linux.git/net/core/skmsg.c, branch v6.16

bpf, sockmap: Avoid using sk_socket after free when sending

2025-05-22T23:16:37+00:00

The sk->sk_socket is not locked or referenced in backlog thread, and
during the call to skb_send_sock(), there is a race condition with
the release of sk_socket. All types of sockets(tcp/udp/unix/vsock)
will be affected.

Race conditions:
'''
CPU0                               CPU1

backlog::skb_send_sock
  sendmsg_unlocked
    sock_sendmsg
      sock_sendmsg_nosec
                                   close(fd):
                                     ...
                                     ops->release() -> sock_map_close()
                                     sk_socket->ops = NULL
                                     free(socket)
      sock->ops->sendmsg
            ^
            panic here
'''

The ref of psock become 0 after sock_map_close() executed.
'''
void sock_map_close()
{
    ...
    if (likely(psock)) {
    ...
    // !! here we remove psock and the ref of psock become 0
    sock_map_remove_links(sk, psock)
    psock = sk_psock_get(sk);
    if (unlikely(!psock))
        goto no_psock; <=== Control jumps here via goto
        ...
        cancel_delayed_work_sync(&psock->work); <=== not executed
        sk_psock_put(sk, psock);
        ...
}
'''

Based on the fact that we already wait for the workqueue to finish in
sock_map_close() if psock is held, we simply increase the psock
reference count to avoid race conditions.

With this patch, if the backlog thread is running, sock_map_close() will
wait for the backlog thread to complete and cancel all pending work.

If no backlog running, any pending work that hasn't started by then will
fail when invoked by sk_psock_get(), as the psock reference count have
been zeroed, and sk_psock_drop() will cancel all jobs via
cancel_delayed_work_sync().

In summary, we require synchronization to coordinate the backlog thread
and close() thread.

The panic I catched:
'''
Workqueue: events sk_psock_backlog
RIP: 0010:sock_sendmsg+0x21d/0x440
RAX: 0000000000000000 RBX: ffffc9000521fad8 RCX: 0000000000000001
...
Call Trace:
 
 ? die_addr+0x40/0xa0
 ? exc_general_protection+0x14c/0x230
 ? asm_exc_general_protection+0x26/0x30
 ? sock_sendmsg+0x21d/0x440
 ? sock_sendmsg+0x3e0/0x440
 ? __pfx_sock_sendmsg+0x10/0x10
 __skb_send_sock+0x543/0xb70
 sk_psock_backlog+0x247/0xb80
...
'''

Fixes: 4b4647add7d3 ("sock_map: avoid race between sock_map_close and sk_psock_put")
Reported-by: Michal Luczaj 
Signed-off-by: Jiayuan Chen 
Signed-off-by: Martin KaFai Lau 
Reviewed-by: John Fastabend 
Link: https://lore.kernel.org/r/20250516141713.291150-1-jiayuan.chen@linux.dev

bpf, sockmap: Fix panic when calling skb_linearize

2025-04-10T02:59:00+00:00

The panic can be reproduced by executing the command:
./bench sockmap -c 2 -p 1 -a --rx-verdict-ingress --rx-strp 100000

Then a kernel panic was captured:
'''
[  657.460555] kernel BUG at net/core/skbuff.c:2178!
[  657.462680] Tainted: [W]=WARN
[  657.463287] Workqueue: events sk_psock_backlog
...
[  657.469610]  
[  657.469738]  ? die+0x36/0x90
[  657.469916]  ? do_trap+0x1d0/0x270
[  657.470118]  ? pskb_expand_head+0x612/0xf40
[  657.470376]  ? pskb_expand_head+0x612/0xf40
[  657.470620]  ? do_error_trap+0xa3/0x170
[  657.470846]  ? pskb_expand_head+0x612/0xf40
[  657.471092]  ? handle_invalid_op+0x2c/0x40
[  657.471335]  ? pskb_expand_head+0x612/0xf40
[  657.471579]  ? exc_invalid_op+0x2d/0x40
[  657.471805]  ? asm_exc_invalid_op+0x1a/0x20
[  657.472052]  ? pskb_expand_head+0xd1/0xf40
[  657.472292]  ? pskb_expand_head+0x612/0xf40
[  657.472540]  ? lock_acquire+0x18f/0x4e0
[  657.472766]  ? find_held_lock+0x2d/0x110
[  657.472999]  ? __pfx_pskb_expand_head+0x10/0x10
[  657.473263]  ? __kmalloc_cache_noprof+0x5b/0x470
[  657.473537]  ? __pfx___lock_release.isra.0+0x10/0x10
[  657.473826]  __pskb_pull_tail+0xfd/0x1d20
[  657.474062]  ? __kasan_slab_alloc+0x4e/0x90
[  657.474707]  sk_psock_skb_ingress_enqueue+0x3bf/0x510
[  657.475392]  ? __kasan_kmalloc+0xaa/0xb0
[  657.476010]  sk_psock_backlog+0x5cf/0xd70
[  657.476637]  process_one_work+0x858/0x1a20
'''

The panic originates from the assertion BUG_ON(skb_shared(skb)) in
skb_linearize(). A previous commit(see Fixes tag) introduced skb_get()
to avoid race conditions between skb operations in the backlog and skb
release in the recvmsg path. However, this caused the panic to always
occur when skb_linearize is executed.

The "--rx-strp 100000" parameter forces the RX path to use the strparser
module which aggregates data until it reaches 100KB before calling sockmap
logic. The 100KB payload exceeds MAX_MSG_FRAGS, triggering skb_linearize.

To fix this issue, just move skb_get into sk_psock_skb_ingress_enqueue.

'''
sk_psock_backlog:
    sk_psock_handle_skb
       skb_get(skb) <== we move it into 'sk_psock_skb_ingress_enqueue'
       sk_psock_skb_ingress____________
                                       ↓
                                       |
                                       | → sk_psock_skb_ingress_self
                                       |      sk_psock_skb_ingress_enqueue
sk_psock_verdict_apply_________________↑          skb_linearize
'''

Note that for verdict_apply path, the skb_get operation is unnecessary so
we add 'take_ref' param to control it's behavior.

Fixes: a454d84ee20b ("bpf, sockmap: Fix skb refcnt race after locking changes")
Signed-off-by: Jiayuan Chen 
Link: https://lore.kernel.org/r/20250407142234.47591-4-jiayuan.chen@linux.dev
Signed-off-by: Alexei Starovoitov

bpf, sockmap: fix duplicated data transmission

2025-04-10T02:58:59+00:00

In the !ingress path under sk_psock_handle_skb(), when sending data to the
remote under snd_buf limitations, partial skb data might be transmitted.

Although we preserved the partial transmission state (offset/length), the
state wasn't properly consumed during retries. This caused the retry path
to resend the entire skb data instead of continuing from the previous
offset, resulting in data overlap at the receiver side.

Fixes: 405df89dd52c ("bpf, sockmap: Improved check for empty queue")
Signed-off-by: Jiayuan Chen 
Link: https://lore.kernel.org/r/20250407142234.47591-3-jiayuan.chen@linux.dev
Signed-off-by: Alexei Starovoitov

bpf, sockmap: Fix data lost during EAGAIN retries

2025-04-10T02:58:59+00:00

We call skb_bpf_redirect_clear() to clean _sk_redir before handling skb in
backlog, but when sk_psock_handle_skb() return EAGAIN due to sk_rcvbuf
limit, the redirect info in _sk_redir is not recovered.

Fix skb redir loss during EAGAIN retries by restoring _sk_redir
information using skb_bpf_set_redir().

Before this patch:
'''
./bench sockmap -c 2 -p 1 -a --rx-verdict-ingress
Setting up benchmark 'sockmap'...
create socket fd c1:13 p1:14 c2:15 p2:16
Benchmark 'sockmap' started.
Send Speed 1343.172 MB/s, BPF Speed 1343.238 MB/s, Rcv Speed   65.271 MB/s
Send Speed 1352.022 MB/s, BPF Speed 1352.088 MB/s, Rcv Speed   0 MB/s
Send Speed 1354.105 MB/s, BPF Speed 1354.105 MB/s, Rcv Speed   0 MB/s
Send Speed 1355.018 MB/s, BPF Speed 1354.887 MB/s, Rcv Speed   0 MB/s
'''
Due to the high send rate, the RX processing path may frequently hit the
sk_rcvbuf limit. Once triggered, incorrect _sk_redir will cause the flow
to mistakenly enter the "!ingress" path, leading to send failures.
(The Rcv speed depends on tcp_rmem).

After this patch:
'''
./bench sockmap -c 2 -p 1 -a --rx-verdict-ingress
Setting up benchmark 'sockmap'...
create socket fd c1:13 p1:14 c2:15 p2:16
Benchmark 'sockmap' started.
Send Speed 1347.236 MB/s, BPF Speed 1347.367 MB/s, Rcv Speed   65.402 MB/s
Send Speed 1353.320 MB/s, BPF Speed 1353.320 MB/s, Rcv Speed   65.536 MB/s
Send Speed 1353.186 MB/s, BPF Speed 1353.121 MB/s, Rcv Speed   65.536 MB/s
'''

Signed-off-by: Jiayuan Chen 
Link: https://lore.kernel.org/r/20250407142234.47591-2-jiayuan.chen@linux.dev
Signed-off-by: Alexei Starovoitov

bpf: Fix wrong copied_seq calculation

2025-01-29T21:32:23+00:00

'sk->copied_seq' was updated in the tcp_eat_skb() function when the action
of a BPF program was SK_REDIRECT. For other actions, like SK_PASS, the
update logic for 'sk->copied_seq' was moved to tcp_bpf_recvmsg_parser()
to ensure the accuracy of the 'fionread' feature.

It works for a single stream_verdict scenario, as it also modified
sk_data_ready->sk_psock_verdict_data_ready->tcp_read_skb
to remove updating 'sk->copied_seq'.

However, for programs where both stream_parser and stream_verdict are
active (strparser purpose), tcp_read_sock() was used instead of
tcp_read_skb() (sk_data_ready->strp_data_ready->tcp_read_sock).
tcp_read_sock() now still updates 'sk->copied_seq', leading to duplicate
updates.

In summary, for strparser + SK_PASS, copied_seq is redundantly calculated
in both tcp_read_sock() and tcp_bpf_recvmsg_parser().

The issue causes incorrect copied_seq calculations, which prevent
correct data reads from the recv() interface in user-land.

We do not want to add new proto_ops to implement a new version of
tcp_read_sock, as this would introduce code complexity [1].

We could have added noack and copied_seq to desc, and then called
ops->read_sock. However, unfortunately, other modules didn’t fully
initialize desc to zero. So, for now, we are directly calling
tcp_read_sock_noack() in tcp_bpf.c.

[1]: https://lore.kernel.org/bpf/20241218053408.437295-1-mrpre@163.com

Fixes: e5c6de5fa025 ("bpf, sockmap: Incorrectly handling copied_seq")
Suggested-by: Jakub Sitnicki 
Signed-off-by: Jiayuan Chen 
Signed-off-by: Martin KaFai Lau 
Reviewed-by: Jakub Sitnicki 
Acked-by: John Fastabend 
Link: https://patch.msgid.link/20250122100917.49845-3-mrpre@163.com

skmsg: Return copied bytes in sk_msg_memcopy_from_iter

2024-12-20T21:53:36+00:00

Previously sk_msg_memcopy_from_iter returns the copied bytes from the
last copy_from_iter{,_nocache} call upon success.

This commit changes it to return the total number of copied bytes on
success.

Signed-off-by: Levi Zim 
Signed-off-by: Daniel Borkmann 
Tested-by: Björn Töpel 
Reviewed-by: John Fastabend 
Link: https://lore.kernel.org/bpf/20241130-tcp-bpf-sendmsg-v1-1-bae583d014f3@outlook.com

tcp_bpf: Add sk_rmem_alloc related logic for tcp_bpf ingress redirection

2024-12-20T16:59:47+00:00

When we do sk_psock_verdict_apply->sk_psock_skb_ingress, an sk_msg will
be created out of the skb, and the rmem accounting of the sk_msg will be
handled by the skb.

For skmsgs in __SK_REDIRECT case of tcp_bpf_send_verdict, when redirecting
to the ingress of a socket, although we sk_rmem_schedule and add sk_msg to
the ingress_msg of sk_redir, we do not update sk_rmem_alloc. As a result,
except for the global memory limit, the rmem of sk_redir is nearly
unlimited. Thus, add sk_rmem_alloc related logic to limit the recv buffer.

Since the function sk_msg_recvmsg and __sk_psock_purge_ingress_msg are
used in these two paths. We use "msg->skb" to test whether the sk_msg is
skb backed up. If it's not, we shall do the memory accounting explicitly.

Fixes: 604326b41a6f ("bpf, sockmap: convert to generic sk_msg interface")
Signed-off-by: Zijian Zhang 
Signed-off-by: Daniel Borkmann 
Reviewed-by: John Fastabend 
Link: https://lore.kernel.org/bpf/20241210012039.1669389-3-zijianzhang@bytedance.com

bpf: fix recursive lock when verdict program return SK_PASS

2024-11-19T03:39:59+00:00

When the stream_verdict program returns SK_PASS, it places the received skb
into its own receive queue, but a recursive lock eventually occurs, leading
to an operating system deadlock. This issue has been present since v6.9.

'''
sk_psock_strp_data_ready
    write_lock_bh(&sk->sk_callback_lock)
    strp_data_ready
      strp_read_sock
        read_sock -> tcp_read_sock
          strp_recv
            cb.rcv_msg -> sk_psock_strp_read
              # now stream_verdict return SK_PASS without peer sock assign
              __SK_PASS = sk_psock_map_verd(SK_PASS, NULL)
              sk_psock_verdict_apply
                sk_psock_skb_ingress_self
                  sk_psock_skb_ingress_enqueue
                    sk_psock_data_ready
                      read_lock_bh(&sk->sk_callback_lock) <= dead lock

'''

This topic has been discussed before, but it has not been fixed.
Previous discussion:
https://lore.kernel.org/all/6684a5864ec86_403d20898@john.notmuch

Fixes: 6648e613226e ("bpf, skmsg: Fix NULL pointer dereference in sk_psock_skb_ingress_enqueue")
Reported-by: Vincent Whitchurch 
Signed-off-by: Jiayuan Chen 
Signed-off-by: John Fastabend 
Acked-by: Martin KaFai Lau 
Link: https://patch.msgid.link/20241118030910.36230-2-mrpre@163.com
Signed-off-by: Jakub Kicinski

bpf, sockmap: Correct spelling skmsg.c

2024-09-02T16:43:58+00:00

Correct spelling in skmsg.c. As reported by codespell.

Signed-off-by: Simon Horman 
Signed-off-by: Daniel Borkmann 
Acked-by: Stanislav Fomichev 
Link: https://lore.kernel.org/bpf/20240829-sockmap-spell-v1-1-a614d76564cc@kernel.org

skmsg: Skip zero length skb in sk_msg_recvmsg

2024-07-09T08:24:50+00:00

When running BPF selftests (./test_progs -t sockmap_basic) on a Loongarch
platform, the following kernel panic occurs:

  [...]
  Oops[#1]:
  CPU: 22 PID: 2824 Comm: test_progs Tainted: G           OE  6.10.0-rc2+ #18
  Hardware name: LOONGSON Dabieshan/Loongson-TC542F0, BIOS Loongson-UDK2018
     ... ...
     ra: 90000000048bf6c0 sk_msg_recvmsg+0x120/0x560
    ERA: 9000000004162774 copy_page_to_iter+0x74/0x1c0
   CRMD: 000000b0 (PLV0 -IE -DA +PG DACF=CC DACM=CC -WE)
   PRMD: 0000000c (PPLV0 +PIE +PWE)
   EUEN: 00000007 (+FPE +SXE +ASXE -BTE)
   ECFG: 00071c1d (LIE=0,2-4,10-12 VS=7)
  ESTAT: 00010000 [PIL] (IS= ECode=1 EsubCode=0)
   BADV: 0000000000000040
   PRID: 0014c011 (Loongson-64bit, Loongson-3C5000)
  Modules linked in: bpf_testmod(OE) xt_CHECKSUM xt_MASQUERADE xt_conntrack
  Process test_progs (pid: 2824, threadinfo=0000000000863a31, task=...)
  Stack : ...
  Call Trace:
  [<9000000004162774>] copy_page_to_iter+0x74/0x1c0
  [<90000000048bf6c0>] sk_msg_recvmsg+0x120/0x560
  [<90000000049f2b90>] tcp_bpf_recvmsg_parser+0x170/0x4e0
  [<90000000049aae34>] inet_recvmsg+0x54/0x100
  [<900000000481ad5c>] sock_recvmsg+0x7c/0xe0
  [<900000000481e1a8>] __sys_recvfrom+0x108/0x1c0
  [<900000000481e27c>] sys_recvfrom+0x1c/0x40
  [<9000000004c076ec>] do_syscall+0x8c/0xc0
  [<9000000003731da4>] handle_syscall+0xc4/0x160
  Code: ...
  ---[ end trace 0000000000000000 ]---
  Kernel panic - not syncing: Fatal exception
  Kernel relocated by 0x3510000
   .text @ 0x9000000003710000
   .data @ 0x9000000004d70000
   .bss  @ 0x9000000006469400
  ---[ end Kernel panic - not syncing: Fatal exception ]---
  [...]

This crash happens every time when running sockmap_skb_verdict_shutdown
subtest in sockmap_basic.

This crash is because a NULL pointer is passed to page_address() in the
sk_msg_recvmsg(). Due to the different implementations depending on the
architecture, page_address(NULL) will trigger a panic on Loongarch
platform but not on x86 platform. So this bug was hidden on x86 platform
for a while, but now it is exposed on Loongarch platform. The root cause
is that a zero length skb (skb->len == 0) was put on the queue.

This zero length skb is a TCP FIN packet, which was sent by shutdown(),
invoked in test_sockmap_skb_verdict_shutdown():

	shutdown(p1, SHUT_WR);

In this case, in sk_psock_skb_ingress_enqueue(), num_sge is zero, and no
page is put to this sge (see sg_set_page in sg_set_page), but this empty
sge is queued into ingress_msg list.

And in sk_msg_recvmsg(), this empty sge is used, and a NULL page is got by
sg_page(sge). Pass this NULL page to copy_page_to_iter(), which passes it
to kmap_local_page() and to page_address(), then kernel panics.

To solve this, we should skip this zero length skb. So in sk_msg_recvmsg(),
if copy is zero, that means it's a zero length skb, skip invoking
copy_page_to_iter(). We are using the EFAULT return triggered by
copy_page_to_iter to check for is_fin in tcp_bpf.c.

Fixes: 604326b41a6f ("bpf, sockmap: convert to generic sk_msg interface")
Suggested-by: John Fastabend 
Signed-off-by: Geliang Tang 
Signed-off-by: Daniel Borkmann 
Reviewed-by: John Fastabend 
Link: https://lore.kernel.org/bpf/e3a16eacdc6740658ee02a33489b1b9d4912f378.1719992715.git.tanggeliang@kylinos.cn