renegade-project/linux-next
2
0
Fork 0
mirror of https://github.com/edk2-porting/linux-next.git synced 2024-12-23 04:34:11 +08:00
Code
6e830c2f6c
linux-next/net/core/sock_map.c

1091 lines
26 KiB
C
Raw Normal View History

bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
// SPDX-License-Identifier: GPL-2.0
/* Copyright (c) 2017 - 2018 Covalent IO, Inc. http://covalent.io */
#include <linux/bpf.h>
#include <linux/filter.h>
#include <linux/errno.h>
#include <linux/file.h>
#include <linux/net.h>
#include <linux/workqueue.h>
#include <linux/skmsg.h>
#include <linux/list.h>
#include <linux/jhash.h>
struct bpf_stab {
struct bpf_map map;
struct sock **sks;
struct sk_psock_progs progs;
raw_spinlock_t lock;
};
#define SOCK_CREATE_FLAG_MASK \
(BPF_F_NUMA_NODE | BPF_F_RDONLY | BPF_F_WRONLY)
static struct bpf_map *sock_map_alloc(union bpf_attr *attr)
{
struct bpf_stab *stab;
u64 cost;
int err;
if (!capable(CAP_NET_ADMIN))
return ERR_PTR(-EPERM);
if (attr->max_entries == 0 ||
attr->key_size != 4 ||
attr->value_size != 4 ||
attr->map_flags & ~SOCK_CREATE_FLAG_MASK)
return ERR_PTR(-EINVAL);
stab = kzalloc(sizeof(*stab), GFP_USER);
if (!stab)
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 *);
bpf: move memory size checks to bpf_map_charge_init() Most bpf map types doing similar checks and bytes to pages conversion during memory allocation and charging. Let's unify these checks by moving them into bpf_map_charge_init(). Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Song Liu <songliubraving@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-05-30 09:03:59 +08:00
err = bpf_map_charge_init(&stab->map.memory, cost);
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
if (err)
goto free_stab;
stab->sks = bpf_map_area_alloc(stab->map.max_entries *
sizeof(struct sock *),
stab->map.numa_node);
if (stab->sks)
return &stab->map;
err = -ENOMEM;
bpf: rework memlock-based memory accounting for maps In order to unify the existing memlock charging code with the memcg-based memory accounting, which will be added later, let's rework the current scheme. Currently the following design is used: 1) .alloc() callback optionally checks if the allocation will likely succeed using bpf_map_precharge_memlock() 2) .alloc() performs actual allocations 3) .alloc() callback calculates map cost and sets map.memory.pages 4) map_create() calls bpf_map_init_memlock() which sets map.memory.user and performs actual charging; in case of failure the map is destroyed <map is in use> 1) bpf_map_free_deferred() calls bpf_map_release_memlock(), which performs uncharge and releases the user 2) .map_free() callback releases the memory The scheme can be simplified and made more robust: 1) .alloc() calculates map cost and calls bpf_map_charge_init() 2) bpf_map_charge_init() sets map.memory.user and performs actual charge 3) .alloc() performs actual allocations <map is in use> 1) .map_free() callback releases the memory 2) bpf_map_charge_finish() performs uncharge and releases the user The new scheme also allows to reuse bpf_map_charge_init()/finish() functions for memcg-based accounting. Because charges are performed before actual allocations and uncharges after freeing the memory, no bogus memory pressure can be created. In cases when the map structure is not available (e.g. it's not created yet, or is already destroyed), on-stack bpf_map_memory structure is used. The charge can be transferred with the bpf_map_charge_move() function. Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Song Liu <songliubraving@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-05-30 09:03:58 +08:00
bpf_map_charge_finish(&stab->map.memory);
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
free_stab:
kfree(stab);
return ERR_PTR(err);
}
int sock_map_get_from_fd(const union bpf_attr *attr, struct bpf_prog *prog)
{
u32 ufd = attr->target_fd;
struct bpf_map *map;
struct fd f;
int ret;
f = fdget(ufd);
map = __bpf_map_get(f);
if (IS_ERR(map))
return PTR_ERR(map);
ret = sock_map_prog_update(map, prog, attr->attach_type);
fdput(f);
return ret;
}
static void sock_map_sk_acquire(struct sock *sk)
__acquires(&sk->sk_lock.slock)
{
lock_sock(sk);
preempt_disable();
rcu_read_lock();
}
static void sock_map_sk_release(struct sock *sk)
__releases(&sk->sk_lock.slock)
{
rcu_read_unlock();
preempt_enable();
release_sock(sk);
}
static void sock_map_add_link(struct sk_psock *psock,
struct sk_psock_link *link,
struct bpf_map *map, void *link_raw)
{
link->link_raw = link_raw;
link->map = map;
spin_lock_bh(&psock->link_lock);
list_add_tail(&link->list, &psock->link);
spin_unlock_bh(&psock->link_lock);
}
static void sock_map_del_link(struct sock *sk,
struct sk_psock *psock, void *link_raw)
{
struct sk_psock_link *link, *tmp;
bool strp_stop = false;
spin_lock_bh(&psock->link_lock);
list_for_each_entry_safe(link, tmp, &psock->link, list) {
if (link->link_raw == link_raw) {
struct bpf_map *map = link->map;
struct bpf_stab *stab = container_of(map, struct bpf_stab,
map);
if (psock->parser.enabled && stab->progs.skb_parser)
strp_stop = true;
list_del(&link->list);
sk_psock_free_link(link);
}
}
spin_unlock_bh(&psock->link_lock);
if (strp_stop) {
write_lock_bh(&sk->sk_callback_lock);
sk_psock_stop_strp(sk, psock);
write_unlock_bh(&sk->sk_callback_lock);
}
}
static void sock_map_unref(struct sock *sk, void *link_raw)
{
struct sk_psock *psock = sk_psock(sk);
if (likely(psock)) {
sock_map_del_link(sk, psock, link_raw);
sk_psock_put(sk, psock);
}
}
static int sock_map_link(struct bpf_map *map, struct sk_psock_progs *progs,
struct sock *sk)
{
struct bpf_prog *msg_parser, *skb_parser, *skb_verdict;
bool skb_progs, sk_psock_is_new = false;
struct sk_psock *psock;
int ret;
skb_verdict = READ_ONCE(progs->skb_verdict);
skb_parser = READ_ONCE(progs->skb_parser);
skb_progs = skb_parser && skb_verdict;
if (skb_progs) {
skb_verdict = bpf_prog_inc_not_zero(skb_verdict);
if (IS_ERR(skb_verdict))
return PTR_ERR(skb_verdict);
skb_parser = bpf_prog_inc_not_zero(skb_parser);
if (IS_ERR(skb_parser)) {
bpf_prog_put(skb_verdict);
return PTR_ERR(skb_parser);
}
}
msg_parser = READ_ONCE(progs->msg_parser);
if (msg_parser) {
msg_parser = bpf_prog_inc_not_zero(msg_parser);
if (IS_ERR(msg_parser)) {
ret = PTR_ERR(msg_parser);
goto out;
}
}
bpf: skmsg, fix psock create on existing kcm/tls port Before using the psock returned by sk_psock_get() when adding it to a sockmap we need to ensure it is actually a sockmap based psock. Previously we were only checking this after incrementing the reference counter which was an error. This resulted in a slab-out-of-bounds error when the psock was not actually a sockmap type. This moves the check up so the reference counter is only used if it is a sockmap psock. Eric reported the following KASAN BUG, BUG: KASAN: slab-out-of-bounds in atomic_read include/asm-generic/atomic-instrumented.h:21 [inline] BUG: KASAN: slab-out-of-bounds in refcount_inc_not_zero_checked+0x97/0x2f0 lib/refcount.c:120 Read of size 4 at addr ffff88019548be58 by task syz-executor4/22387 CPU: 1 PID: 22387 Comm: syz-executor4 Not tainted 4.19.0-rc7+ #264 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 Call Trace: __dump_stack lib/dump_stack.c:77 [inline] dump_stack+0x1c4/0x2b4 lib/dump_stack.c:113 print_address_description.cold.8+0x9/0x1ff mm/kasan/report.c:256 kasan_report_error mm/kasan/report.c:354 [inline] kasan_report.cold.9+0x242/0x309 mm/kasan/report.c:412 check_memory_region_inline mm/kasan/kasan.c:260 [inline] check_memory_region+0x13e/0x1b0 mm/kasan/kasan.c:267 kasan_check_read+0x11/0x20 mm/kasan/kasan.c:272 atomic_read include/asm-generic/atomic-instrumented.h:21 [inline] refcount_inc_not_zero_checked+0x97/0x2f0 lib/refcount.c:120 sk_psock_get include/linux/skmsg.h:379 [inline] sock_map_link.isra.6+0x41f/0xe30 net/core/sock_map.c:178 sock_hash_update_common+0x19b/0x11e0 net/core/sock_map.c:669 sock_hash_update_elem+0x306/0x470 net/core/sock_map.c:738 map_update_elem+0x819/0xdf0 kernel/bpf/syscall.c:818 Signed-off-by: John Fastabend <john.fastabend@gmail.com> Reported-by: Eric Dumazet <eric.dumazet@gmail.com> Fixes: 604326b41a6f ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-10-19 04:58:35 +08:00
psock = sk_psock_get_checked(sk);
if (IS_ERR(psock)) {
ret = PTR_ERR(psock);
goto out_progs;
}
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
if (psock) {
if ((msg_parser && READ_ONCE(psock->progs.msg_parser)) ||
(skb_progs && READ_ONCE(psock->progs.skb_parser))) {
sk_psock_put(sk, psock);
ret = -EBUSY;
goto out_progs;
}
} else {
psock = sk_psock_init(sk, map->numa_node);
if (!psock) {
ret = -ENOMEM;
goto out_progs;
}
sk_psock_is_new = true;
}
if (msg_parser)
psock_set_prog(&psock->progs.msg_parser, msg_parser);
if (sk_psock_is_new) {
ret = tcp_bpf_init(sk);
if (ret < 0)
goto out_drop;
} else {
tcp_bpf_reinit(sk);
}
write_lock_bh(&sk->sk_callback_lock);
if (skb_progs && !psock->parser.enabled) {
ret = sk_psock_init_strp(sk, psock);
if (ret) {
write_unlock_bh(&sk->sk_callback_lock);
goto out_drop;
}
psock_set_prog(&psock->progs.skb_verdict, skb_verdict);
psock_set_prog(&psock->progs.skb_parser, skb_parser);
sk_psock_start_strp(sk, psock);
}
write_unlock_bh(&sk->sk_callback_lock);
return 0;
out_drop:
sk_psock_put(sk, psock);
out_progs:
if (msg_parser)
bpf_prog_put(msg_parser);
out:
if (skb_progs) {
bpf_prog_put(skb_verdict);
bpf_prog_put(skb_parser);
}
return ret;
}
bpf, sockmap: Don't set up upcalls and progs for listening sockets Now that sockmap/sockhash can hold listening sockets, when setting up the psock we will (i) grab references to verdict/parser progs, and (2) override socket upcalls sk_data_ready and sk_write_space. However, since we cannot redirect to listening sockets so we don't need to link the socket to the BPF progs. And more importantly we don't want the listening socket to have overridden upcalls because they would get inherited by child sockets cloned from it. Introduce a separate initialization path for listening sockets that does not change the upcalls and ignores the BPF progs. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-6-jakub@cloudflare.com
2020-02-19 01:10:17 +08:00
static int sock_map_link_no_progs(struct bpf_map *map, struct sock *sk)
{
struct sk_psock *psock;
int ret;
psock = sk_psock_get_checked(sk);
if (IS_ERR(psock))
return PTR_ERR(psock);
if (psock) {
tcp_bpf_reinit(sk);
return 0;
}
psock = sk_psock_init(sk, map->numa_node);
if (!psock)
return -ENOMEM;
ret = tcp_bpf_init(sk);
if (ret < 0)
sk_psock_put(sk, psock);
return ret;
}
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
static void sock_map_free(struct bpf_map *map)
{
struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
int i;
synchronize_rcu();
raw_spin_lock_bh(&stab->lock);
for (i = 0; i < stab->map.max_entries; i++) {
struct sock **psk = &stab->sks[i];
struct sock *sk;
sk = xchg(psk, NULL);
bpf: Sockmap, ensure sock lock held during tear down The sock_map_free() and sock_hash_free() paths used to delete sockmap and sockhash maps walk the maps and destroy psock and bpf state associated with the socks in the map. When done the socks no longer have BPF programs attached and will function normally. This can happen while the socks in the map are still "live" meaning data may be sent/received during the walk. Currently, though we don't take the sock_lock when the psock and bpf state is removed through this path. Specifically, this means we can be writing into the ops structure pointers such as sendmsg, sendpage, recvmsg, etc. while they are also being called from the networking side. This is not safe, we never used proper READ_ONCE/WRITE_ONCE semantics here if we believed it was safe. Further its not clear to me its even a good idea to try and do this on "live" sockets while networking side might also be using the socket. Instead of trying to reason about using the socks from both sides lets realize that every use case I'm aware of rarely deletes maps, in fact kubernetes/Cilium case builds map at init and never tears it down except on errors. So lets do the simple fix and grab sock lock. This patch wraps sock deletes from maps in sock lock and adds some annotations so we catch any other cases easier. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Song Liu <songliubraving@fb.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/bpf/20200111061206.8028-3-john.fastabend@gmail.com
2020-01-11 14:12:00 +08:00
if (sk) {
lock_sock(sk);
bpf, sockmap: Don't sleep while holding RCU lock on tear-down rcu_read_lock is needed to protect access to psock inside sock_map_unref when tearing down the map. However, we can't afford to sleep in lock_sock while in RCU read-side critical section. Grab the RCU lock only after we have locked the socket. This fixes RCU warnings triggerable on a VM with 1 vCPU when free'ing a sockmap/sockhash that contains at least one socket: | ============================= | WARNING: suspicious RCU usage | 5.5.0-04005-g8fc91b972b73 #450 Not tainted | ----------------------------- | include/linux/rcupdate.h:272 Illegal context switch in RCU read-side critical section! | | other info that might help us debug this: | | | rcu_scheduler_active = 2, debug_locks = 1 | 4 locks held by kworker/0:1/62: | #0: ffff88813b019748 ((wq_completion)events){+.+.}, at: process_one_work+0x1d7/0x5e0 | #1: ffffc900000abe50 ((work_completion)(&map->work)){+.+.}, at: process_one_work+0x1d7/0x5e0 | #2: ffffffff82065d20 (rcu_read_lock){....}, at: sock_map_free+0x5/0x170 | #3: ffff8881368c5df8 (&stab->lock){+...}, at: sock_map_free+0x64/0x170 | | stack backtrace: | CPU: 0 PID: 62 Comm: kworker/0:1 Not tainted 5.5.0-04005-g8fc91b972b73 #450 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_073836-buildvm-ppc64le-16.ppc.fedoraproject.org-3.fc31 04/01/2014 | Workqueue: events bpf_map_free_deferred | Call Trace: | dump_stack+0x71/0xa0 | ___might_sleep+0x105/0x190 | lock_sock_nested+0x28/0x90 | sock_map_free+0x95/0x170 | bpf_map_free_deferred+0x58/0x80 | process_one_work+0x260/0x5e0 | worker_thread+0x4d/0x3e0 | kthread+0x108/0x140 | ? process_one_work+0x5e0/0x5e0 | ? kthread_park+0x90/0x90 | ret_from_fork+0x3a/0x50 | ============================= | WARNING: suspicious RCU usage | 5.5.0-04005-g8fc91b972b73-dirty #452 Not tainted | ----------------------------- | include/linux/rcupdate.h:272 Illegal context switch in RCU read-side critical section! | | other info that might help us debug this: | | | rcu_scheduler_active = 2, debug_locks = 1 | 4 locks held by kworker/0:1/62: | #0: ffff88813b019748 ((wq_completion)events){+.+.}, at: process_one_work+0x1d7/0x5e0 | #1: ffffc900000abe50 ((work_completion)(&map->work)){+.+.}, at: process_one_work+0x1d7/0x5e0 | #2: ffffffff82065d20 (rcu_read_lock){....}, at: sock_hash_free+0x5/0x1d0 | #3: ffff888139966e00 (&htab->buckets[i].lock){+...}, at: sock_hash_free+0x92/0x1d0 | | stack backtrace: | CPU: 0 PID: 62 Comm: kworker/0:1 Not tainted 5.5.0-04005-g8fc91b972b73-dirty #452 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_073836-buildvm-ppc64le-16.ppc.fedoraproject.org-3.fc31 04/01/2014 | Workqueue: events bpf_map_free_deferred | Call Trace: | dump_stack+0x71/0xa0 | ___might_sleep+0x105/0x190 | lock_sock_nested+0x28/0x90 | sock_hash_free+0xec/0x1d0 | bpf_map_free_deferred+0x58/0x80 | process_one_work+0x260/0x5e0 | worker_thread+0x4d/0x3e0 | kthread+0x108/0x140 | ? process_one_work+0x5e0/0x5e0 | ? kthread_park+0x90/0x90 | ret_from_fork+0x3a/0x50 Fixes: 7e81a3530206 ("bpf: Sockmap, ensure sock lock held during tear down") Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200206111652.694507-2-jakub@cloudflare.com
2020-02-06 19:16:50 +08:00
rcu_read_lock();
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
sock_map_unref(sk, psk);
bpf, sockmap: Don't sleep while holding RCU lock on tear-down rcu_read_lock is needed to protect access to psock inside sock_map_unref when tearing down the map. However, we can't afford to sleep in lock_sock while in RCU read-side critical section. Grab the RCU lock only after we have locked the socket. This fixes RCU warnings triggerable on a VM with 1 vCPU when free'ing a sockmap/sockhash that contains at least one socket: | ============================= | WARNING: suspicious RCU usage | 5.5.0-04005-g8fc91b972b73 #450 Not tainted | ----------------------------- | include/linux/rcupdate.h:272 Illegal context switch in RCU read-side critical section! | | other info that might help us debug this: | | | rcu_scheduler_active = 2, debug_locks = 1 | 4 locks held by kworker/0:1/62: | #0: ffff88813b019748 ((wq_completion)events){+.+.}, at: process_one_work+0x1d7/0x5e0 | #1: ffffc900000abe50 ((work_completion)(&map->work)){+.+.}, at: process_one_work+0x1d7/0x5e0 | #2: ffffffff82065d20 (rcu_read_lock){....}, at: sock_map_free+0x5/0x170 | #3: ffff8881368c5df8 (&stab->lock){+...}, at: sock_map_free+0x64/0x170 | | stack backtrace: | CPU: 0 PID: 62 Comm: kworker/0:1 Not tainted 5.5.0-04005-g8fc91b972b73 #450 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_073836-buildvm-ppc64le-16.ppc.fedoraproject.org-3.fc31 04/01/2014 | Workqueue: events bpf_map_free_deferred | Call Trace: | dump_stack+0x71/0xa0 | ___might_sleep+0x105/0x190 | lock_sock_nested+0x28/0x90 | sock_map_free+0x95/0x170 | bpf_map_free_deferred+0x58/0x80 | process_one_work+0x260/0x5e0 | worker_thread+0x4d/0x3e0 | kthread+0x108/0x140 | ? process_one_work+0x5e0/0x5e0 | ? kthread_park+0x90/0x90 | ret_from_fork+0x3a/0x50 | ============================= | WARNING: suspicious RCU usage | 5.5.0-04005-g8fc91b972b73-dirty #452 Not tainted | ----------------------------- | include/linux/rcupdate.h:272 Illegal context switch in RCU read-side critical section! | | other info that might help us debug this: | | | rcu_scheduler_active = 2, debug_locks = 1 | 4 locks held by kworker/0:1/62: | #0: ffff88813b019748 ((wq_completion)events){+.+.}, at: process_one_work+0x1d7/0x5e0 | #1: ffffc900000abe50 ((work_completion)(&map->work)){+.+.}, at: process_one_work+0x1d7/0x5e0 | #2: ffffffff82065d20 (rcu_read_lock){....}, at: sock_hash_free+0x5/0x1d0 | #3: ffff888139966e00 (&htab->buckets[i].lock){+...}, at: sock_hash_free+0x92/0x1d0 | | stack backtrace: | CPU: 0 PID: 62 Comm: kworker/0:1 Not tainted 5.5.0-04005-g8fc91b972b73-dirty #452 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_073836-buildvm-ppc64le-16.ppc.fedoraproject.org-3.fc31 04/01/2014 | Workqueue: events bpf_map_free_deferred | Call Trace: | dump_stack+0x71/0xa0 | ___might_sleep+0x105/0x190 | lock_sock_nested+0x28/0x90 | sock_hash_free+0xec/0x1d0 | bpf_map_free_deferred+0x58/0x80 | process_one_work+0x260/0x5e0 | worker_thread+0x4d/0x3e0 | kthread+0x108/0x140 | ? process_one_work+0x5e0/0x5e0 | ? kthread_park+0x90/0x90 | ret_from_fork+0x3a/0x50 Fixes: 7e81a3530206 ("bpf: Sockmap, ensure sock lock held during tear down") Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200206111652.694507-2-jakub@cloudflare.com
2020-02-06 19:16:50 +08:00
rcu_read_unlock();
bpf: Sockmap, ensure sock lock held during tear down The sock_map_free() and sock_hash_free() paths used to delete sockmap and sockhash maps walk the maps and destroy psock and bpf state associated with the socks in the map. When done the socks no longer have BPF programs attached and will function normally. This can happen while the socks in the map are still "live" meaning data may be sent/received during the walk. Currently, though we don't take the sock_lock when the psock and bpf state is removed through this path. Specifically, this means we can be writing into the ops structure pointers such as sendmsg, sendpage, recvmsg, etc. while they are also being called from the networking side. This is not safe, we never used proper READ_ONCE/WRITE_ONCE semantics here if we believed it was safe. Further its not clear to me its even a good idea to try and do this on "live" sockets while networking side might also be using the socket. Instead of trying to reason about using the socks from both sides lets realize that every use case I'm aware of rarely deletes maps, in fact kubernetes/Cilium case builds map at init and never tears it down except on errors. So lets do the simple fix and grab sock lock. This patch wraps sock deletes from maps in sock lock and adds some annotations so we catch any other cases easier. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Song Liu <songliubraving@fb.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/bpf/20200111061206.8028-3-john.fastabend@gmail.com
2020-01-11 14:12:00 +08:00
release_sock(sk);
}
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
}
raw_spin_unlock_bh(&stab->lock);
bpf, sockhash: Synchronize_rcu before free'ing map We need to have a synchronize_rcu before free'ing the sockhash because any outstanding psock references will have a pointer to the map and when they use it, this could trigger a use after free. This is a sister fix for sockhash, following commit 2bb90e5cc90e ("bpf: sockmap, synchronize_rcu before free'ing map") which addressed sockmap, which comes from a manual audit. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200206111652.694507-3-jakub@cloudflare.com
2020-02-06 19:16:51 +08:00
/* wait for psock readers accessing its map link */
bpf: sockmap, synchronize_rcu before free'ing map We need to have a synchronize_rcu before free'ing the sockmap because any outstanding psock references will have a pointer to the map and when they use this could trigger a use after free. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-07-20 01:29:20 +08:00
synchronize_rcu();
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
bpf_map_area_free(stab->sks);
kfree(stab);
}
static void sock_map_release_progs(struct bpf_map *map)
{
psock_progs_drop(&container_of(map, struct bpf_stab, map)->progs);
}
static struct sock *__sock_map_lookup_elem(struct bpf_map *map, u32 key)
{
struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
WARN_ON_ONCE(!rcu_read_lock_held());
if (unlikely(key >= map->max_entries))
return NULL;
return READ_ONCE(stab->sks[key]);
}
static void *sock_map_lookup(struct bpf_map *map, void *key)
{
return ERR_PTR(-EOPNOTSUPP);
}
static int __sock_map_delete(struct bpf_stab *stab, struct sock *sk_test,
struct sock **psk)
{
struct sock *sk;
bpf: sockmap, sock_map_delete needs to use xchg __sock_map_delete() may be called from a tcp event such as unhash or close from the following trace, tcp_bpf_close() tcp_bpf_remove() sk_psock_unlink() sock_map_delete_from_link() __sock_map_delete() In this case the sock lock is held but this only protects against duplicate removals on the TCP side. If the map is free'd then we have this trace, sock_map_free xchg() <- replaces map entry sock_map_unref() sk_psock_put() sock_map_del_link() The __sock_map_delete() call however uses a read, test, null over the map entry which can result in both paths trying to free the map entry. To fix use xchg in TCP paths as well so we avoid having two references to the same map entry. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-07-20 01:29:19 +08:00
int err = 0;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
raw_spin_lock_bh(&stab->lock);
sk = *psk;
if (!sk_test || sk_test == sk)
bpf: sockmap, sock_map_delete needs to use xchg __sock_map_delete() may be called from a tcp event such as unhash or close from the following trace, tcp_bpf_close() tcp_bpf_remove() sk_psock_unlink() sock_map_delete_from_link() __sock_map_delete() In this case the sock lock is held but this only protects against duplicate removals on the TCP side. If the map is free'd then we have this trace, sock_map_free xchg() <- replaces map entry sock_map_unref() sk_psock_put() sock_map_del_link() The __sock_map_delete() call however uses a read, test, null over the map entry which can result in both paths trying to free the map entry. To fix use xchg in TCP paths as well so we avoid having two references to the same map entry. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-07-20 01:29:19 +08:00
sk = xchg(psk, NULL);
if (likely(sk))
sock_map_unref(sk, psk);
else
err = -EINVAL;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
raw_spin_unlock_bh(&stab->lock);
bpf: sockmap, sock_map_delete needs to use xchg __sock_map_delete() may be called from a tcp event such as unhash or close from the following trace, tcp_bpf_close() tcp_bpf_remove() sk_psock_unlink() sock_map_delete_from_link() __sock_map_delete() In this case the sock lock is held but this only protects against duplicate removals on the TCP side. If the map is free'd then we have this trace, sock_map_free xchg() <- replaces map entry sock_map_unref() sk_psock_put() sock_map_del_link() The __sock_map_delete() call however uses a read, test, null over the map entry which can result in both paths trying to free the map entry. To fix use xchg in TCP paths as well so we avoid having two references to the same map entry. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-07-20 01:29:19 +08:00
return err;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
}
static void sock_map_delete_from_link(struct bpf_map *map, struct sock *sk,
void *link_raw)
{
struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
__sock_map_delete(stab, sk, link_raw);
}
static int sock_map_delete_elem(struct bpf_map *map, void *key)
{
struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
u32 i = *(u32 *)key;
struct sock **psk;
if (unlikely(i >= map->max_entries))
return -EINVAL;
psk = &stab->sks[i];
return __sock_map_delete(stab, NULL, psk);
}
static int sock_map_get_next_key(struct bpf_map *map, void *key, void *next)
{
struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
u32 i = key ? *(u32 *)key : U32_MAX;
u32 *key_next = next;
if (i == stab->map.max_entries - 1)
return -ENOENT;
if (i >= stab->map.max_entries)
*key_next = 0;
else
*key_next = i + 1;
return 0;
}
bpf, sockmap: Don't set up upcalls and progs for listening sockets Now that sockmap/sockhash can hold listening sockets, when setting up the psock we will (i) grab references to verdict/parser progs, and (2) override socket upcalls sk_data_ready and sk_write_space. However, since we cannot redirect to listening sockets so we don't need to link the socket to the BPF progs. And more importantly we don't want the listening socket to have overridden upcalls because they would get inherited by child sockets cloned from it. Introduce a separate initialization path for listening sockets that does not change the upcalls and ignores the BPF progs. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-6-jakub@cloudflare.com
2020-02-19 01:10:17 +08:00
static bool sock_map_redirect_allowed(const struct sock *sk)
{
return sk->sk_state != TCP_LISTEN;
}
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
static int sock_map_update_common(struct bpf_map *map, u32 idx,
struct sock *sk, u64 flags)
{
struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
bpf: sockmap, only create entry if ulp is not already enabled Sockmap does not currently support adding sockets after TLS has been enabled. There never was a real use case for this so it was never added. But, we lost the test for ULP at some point so add it here and fail the socket insert if TLS is enabled. Future work could make sockmap support this use case but fixup the bug here. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-07-20 01:29:21 +08:00
struct inet_connection_sock *icsk = inet_csk(sk);
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
struct sk_psock_link *link;
struct sk_psock *psock;
struct sock *osk;
int ret;
WARN_ON_ONCE(!rcu_read_lock_held());
if (unlikely(flags > BPF_EXIST))
return -EINVAL;
if (unlikely(idx >= map->max_entries))
return -E2BIG;
net/tls: use RCU protection on icsk->icsk_ulp_data We need to make sure context does not get freed while diag code is interrogating it. Free struct tls_context with kfree_rcu(). We add the __rcu annotation directly in icsk, and cast it away in the datapath accessor. Presumably all ULPs will do a similar thing. Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-08-30 18:25:47 +08:00
if (unlikely(rcu_access_pointer(icsk->icsk_ulp_data)))
bpf: sockmap, only create entry if ulp is not already enabled Sockmap does not currently support adding sockets after TLS has been enabled. There never was a real use case for this so it was never added. But, we lost the test for ULP at some point so add it here and fail the socket insert if TLS is enabled. Future work could make sockmap support this use case but fixup the bug here. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-07-20 01:29:21 +08:00
return -EINVAL;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
link = sk_psock_init_link();
if (!link)
return -ENOMEM;
bpf, sockmap: Don't set up upcalls and progs for listening sockets Now that sockmap/sockhash can hold listening sockets, when setting up the psock we will (i) grab references to verdict/parser progs, and (2) override socket upcalls sk_data_ready and sk_write_space. However, since we cannot redirect to listening sockets so we don't need to link the socket to the BPF progs. And more importantly we don't want the listening socket to have overridden upcalls because they would get inherited by child sockets cloned from it. Introduce a separate initialization path for listening sockets that does not change the upcalls and ignores the BPF progs. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-6-jakub@cloudflare.com
2020-02-19 01:10:17 +08:00
/* Only sockets we can redirect into/from in BPF need to hold
* refs to parser/verdict progs and have their sk_data_ready
* and sk_write_space callbacks overridden.
*/
if (sock_map_redirect_allowed(sk))
ret = sock_map_link(map, &stab->progs, sk);
else
ret = sock_map_link_no_progs(map, sk);
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
if (ret < 0)
goto out_free;
psock = sk_psock(sk);
WARN_ON_ONCE(!psock);
raw_spin_lock_bh(&stab->lock);
osk = stab->sks[idx];
if (osk && flags == BPF_NOEXIST) {
ret = -EEXIST;
goto out_unlock;
} else if (!osk && flags == BPF_EXIST) {
ret = -ENOENT;
goto out_unlock;
}
sock_map_add_link(psock, link, map, &stab->sks[idx]);
stab->sks[idx] = sk;
if (osk)
sock_map_unref(osk, &stab->sks[idx]);
raw_spin_unlock_bh(&stab->lock);
return 0;
out_unlock:
raw_spin_unlock_bh(&stab->lock);
if (psock)
sk_psock_put(sk, psock);
out_free:
sk_psock_free_link(link);
return ret;
}
static bool sock_map_op_okay(const struct bpf_sock_ops_kern *ops)
{
return ops->op == BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB ||
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
ops->op == BPF_SOCK_OPS_ACTIVE_ESTABLISHED_CB ||
ops->op == BPF_SOCK_OPS_TCP_LISTEN_CB;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
}
static bool sock_map_sk_is_suitable(const struct sock *sk)
{
return sk->sk_type == SOCK_STREAM &&
sk->sk_protocol == IPPROTO_TCP;
}
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
static bool sock_map_sk_state_allowed(const struct sock *sk)
{
return (1 << sk->sk_state) & (TCPF_ESTABLISHED | TCPF_LISTEN);
}
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
static int sock_map_update_elem(struct bpf_map *map, void *key,
void *value, u64 flags)
{
u32 ufd = *(u32 *)value;
u32 idx = *(u32 *)key;
struct socket *sock;
struct sock *sk;
int ret;
sock = sockfd_lookup(ufd, &ret);
if (!sock)
return ret;
sk = sock->sk;
if (!sk) {
ret = -EINVAL;
goto out;
}
bpf, sockmap: Check update requirements after locking It's currently possible to insert sockets in unexpected states into a sockmap, due to a TOCTTOU when updating the map from a syscall. sock_map_update_elem checks that sk->sk_state == TCP_ESTABLISHED, locks the socket and then calls sock_map_update_common. At this point, the socket may have transitioned into another state, and the earlier assumptions don't hold anymore. Crucially, it's conceivable (though very unlikely) that a socket has become unhashed. This breaks the sockmap's assumption that it will get a callback via sk->sk_prot->unhash. Fix this by checking the (fixed) sk_type and sk_protocol without the lock, followed by a locked check of sk_state. Unfortunately it's not possible to push the check down into sock_(map|hash)_update_common, since BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB run before the socket has transitioned from TCP_SYN_RECV into TCP_ESTABLISHED. Fixes: 604326b41a6f ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: Lorenz Bauer <lmb@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com> Link: https://lore.kernel.org/bpf/20200207103713.28175-1-lmb@cloudflare.com
2020-02-07 18:37:12 +08:00
if (!sock_map_sk_is_suitable(sk)) {
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
ret = -EOPNOTSUPP;
goto out;
}
sock_map_sk_acquire(sk);
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
if (!sock_map_sk_state_allowed(sk))
bpf, sockmap: Check update requirements after locking It's currently possible to insert sockets in unexpected states into a sockmap, due to a TOCTTOU when updating the map from a syscall. sock_map_update_elem checks that sk->sk_state == TCP_ESTABLISHED, locks the socket and then calls sock_map_update_common. At this point, the socket may have transitioned into another state, and the earlier assumptions don't hold anymore. Crucially, it's conceivable (though very unlikely) that a socket has become unhashed. This breaks the sockmap's assumption that it will get a callback via sk->sk_prot->unhash. Fix this by checking the (fixed) sk_type and sk_protocol without the lock, followed by a locked check of sk_state. Unfortunately it's not possible to push the check down into sock_(map|hash)_update_common, since BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB run before the socket has transitioned from TCP_SYN_RECV into TCP_ESTABLISHED. Fixes: 604326b41a6f ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: Lorenz Bauer <lmb@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com> Link: https://lore.kernel.org/bpf/20200207103713.28175-1-lmb@cloudflare.com
2020-02-07 18:37:12 +08:00
ret = -EOPNOTSUPP;
else
ret = sock_map_update_common(map, idx, sk, flags);
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
sock_map_sk_release(sk);
out:
fput(sock->file);
return ret;
}
BPF_CALL_4(bpf_sock_map_update, struct bpf_sock_ops_kern *, sops,
struct bpf_map *, map, void *, key, u64, flags)
{
WARN_ON_ONCE(!rcu_read_lock_held());
if (likely(sock_map_sk_is_suitable(sops->sk) &&
sock_map_op_okay(sops)))
return sock_map_update_common(map, *(u32 *)key, sops->sk,
flags);
return -EOPNOTSUPP;
}
const struct bpf_func_proto bpf_sock_map_update_proto = {
.func = bpf_sock_map_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,
};
BPF_CALL_4(bpf_sk_redirect_map, struct sk_buff *, skb,
struct bpf_map *, map, u32, key, u64, flags)
{
struct tcp_skb_cb *tcb = TCP_SKB_CB(skb);
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
struct sock *sk;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
if (unlikely(flags & ~(BPF_F_INGRESS)))
return SK_DROP;
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
sk = __sock_map_lookup_elem(map, key);
if (unlikely(!sk || !sock_map_redirect_allowed(sk)))
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
return SK_DROP;
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
tcb->bpf.flags = flags;
tcb->bpf.sk_redir = sk;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
return SK_PASS;
}
const struct bpf_func_proto bpf_sk_redirect_map_proto = {
.func = bpf_sk_redirect_map,
.gpl_only = false,
.ret_type = RET_INTEGER,
.arg1_type = ARG_PTR_TO_CTX,
.arg2_type = ARG_CONST_MAP_PTR,
.arg3_type = ARG_ANYTHING,
.arg4_type = ARG_ANYTHING,
};
BPF_CALL_4(bpf_msg_redirect_map, struct sk_msg *, msg,
struct bpf_map *, map, u32, key, u64, flags)
{
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
struct sock *sk;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
if (unlikely(flags & ~(BPF_F_INGRESS)))
return SK_DROP;
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
sk = __sock_map_lookup_elem(map, key);
if (unlikely(!sk || !sock_map_redirect_allowed(sk)))
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
return SK_DROP;
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
msg->flags = flags;
msg->sk_redir = sk;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
return SK_PASS;
}
const struct bpf_func_proto bpf_msg_redirect_map_proto = {
.func = bpf_msg_redirect_map,
.gpl_only = false,
.ret_type = RET_INTEGER,
.arg1_type = ARG_PTR_TO_CTX,
.arg2_type = ARG_CONST_MAP_PTR,
.arg3_type = ARG_ANYTHING,
.arg4_type = ARG_ANYTHING,
};
const struct bpf_map_ops sock_map_ops = {
.map_alloc = sock_map_alloc,
.map_free = sock_map_free,
.map_get_next_key = sock_map_get_next_key,
.map_update_elem = sock_map_update_elem,
.map_delete_elem = sock_map_delete_elem,
.map_lookup_elem = sock_map_lookup,
.map_release_uref = sock_map_release_progs,
.map_check_btf = map_check_no_btf,
};
struct bpf_htab_elem {
struct rcu_head rcu;
u32 hash;
struct sock *sk;
struct hlist_node node;
bpf, sockmap: Replace zero-length array with flexible-array member The current codebase makes use of the zero-length array language extension to the C90 standard, but the preferred mechanism to declare variable-length types such as these ones is a flexible array member[1][2], introduced in C99: struct foo { int stuff; struct boo array[]; }; By making use of the mechanism above, we will get a compiler warning in case the flexible array does not occur last in the structure, which will help us prevent some kind of undefined behavior bugs from being inadvertently introduced[3] to the codebase from now on. Also, notice that, dynamic memory allocations won't be affected by this change: "Flexible array members have incomplete type, and so the sizeof operator may not be applied. As a quirk of the original implementation of zero-length arrays, sizeof evaluates to zero."[1] This issue was found with the help of Coccinelle. [1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html [2] https://github.com/KSPP/linux/issues/21 [3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour") Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-02-18 04:01:11 +08:00
u8 key[];
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
};
struct bpf_htab_bucket {
struct hlist_head head;
raw_spinlock_t lock;
};
struct bpf_htab {
struct bpf_map map;
struct bpf_htab_bucket *buckets;
u32 buckets_num;
u32 elem_size;
struct sk_psock_progs progs;
atomic_t count;
};
static inline u32 sock_hash_bucket_hash(const void *key, u32 len)
{
return jhash(key, len, 0);
}
static struct bpf_htab_bucket *sock_hash_select_bucket(struct bpf_htab *htab,
u32 hash)
{
return &htab->buckets[hash & (htab->buckets_num - 1)];
}
static struct bpf_htab_elem *
sock_hash_lookup_elem_raw(struct hlist_head *head, u32 hash, void *key,
u32 key_size)
{
struct bpf_htab_elem *elem;
hlist_for_each_entry_rcu(elem, head, node) {
if (elem->hash == hash &&
!memcmp(&elem->key, key, key_size))
return elem;
}
return NULL;
}
static struct sock *__sock_hash_lookup_elem(struct bpf_map *map, void *key)
{
struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
u32 key_size = map->key_size, hash;
struct bpf_htab_bucket *bucket;
struct bpf_htab_elem *elem;
WARN_ON_ONCE(!rcu_read_lock_held());
hash = sock_hash_bucket_hash(key, key_size);
bucket = sock_hash_select_bucket(htab, hash);
elem = sock_hash_lookup_elem_raw(&bucket->head, hash, key, key_size);
return elem ? elem->sk : NULL;
}
static void sock_hash_free_elem(struct bpf_htab *htab,
struct bpf_htab_elem *elem)
{
atomic_dec(&htab->count);
kfree_rcu(elem, rcu);
}
static void sock_hash_delete_from_link(struct bpf_map *map, struct sock *sk,
void *link_raw)
{
struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
struct bpf_htab_elem *elem_probe, *elem = link_raw;
struct bpf_htab_bucket *bucket;
WARN_ON_ONCE(!rcu_read_lock_held());
bucket = sock_hash_select_bucket(htab, elem->hash);
/* elem may be deleted in parallel from the map, but access here
* is okay since it's going away only after RCU grace period.
* However, we need to check whether it's still present.
*/
raw_spin_lock_bh(&bucket->lock);
elem_probe = sock_hash_lookup_elem_raw(&bucket->head, elem->hash,
elem->key, map->key_size);
if (elem_probe && elem_probe == elem) {
hlist_del_rcu(&elem->node);
sock_map_unref(elem->sk, elem);
sock_hash_free_elem(htab, elem);
}
raw_spin_unlock_bh(&bucket->lock);
}
static int sock_hash_delete_elem(struct bpf_map *map, void *key)
{
struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
u32 hash, key_size = map->key_size;
struct bpf_htab_bucket *bucket;
struct bpf_htab_elem *elem;
int ret = -ENOENT;
hash = sock_hash_bucket_hash(key, key_size);
bucket = sock_hash_select_bucket(htab, hash);
raw_spin_lock_bh(&bucket->lock);
elem = sock_hash_lookup_elem_raw(&bucket->head, hash, key, key_size);
if (elem) {
hlist_del_rcu(&elem->node);
sock_map_unref(elem->sk, elem);
sock_hash_free_elem(htab, elem);
ret = 0;
}
raw_spin_unlock_bh(&bucket->lock);
return ret;
}
static struct bpf_htab_elem *sock_hash_alloc_elem(struct bpf_htab *htab,
void *key, u32 key_size,
u32 hash, struct sock *sk,
struct bpf_htab_elem *old)
{
struct bpf_htab_elem *new;
if (atomic_inc_return(&htab->count) > htab->map.max_entries) {
if (!old) {
atomic_dec(&htab->count);
return ERR_PTR(-E2BIG);
}
}
new = kmalloc_node(htab->elem_size, GFP_ATOMIC | __GFP_NOWARN,
htab->map.numa_node);
if (!new) {
atomic_dec(&htab->count);
return ERR_PTR(-ENOMEM);
}
memcpy(new->key, key, key_size);
new->sk = sk;
new->hash = hash;
return new;
}
static int sock_hash_update_common(struct bpf_map *map, void *key,
struct sock *sk, u64 flags)
{
struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
net: sock_map, fix missing ulp check in sock hash case sock_map and ULP only work together when ULP is loaded after the sock map is loaded. In the sock_map case we added a check for this to fail the load if ULP is already set. However, we missed the check on the sock_hash side. Add a ULP check to the sock_hash update path. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Reported-by: syzbot+7a6ee4d0078eac6bf782@syzkaller.appspotmail.com Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-09-04 04:24:50 +08:00
struct inet_connection_sock *icsk = inet_csk(sk);
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
u32 key_size = map->key_size, hash;
struct bpf_htab_elem *elem, *elem_new;
struct bpf_htab_bucket *bucket;
struct sk_psock_link *link;
struct sk_psock *psock;
int ret;
WARN_ON_ONCE(!rcu_read_lock_held());
if (unlikely(flags > BPF_EXIST))
return -EINVAL;
net: sock_map, fix missing ulp check in sock hash case sock_map and ULP only work together when ULP is loaded after the sock map is loaded. In the sock_map case we added a check for this to fail the load if ULP is already set. However, we missed the check on the sock_hash side. Add a ULP check to the sock_hash update path. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Reported-by: syzbot+7a6ee4d0078eac6bf782@syzkaller.appspotmail.com Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-09-04 04:24:50 +08:00
if (unlikely(icsk->icsk_ulp_data))
return -EINVAL;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
link = sk_psock_init_link();
if (!link)
return -ENOMEM;
bpf, sockmap: Don't set up upcalls and progs for listening sockets Now that sockmap/sockhash can hold listening sockets, when setting up the psock we will (i) grab references to verdict/parser progs, and (2) override socket upcalls sk_data_ready and sk_write_space. However, since we cannot redirect to listening sockets so we don't need to link the socket to the BPF progs. And more importantly we don't want the listening socket to have overridden upcalls because they would get inherited by child sockets cloned from it. Introduce a separate initialization path for listening sockets that does not change the upcalls and ignores the BPF progs. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-6-jakub@cloudflare.com
2020-02-19 01:10:17 +08:00
/* Only sockets we can redirect into/from in BPF need to hold
* refs to parser/verdict progs and have their sk_data_ready
* and sk_write_space callbacks overridden.
*/
if (sock_map_redirect_allowed(sk))
ret = sock_map_link(map, &htab->progs, sk);
else
ret = sock_map_link_no_progs(map, sk);
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
if (ret < 0)
goto out_free;
psock = sk_psock(sk);
WARN_ON_ONCE(!psock);
hash = sock_hash_bucket_hash(key, key_size);
bucket = sock_hash_select_bucket(htab, hash);
raw_spin_lock_bh(&bucket->lock);
elem = sock_hash_lookup_elem_raw(&bucket->head, hash, key, key_size);
if (elem && flags == BPF_NOEXIST) {
ret = -EEXIST;
goto out_unlock;
} else if (!elem && flags == BPF_EXIST) {
ret = -ENOENT;
goto out_unlock;
}
elem_new = sock_hash_alloc_elem(htab, key, key_size, hash, sk, elem);
if (IS_ERR(elem_new)) {
ret = PTR_ERR(elem_new);
goto out_unlock;
}
sock_map_add_link(psock, link, map, elem_new);
/* Add new element to the head of the list, so that
* concurrent search will find it before old elem.
*/
hlist_add_head_rcu(&elem_new->node, &bucket->head);
if (elem) {
hlist_del_rcu(&elem->node);
sock_map_unref(elem->sk, elem);
sock_hash_free_elem(htab, elem);
}
raw_spin_unlock_bh(&bucket->lock);
return 0;
out_unlock:
raw_spin_unlock_bh(&bucket->lock);
sk_psock_put(sk, psock);
out_free:
sk_psock_free_link(link);
return ret;
}
static int sock_hash_update_elem(struct bpf_map *map, void *key,
void *value, u64 flags)
{
u32 ufd = *(u32 *)value;
struct socket *sock;
struct sock *sk;
int ret;
sock = sockfd_lookup(ufd, &ret);
if (!sock)
return ret;
sk = sock->sk;
if (!sk) {
ret = -EINVAL;
goto out;
}
bpf, sockmap: Check update requirements after locking It's currently possible to insert sockets in unexpected states into a sockmap, due to a TOCTTOU when updating the map from a syscall. sock_map_update_elem checks that sk->sk_state == TCP_ESTABLISHED, locks the socket and then calls sock_map_update_common. At this point, the socket may have transitioned into another state, and the earlier assumptions don't hold anymore. Crucially, it's conceivable (though very unlikely) that a socket has become unhashed. This breaks the sockmap's assumption that it will get a callback via sk->sk_prot->unhash. Fix this by checking the (fixed) sk_type and sk_protocol without the lock, followed by a locked check of sk_state. Unfortunately it's not possible to push the check down into sock_(map|hash)_update_common, since BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB run before the socket has transitioned from TCP_SYN_RECV into TCP_ESTABLISHED. Fixes: 604326b41a6f ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: Lorenz Bauer <lmb@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com> Link: https://lore.kernel.org/bpf/20200207103713.28175-1-lmb@cloudflare.com
2020-02-07 18:37:12 +08:00
if (!sock_map_sk_is_suitable(sk)) {
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
ret = -EOPNOTSUPP;
goto out;
}
sock_map_sk_acquire(sk);
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
if (!sock_map_sk_state_allowed(sk))
bpf, sockmap: Check update requirements after locking It's currently possible to insert sockets in unexpected states into a sockmap, due to a TOCTTOU when updating the map from a syscall. sock_map_update_elem checks that sk->sk_state == TCP_ESTABLISHED, locks the socket and then calls sock_map_update_common. At this point, the socket may have transitioned into another state, and the earlier assumptions don't hold anymore. Crucially, it's conceivable (though very unlikely) that a socket has become unhashed. This breaks the sockmap's assumption that it will get a callback via sk->sk_prot->unhash. Fix this by checking the (fixed) sk_type and sk_protocol without the lock, followed by a locked check of sk_state. Unfortunately it's not possible to push the check down into sock_(map|hash)_update_common, since BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB run before the socket has transitioned from TCP_SYN_RECV into TCP_ESTABLISHED. Fixes: 604326b41a6f ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: Lorenz Bauer <lmb@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com> Link: https://lore.kernel.org/bpf/20200207103713.28175-1-lmb@cloudflare.com
2020-02-07 18:37:12 +08:00
ret = -EOPNOTSUPP;
else
ret = sock_hash_update_common(map, key, sk, flags);
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
sock_map_sk_release(sk);
out:
fput(sock->file);
return ret;
}
static int sock_hash_get_next_key(struct bpf_map *map, void *key,
void *key_next)
{
struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
struct bpf_htab_elem *elem, *elem_next;
u32 hash, key_size = map->key_size;
struct hlist_head *head;
int i = 0;
if (!key)
goto find_first_elem;
hash = sock_hash_bucket_hash(key, key_size);
head = &sock_hash_select_bucket(htab, hash)->head;
elem = sock_hash_lookup_elem_raw(head, hash, key, key_size);
if (!elem)
goto find_first_elem;
elem_next = hlist_entry_safe(rcu_dereference_raw(hlist_next_rcu(&elem->node)),
struct bpf_htab_elem, node);
if (elem_next) {
memcpy(key_next, elem_next->key, key_size);
return 0;
}
i = hash & (htab->buckets_num - 1);
i++;
find_first_elem:
for (; i < htab->buckets_num; i++) {
head = &sock_hash_select_bucket(htab, i)->head;
elem_next = hlist_entry_safe(rcu_dereference_raw(hlist_first_rcu(head)),
struct bpf_htab_elem, node);
if (elem_next) {
memcpy(key_next, elem_next->key, key_size);
return 0;
}
}
return -ENOENT;
}
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);
if (attr->max_entries == 0 ||
attr->key_size == 0 ||
attr->value_size != 4 ||
attr->map_flags & ~SOCK_CREATE_FLAG_MASK)
return ERR_PTR(-EINVAL);
if (attr->key_size > MAX_BPF_STACK)
return ERR_PTR(-E2BIG);
htab = kzalloc(sizeof(*htab), GFP_USER);
if (!htab)
return ERR_PTR(-ENOMEM);
bpf_map_init_from_attr(&htab->map, attr);
htab->buckets_num = roundup_pow_of_two(htab->map.max_entries);
htab->elem_size = sizeof(struct bpf_htab_elem) +
round_up(htab->map.key_size, 8);
if (htab->buckets_num == 0 ||
htab->buckets_num > U32_MAX / sizeof(struct bpf_htab_bucket)) {
err = -EINVAL;
goto free_htab;
}
cost = (u64) htab->buckets_num * sizeof(struct bpf_htab_bucket) +
(u64) htab->elem_size * htab->map.max_entries;
if (cost >= U32_MAX - PAGE_SIZE) {
err = -EINVAL;
goto free_htab;
}
htab->buckets = bpf_map_area_alloc(htab->buckets_num *
sizeof(struct bpf_htab_bucket),
htab->map.numa_node);
if (!htab->buckets) {
err = -ENOMEM;
goto free_htab;
}
for (i = 0; i < htab->buckets_num; 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 void sock_hash_free(struct bpf_map *map)
{
struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
struct bpf_htab_bucket *bucket;
struct bpf_htab_elem *elem;
struct hlist_node *node;
int i;
synchronize_rcu();
for (i = 0; i < htab->buckets_num; i++) {
bucket = sock_hash_select_bucket(htab, i);
raw_spin_lock_bh(&bucket->lock);
hlist_for_each_entry_safe(elem, node, &bucket->head, node) {
hlist_del_rcu(&elem->node);
bpf: Sockmap, ensure sock lock held during tear down The sock_map_free() and sock_hash_free() paths used to delete sockmap and sockhash maps walk the maps and destroy psock and bpf state associated with the socks in the map. When done the socks no longer have BPF programs attached and will function normally. This can happen while the socks in the map are still "live" meaning data may be sent/received during the walk. Currently, though we don't take the sock_lock when the psock and bpf state is removed through this path. Specifically, this means we can be writing into the ops structure pointers such as sendmsg, sendpage, recvmsg, etc. while they are also being called from the networking side. This is not safe, we never used proper READ_ONCE/WRITE_ONCE semantics here if we believed it was safe. Further its not clear to me its even a good idea to try and do this on "live" sockets while networking side might also be using the socket. Instead of trying to reason about using the socks from both sides lets realize that every use case I'm aware of rarely deletes maps, in fact kubernetes/Cilium case builds map at init and never tears it down except on errors. So lets do the simple fix and grab sock lock. This patch wraps sock deletes from maps in sock lock and adds some annotations so we catch any other cases easier. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Song Liu <songliubraving@fb.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/bpf/20200111061206.8028-3-john.fastabend@gmail.com
2020-01-11 14:12:00 +08:00
lock_sock(elem->sk);
bpf, sockmap: Don't sleep while holding RCU lock on tear-down rcu_read_lock is needed to protect access to psock inside sock_map_unref when tearing down the map. However, we can't afford to sleep in lock_sock while in RCU read-side critical section. Grab the RCU lock only after we have locked the socket. This fixes RCU warnings triggerable on a VM with 1 vCPU when free'ing a sockmap/sockhash that contains at least one socket: | ============================= | WARNING: suspicious RCU usage | 5.5.0-04005-g8fc91b972b73 #450 Not tainted | ----------------------------- | include/linux/rcupdate.h:272 Illegal context switch in RCU read-side critical section! | | other info that might help us debug this: | | | rcu_scheduler_active = 2, debug_locks = 1 | 4 locks held by kworker/0:1/62: | #0: ffff88813b019748 ((wq_completion)events){+.+.}, at: process_one_work+0x1d7/0x5e0 | #1: ffffc900000abe50 ((work_completion)(&map->work)){+.+.}, at: process_one_work+0x1d7/0x5e0 | #2: ffffffff82065d20 (rcu_read_lock){....}, at: sock_map_free+0x5/0x170 | #3: ffff8881368c5df8 (&stab->lock){+...}, at: sock_map_free+0x64/0x170 | | stack backtrace: | CPU: 0 PID: 62 Comm: kworker/0:1 Not tainted 5.5.0-04005-g8fc91b972b73 #450 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_073836-buildvm-ppc64le-16.ppc.fedoraproject.org-3.fc31 04/01/2014 | Workqueue: events bpf_map_free_deferred | Call Trace: | dump_stack+0x71/0xa0 | ___might_sleep+0x105/0x190 | lock_sock_nested+0x28/0x90 | sock_map_free+0x95/0x170 | bpf_map_free_deferred+0x58/0x80 | process_one_work+0x260/0x5e0 | worker_thread+0x4d/0x3e0 | kthread+0x108/0x140 | ? process_one_work+0x5e0/0x5e0 | ? kthread_park+0x90/0x90 | ret_from_fork+0x3a/0x50 | ============================= | WARNING: suspicious RCU usage | 5.5.0-04005-g8fc91b972b73-dirty #452 Not tainted | ----------------------------- | include/linux/rcupdate.h:272 Illegal context switch in RCU read-side critical section! | | other info that might help us debug this: | | | rcu_scheduler_active = 2, debug_locks = 1 | 4 locks held by kworker/0:1/62: | #0: ffff88813b019748 ((wq_completion)events){+.+.}, at: process_one_work+0x1d7/0x5e0 | #1: ffffc900000abe50 ((work_completion)(&map->work)){+.+.}, at: process_one_work+0x1d7/0x5e0 | #2: ffffffff82065d20 (rcu_read_lock){....}, at: sock_hash_free+0x5/0x1d0 | #3: ffff888139966e00 (&htab->buckets[i].lock){+...}, at: sock_hash_free+0x92/0x1d0 | | stack backtrace: | CPU: 0 PID: 62 Comm: kworker/0:1 Not tainted 5.5.0-04005-g8fc91b972b73-dirty #452 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_073836-buildvm-ppc64le-16.ppc.fedoraproject.org-3.fc31 04/01/2014 | Workqueue: events bpf_map_free_deferred | Call Trace: | dump_stack+0x71/0xa0 | ___might_sleep+0x105/0x190 | lock_sock_nested+0x28/0x90 | sock_hash_free+0xec/0x1d0 | bpf_map_free_deferred+0x58/0x80 | process_one_work+0x260/0x5e0 | worker_thread+0x4d/0x3e0 | kthread+0x108/0x140 | ? process_one_work+0x5e0/0x5e0 | ? kthread_park+0x90/0x90 | ret_from_fork+0x3a/0x50 Fixes: 7e81a3530206 ("bpf: Sockmap, ensure sock lock held during tear down") Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200206111652.694507-2-jakub@cloudflare.com
2020-02-06 19:16:50 +08:00
rcu_read_lock();
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
sock_map_unref(elem->sk, elem);
bpf, sockmap: Don't sleep while holding RCU lock on tear-down rcu_read_lock is needed to protect access to psock inside sock_map_unref when tearing down the map. However, we can't afford to sleep in lock_sock while in RCU read-side critical section. Grab the RCU lock only after we have locked the socket. This fixes RCU warnings triggerable on a VM with 1 vCPU when free'ing a sockmap/sockhash that contains at least one socket: | ============================= | WARNING: suspicious RCU usage | 5.5.0-04005-g8fc91b972b73 #450 Not tainted | ----------------------------- | include/linux/rcupdate.h:272 Illegal context switch in RCU read-side critical section! | | other info that might help us debug this: | | | rcu_scheduler_active = 2, debug_locks = 1 | 4 locks held by kworker/0:1/62: | #0: ffff88813b019748 ((wq_completion)events){+.+.}, at: process_one_work+0x1d7/0x5e0 | #1: ffffc900000abe50 ((work_completion)(&map->work)){+.+.}, at: process_one_work+0x1d7/0x5e0 | #2: ffffffff82065d20 (rcu_read_lock){....}, at: sock_map_free+0x5/0x170 | #3: ffff8881368c5df8 (&stab->lock){+...}, at: sock_map_free+0x64/0x170 | | stack backtrace: | CPU: 0 PID: 62 Comm: kworker/0:1 Not tainted 5.5.0-04005-g8fc91b972b73 #450 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_073836-buildvm-ppc64le-16.ppc.fedoraproject.org-3.fc31 04/01/2014 | Workqueue: events bpf_map_free_deferred | Call Trace: | dump_stack+0x71/0xa0 | ___might_sleep+0x105/0x190 | lock_sock_nested+0x28/0x90 | sock_map_free+0x95/0x170 | bpf_map_free_deferred+0x58/0x80 | process_one_work+0x260/0x5e0 | worker_thread+0x4d/0x3e0 | kthread+0x108/0x140 | ? process_one_work+0x5e0/0x5e0 | ? kthread_park+0x90/0x90 | ret_from_fork+0x3a/0x50 | ============================= | WARNING: suspicious RCU usage | 5.5.0-04005-g8fc91b972b73-dirty #452 Not tainted | ----------------------------- | include/linux/rcupdate.h:272 Illegal context switch in RCU read-side critical section! | | other info that might help us debug this: | | | rcu_scheduler_active = 2, debug_locks = 1 | 4 locks held by kworker/0:1/62: | #0: ffff88813b019748 ((wq_completion)events){+.+.}, at: process_one_work+0x1d7/0x5e0 | #1: ffffc900000abe50 ((work_completion)(&map->work)){+.+.}, at: process_one_work+0x1d7/0x5e0 | #2: ffffffff82065d20 (rcu_read_lock){....}, at: sock_hash_free+0x5/0x1d0 | #3: ffff888139966e00 (&htab->buckets[i].lock){+...}, at: sock_hash_free+0x92/0x1d0 | | stack backtrace: | CPU: 0 PID: 62 Comm: kworker/0:1 Not tainted 5.5.0-04005-g8fc91b972b73-dirty #452 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_073836-buildvm-ppc64le-16.ppc.fedoraproject.org-3.fc31 04/01/2014 | Workqueue: events bpf_map_free_deferred | Call Trace: | dump_stack+0x71/0xa0 | ___might_sleep+0x105/0x190 | lock_sock_nested+0x28/0x90 | sock_hash_free+0xec/0x1d0 | bpf_map_free_deferred+0x58/0x80 | process_one_work+0x260/0x5e0 | worker_thread+0x4d/0x3e0 | kthread+0x108/0x140 | ? process_one_work+0x5e0/0x5e0 | ? kthread_park+0x90/0x90 | ret_from_fork+0x3a/0x50 Fixes: 7e81a3530206 ("bpf: Sockmap, ensure sock lock held during tear down") Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200206111652.694507-2-jakub@cloudflare.com
2020-02-06 19:16:50 +08:00
rcu_read_unlock();
bpf: Sockmap, ensure sock lock held during tear down The sock_map_free() and sock_hash_free() paths used to delete sockmap and sockhash maps walk the maps and destroy psock and bpf state associated with the socks in the map. When done the socks no longer have BPF programs attached and will function normally. This can happen while the socks in the map are still "live" meaning data may be sent/received during the walk. Currently, though we don't take the sock_lock when the psock and bpf state is removed through this path. Specifically, this means we can be writing into the ops structure pointers such as sendmsg, sendpage, recvmsg, etc. while they are also being called from the networking side. This is not safe, we never used proper READ_ONCE/WRITE_ONCE semantics here if we believed it was safe. Further its not clear to me its even a good idea to try and do this on "live" sockets while networking side might also be using the socket. Instead of trying to reason about using the socks from both sides lets realize that every use case I'm aware of rarely deletes maps, in fact kubernetes/Cilium case builds map at init and never tears it down except on errors. So lets do the simple fix and grab sock lock. This patch wraps sock deletes from maps in sock lock and adds some annotations so we catch any other cases easier. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Song Liu <songliubraving@fb.com> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/bpf/20200111061206.8028-3-john.fastabend@gmail.com
2020-01-11 14:12:00 +08:00
release_sock(elem->sk);
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
}
raw_spin_unlock_bh(&bucket->lock);
}
bpf, sockhash: Synchronize_rcu before free'ing map We need to have a synchronize_rcu before free'ing the sockhash because any outstanding psock references will have a pointer to the map and when they use it, this could trigger a use after free. This is a sister fix for sockhash, following commit 2bb90e5cc90e ("bpf: sockmap, synchronize_rcu before free'ing map") which addressed sockmap, which comes from a manual audit. Fixes: 604326b41a6fb ("bpf, sockmap: convert to generic sk_msg interface") Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200206111652.694507-3-jakub@cloudflare.com
2020-02-06 19:16:51 +08:00
/* wait for psock readers accessing its map link */
synchronize_rcu();
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
bpf_map_area_free(htab->buckets);
kfree(htab);
}
static void sock_hash_release_progs(struct bpf_map *map)
{
psock_progs_drop(&container_of(map, struct bpf_htab, map)->progs);
}
BPF_CALL_4(bpf_sock_hash_update, struct bpf_sock_ops_kern *, sops,
struct bpf_map *, map, void *, key, u64, flags)
{
WARN_ON_ONCE(!rcu_read_lock_held());
if (likely(sock_map_sk_is_suitable(sops->sk) &&
sock_map_op_okay(sops)))
return sock_hash_update_common(map, key, sops->sk, flags);
return -EOPNOTSUPP;
}
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,
};
BPF_CALL_4(bpf_sk_redirect_hash, struct sk_buff *, skb,
struct bpf_map *, map, void *, key, u64, flags)
{
struct tcp_skb_cb *tcb = TCP_SKB_CB(skb);
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
struct sock *sk;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
if (unlikely(flags & ~(BPF_F_INGRESS)))
return SK_DROP;
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
sk = __sock_hash_lookup_elem(map, key);
if (unlikely(!sk || !sock_map_redirect_allowed(sk)))
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
return SK_DROP;
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
tcb->bpf.flags = flags;
tcb->bpf.sk_redir = sk;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
return SK_PASS;
}
const struct bpf_func_proto bpf_sk_redirect_hash_proto = {
.func = bpf_sk_redirect_hash,
.gpl_only = false,
.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,
};
BPF_CALL_4(bpf_msg_redirect_hash, struct sk_msg *, msg,
struct bpf_map *, map, void *, key, u64, flags)
{
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
struct sock *sk;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
if (unlikely(flags & ~(BPF_F_INGRESS)))
return SK_DROP;
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
sk = __sock_hash_lookup_elem(map, key);
if (unlikely(!sk || !sock_map_redirect_allowed(sk)))
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
return SK_DROP;
bpf, sockmap: Allow inserting listening TCP sockets into sockmap In order for sockmap/sockhash types to become generic collections for storing TCP sockets we need to loosen the checks during map update, while tightening the checks in redirect helpers. Currently sock{map,hash} require the TCP socket to be in established state, which prevents inserting listening sockets. Change the update pre-checks so the socket can also be in listening state. Since it doesn't make sense to redirect with sock{map,hash} to listening sockets, add appropriate socket state checks to BPF redirect helpers too. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/bpf/20200218171023.844439-5-jakub@cloudflare.com
2020-02-19 01:10:16 +08:00
msg->flags = flags;
msg->sk_redir = sk;
bpf, sockmap: convert to generic sk_msg interface Add a generic sk_msg layer, and convert current sockmap and later kTLS over to make use of it. While sk_buff handles network packet representation from netdevice up to socket, sk_msg handles data representation from application to socket layer. This means that sk_msg framework spans across ULP users in the kernel, and enables features such as introspection or filtering of data with the help of BPF programs that operate on this data structure. Latter becomes in particular useful for kTLS where data encryption is deferred into the kernel, and as such enabling the kernel to perform L7 introspection and policy based on BPF for TLS connections where the record is being encrypted after BPF has run and came to a verdict. In order to get there, first step is to transform open coding of scatter-gather list handling into a common core framework that subsystems can use. The code itself has been split and refactored into three bigger pieces: i) the generic sk_msg API which deals with managing the scatter gather ring, providing helpers for walking and mangling, transferring application data from user space into it, and preparing it for BPF pre/post-processing, ii) the plain sock map itself where sockets can be attached to or detached from; these bits are independent of i) which can now be used also without sock map, and iii) the integration with plain TCP as one protocol to be used for processing L7 application data (later this could e.g. also be extended to other protocols like UDP). The semantics are the same with the old sock map code and therefore no change of user facing behavior or APIs. While pursuing this work it also helped finding a number of bugs in the old sockmap code that we've fixed already in earlier commits. The test_sockmap kselftest suite passes through fine as well. Joint work with John. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-10-13 08:45:58 +08:00
return SK_PASS;
}
const struct bpf_func_proto bpf_msg_redirect_hash_proto = {
.func = bpf_msg_redirect_hash,
.gpl_only = false,
.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,
};
const struct bpf_map_ops sock_hash_ops = {
.map_alloc = sock_hash_alloc,
.map_free = sock_hash_free,
.map_get_next_key = sock_hash_get_next_key,
.map_update_elem = sock_hash_update_elem,
.map_delete_elem = sock_hash_delete_elem,
.map_lookup_elem = sock_map_lookup,
.map_release_uref = sock_hash_release_progs,
.map_check_btf = map_check_no_btf,
};
static struct sk_psock_progs *sock_map_progs(struct bpf_map *map)
{
switch (map->map_type) {
case BPF_MAP_TYPE_SOCKMAP:
return &container_of(map, struct bpf_stab, map)->progs;
case BPF_MAP_TYPE_SOCKHASH:
return &container_of(map, struct bpf_htab, map)->progs;
default:
break;
}
return NULL;
}
int sock_map_prog_update(struct bpf_map *map, struct bpf_prog *prog,
u32 which)
{
struct sk_psock_progs *progs = sock_map_progs(map);
if (!progs)
return -EOPNOTSUPP;
switch (which) {
case BPF_SK_MSG_VERDICT:
psock_set_prog(&progs->msg_parser, prog);
break;
case BPF_SK_SKB_STREAM_PARSER:
psock_set_prog(&progs->skb_parser, prog);
break;
case BPF_SK_SKB_STREAM_VERDICT:
psock_set_prog(&progs->skb_verdict, prog);
break;
default:
return -EOPNOTSUPP;
}
return 0;
}
void sk_psock_unlink(struct sock *sk, struct sk_psock_link *link)
{
switch (link->map->map_type) {
case BPF_MAP_TYPE_SOCKMAP:
return sock_map_delete_from_link(link->map, sk,
link->link_raw);
case BPF_MAP_TYPE_SOCKHASH:
return sock_hash_delete_from_link(link->map, sk,
link->link_raw);
default:
break;
}
}
Copy Permalink