_ptr(&up_read_work, cpu); + init_irq_work(&work->irq_work, do_up_read); + } + return 0; +} +subsys_initcall(stack_map_init); diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c index c286e75ec087..bfcde949c7f8 100644 --- a/kernel/bpf/syscall.c +++ b/kernel/bpf/syscall.c @@ -255,7 +255,6 @@ static void bpf_map_free_deferred(struct work_struct *work) bpf_map_uncharge_memlock(map); security_bpf_map_free(map); - btf_put(map->btf); /* implementation dependent freeing */ map->ops->map_free(map); } @@ -276,6 +275,7 @@ static void __bpf_map_put(struct bpf_map *map, bool do_idr_lock) if (atomic_dec_and_test(&map->refcnt)) { /* bpf_map_free_id() must be called first */ bpf_map_free_id(map, do_idr_lock); + btf_put(map->btf); INIT_WORK(&map->work, bpf_map_free_deferred); schedule_work(&map->work); } @@ -2011,6 +2011,12 @@ static int bpf_map_get_info_by_fd(struct bpf_map *map, info.map_flags = map->map_flags; memcpy(info.name, map->name, sizeof(map->name)); + if (map->btf) { + info.btf_id = btf_id(map->btf); + info.btf_key_id = map->btf_key_id; + info.btf_value_id = map->btf_value_id; + } + if (bpf_map_is_dev_bound(map)) { err = bpf_map_offload_info_fill(&info, map); if (err) @@ -2024,6 +2030,21 @@ static int bpf_map_get_info_by_fd(struct bpf_map *map, return 0; } +static int bpf_btf_get_info_by_fd(struct btf *btf, + const union bpf_attr *attr, + union bpf_attr __user *uattr) +{ + struct bpf_btf_info __user *uinfo = u64_to_user_ptr(attr->info.info); + u32 info_len = attr->info.info_len; + int err; + + err = check_uarg_tail_zero(uinfo, sizeof(*uinfo), info_len); + if (err) + return err; + + return btf_get_info_by_fd(btf, attr, uattr); +} + #define BPF_OBJ_GET_INFO_BY_FD_LAST_FIELD info.info static int bpf_obj_get_info_by_fd(const union bpf_attr *attr, @@ -2047,7 +2068,7 @@ static int bpf_obj_get_info_by_fd(const union bpf_attr *attr, err = bpf_map_get_info_by_fd(f.file->private_data, attr, uattr); else if (f.file->f_op == &btf_fops) - err = btf_get_info_by_fd(f.file->private_data, attr, uattr); + err = bpf_btf_get_info_by_fd(f.file->private_data, attr, uattr); else err = -EINVAL; @@ -2068,6 +2089,19 @@ static int bpf_btf_load(const union bpf_attr *attr) return btf_new_fd(attr); } +#define BPF_BTF_GET_FD_BY_ID_LAST_FIELD btf_id + +static int bpf_btf_get_fd_by_id(const union bpf_attr *attr) +{ + if (CHECK_ATTR(BPF_BTF_GET_FD_BY_ID)) + return -EINVAL; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + return btf_get_fd_by_id(attr->btf_id); +} + SYSCALL_DEFINE3(bpf, int, cmd, union bpf_attr __user *, uattr, unsigned int, size) { union bpf_attr attr = {}; @@ -2151,6 +2185,9 @@ SYSCALL_DEFINE3(bpf, int, cmd, union bpf_attr __user *, uattr, unsigned int, siz case BPF_BTF_LOAD: err = bpf_btf_load(&attr); break; + case BPF_BTF_GET_FD_BY_ID: + err = bpf_btf_get_fd_by_id(&attr); + break; default: err = -EINVAL; break; diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c index d5e1a6c4165d..a9e4b1372da6 100644 --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c @@ -2093,6 +2093,13 @@ static int check_map_func_compatibility(struct bpf_verifier_env *env, func_id != BPF_FUNC_msg_redirect_map) goto error; break; + case BPF_MAP_TYPE_SOCKHASH: + if (func_id != BPF_FUNC_sk_redirect_hash && + func_id != BPF_FUNC_sock_hash_update && + func_id != BPF_FUNC_map_delete_elem && + func_id != BPF_FUNC_msg_redirect_hash) + goto error; + break; default: break; } @@ -2130,11 +2137,14 @@ static int check_map_func_compatibility(struct bpf_verifier_env *env, break; case BPF_FUNC_sk_redirect_map: case BPF_FUNC_msg_redirect_map: + case BPF_FUNC_sock_map_update: if (map->map_type != BPF_MAP_TYPE_SOCKMAP) goto error; break; - case BPF_FUNC_sock_map_update: - if (map->map_type != BPF_MAP_TYPE_SOCKMAP) + case BPF_FUNC_sk_redirect_hash: + case BPF_FUNC_msg_redirect_hash: + case BPF_FUNC_sock_hash_update: + if (map->map_type != BPF_MAP_TYPE_SOCKHASH) goto error; break; default: @@ -5215,7 +5225,7 @@ static int convert_ctx_accesses(struct bpf_verifier_env *env) } } - if (!ops->convert_ctx_access) + if (!ops->convert_ctx_access || bpf_prog_is_dev_bound(env->prog->aux)) return 0; insn = env->prog->insnsi + delta; |