[PATCH bpf v2 0/2] bpf, sockmap fixes

[PATCH bpf v2 0/2] bpf, sockmap fixes

[John Fastabend]

This addresses an issue found while reviewing latest round of sock
map patches and an issue reported from CI via Andrii. After this
CI ./test_maps is stable for me.

The CI discovered issue was introduced by over correcting our
previously broken memory accounting. After the fix, "bpf, sockmap:
Avoid returning unneeded EAGAIN when redirecting to self" we fixed
a dropped packet and a missing fwd_alloc calculations, but pushed
it too far back into the packet pipeline creating an issue in the
unlikely case socket tear down happens with an enqueued skb. See
patch for details.

Tested with usual suspects: test_sockmap, test_maps, test_progs
and test_progs-no_alu32.

v2: drop skb_orphan its not necessary and use sk directly instead
    of using psock->sk both suggested by Cong

---

John Fastabend (2):
      bpf, sockmap: fix sk->prot unhash op reset
      bpf, sockmap: fix incorrect fwd_alloc accounting


 net/core/skmsg.c | 12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)

--

[PATCH bpf v2 1/2] bpf, sockmap: fix sk->prot unhash op reset

[John Fastabend]

In '4da6a196f93b1' we fixed a potential unhash loop caused when
a TLS socket in a sockmap was removed from the sockmap. This
happened because the unhash operation on the TLS ctx continued
to point at the sockmap implementation of unhash even though the
psock has already been removed. The sockmap unhash handler when a
psock is removed does the following,

 void sock_map_unhash(struct sock *sk)
 {
	void (*saved_unhash)(struct sock *sk);
	struct sk_psock *psock;

	rcu_read_lock();
	psock = sk_psock(sk);
	if (unlikely(!psock)) {
		rcu_read_unlock();
		if (sk->sk_prot->unhash)
			sk->sk_prot->unhash(sk);
		return;
	}
        [...]
 }

The unlikely() case is there to handle the case where psock is detached
but the proto ops have not been updated yet. But, in the above case
with TLS and removed psock we never fixed sk_prot->unhash() and unhash()
points back to sock_map_unhash resulting in a loop. To fix this we added
this bit of code,

 static inline void sk_psock_restore_proto(struct sock *sk,
                                          struct sk_psock *psock)
 {
       sk->sk_prot->unhash = psock->saved_unhash;

This will set the sk_prot->unhash back to its saved value. This is the
correct callback for a TLS socket that has been removed from the sock_map.
Unfortunately, this also overwrites the unhash pointer for all psocks.
We effectively break sockmap unhash handling for any future socks.
Omitting the unhash operation will leave stale entries in the map if
a socket transition through unhash, but does not do close() op.

To fix set unhash correctly before calling into tls_update. This way the
TLS enabled socket will point to the saved unhash() handler.

Fixes: 4da6a196f93b1 ("bpf: Sockmap/tls, during free we may call tcp_bpf_unhash() in loop")
Reported-by: Cong Wang <xiyou.wangcong@gmail.com>
Reported-by: Lorenz Bauer <lmb@cloudflare.com>
Suggested-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
---
 include/linux/skmsg.h |    7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/include/linux/skmsg.h b/include/linux/skmsg.h
index 8edbbf5f2f93..822c048934e3 100644
--- a/include/linux/skmsg.h
+++ b/include/linux/skmsg.h
@@ -349,8 +349,13 @@ static inline void sk_psock_update_proto(struct sock *sk,
 static inline void sk_psock_restore_proto(struct sock *sk,
 					  struct sk_psock *psock)
 {
-	sk->sk_prot->unhash = psock->saved_unhash;
 	if (inet_csk_has_ulp(sk)) {
+		/* TLS does not have an unhash proto in SW cases, but we need
+		 * to ensure we stop using the sock_map unhash routine because
+		 * the associated psock is being removed. So use the original
+		 * unhash handler.
+		 */
+		WRITE_ONCE(sk->sk_prot->unhash, psock->saved_unhash);
 		tcp_update_ulp(sk, psock->sk_proto, psock->saved_write_space);
 	} else {
 		sk->sk_write_space = psock->saved_write_space;

[PATCH bpf v2 2/2] bpf, sockmap: fix incorrect fwd_alloc accounting

[John Fastabend]

Incorrect accounting fwd_alloc can result in a warning when the socket
is torn down,

 [18455.319240] WARNING: CPU: 0 PID: 24075 at net/core/stream.c:208 sk_stream_kill_queues+0x21f/0x230
 [...]
 [18455.319543] Call Trace:
 [18455.319556]  inet_csk_destroy_sock+0xba/0x1f0
 [18455.319577]  tcp_rcv_state_process+0x1b4e/0x2380
 [18455.319593]  ? lock_downgrade+0x3a0/0x3a0
 [18455.319617]  ? tcp_finish_connect+0x1e0/0x1e0
 [18455.319631]  ? sk_reset_timer+0x15/0x70
 [18455.319646]  ? tcp_schedule_loss_probe+0x1b2/0x240
 [18455.319663]  ? lock_release+0xb2/0x3f0
 [18455.319676]  ? __release_sock+0x8a/0x1b0
 [18455.319690]  ? lock_downgrade+0x3a0/0x3a0
 [18455.319704]  ? lock_release+0x3f0/0x3f0
 [18455.319717]  ? __tcp_close+0x2c6/0x790
 [18455.319736]  ? tcp_v4_do_rcv+0x168/0x370
 [18455.319750]  tcp_v4_do_rcv+0x168/0x370
 [18455.319767]  __release_sock+0xbc/0x1b0
 [18455.319785]  __tcp_close+0x2ee/0x790
 [18455.319805]  tcp_close+0x20/0x80

This currently happens because on redirect case we do skb_set_owner_r()
with the original sock. This increments the fwd_alloc memory accounting
on the original sock. Then on redirect we may push this into the queue
of the psock we are redirecting to. When the skb is flushed from the
queue we give the memory back to the original sock. The problem is if
the original sock is destroyed/closed with skbs on another psocks queue
then the original sock will not have a way to reclaim the memory before
being destroyed. Then above warning will be thrown

  sockA                          sockB

  sk_psock_strp_read()
   sk_psock_verdict_apply()
     -- SK_REDIRECT --
     sk_psock_skb_redirect()
                                skb_queue_tail(psock_other->ingress_skb..)

  sk_close()
   sock_map_unref()
     sk_psock_put()
       sk_psock_drop()
         sk_psock_zap_ingress()

At this point we have torn down our own psock, but have the outstanding
skb in psock_other. Note that SK_PASS doesn't have this problem because
the sk_psock_drop() logic releases the skb, its still associated with
our psock.

To resolve lets only account for sockets on the ingress queue that are
still associated with the current socket. On the redirect case we will
check memory limits per 6fa9201a89898, but will omit fwd_alloc accounting
until skb is actually enqueued. When the skb is sent via skb_send_sock_locked
or received with sk_psock_skb_ingress memory will be claimed on psock_other.

Reported-by: Andrii Nakryiko <andrii@kernel.org>
Fixes: 6fa9201a89898 ("bpf, sockmap: Avoid returning unneeded EAGAIN when redirecting to self")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
---
 net/core/skmsg.c |   12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)

diff --git a/net/core/skmsg.c b/net/core/skmsg.c
index 1261512d6807..5def3a2e85be 100644
--- a/net/core/skmsg.c
+++ b/net/core/skmsg.c
@@ -488,6 +488,7 @@ static int sk_psock_skb_ingress_self(struct sk_psock *psock, struct sk_buff *skb
 	if (unlikely(!msg))
 		return -EAGAIN;
 	sk_msg_init(msg);
+	skb_set_owner_r(skb, sk);
 	return sk_psock_skb_ingress_enqueue(skb, psock, sk, msg);
 }
 
@@ -790,7 +791,6 @@ static void sk_psock_tls_verdict_apply(struct sk_buff *skb, struct sock *sk, int
 {
 	switch (verdict) {
 	case __SK_REDIRECT:
-		skb_set_owner_r(skb, sk);
 		sk_psock_skb_redirect(skb);
 		break;
 	case __SK_PASS:
@@ -808,10 +808,6 @@ int sk_psock_tls_strp_read(struct sk_psock *psock, struct sk_buff *skb)
 	rcu_read_lock();
 	prog = READ_ONCE(psock->progs.skb_verdict);
 	if (likely(prog)) {
-		/* We skip full set_owner_r here because if we do a SK_PASS
-		 * or SK_DROP we can skip skb memory accounting and use the
-		 * TLS context.
-		 */
 		skb->sk = psock->sk;
 		tcp_skb_bpf_redirect_clear(skb);
 		ret = sk_psock_bpf_run(psock, prog, skb);
@@ -880,12 +876,13 @@ static void sk_psock_strp_read(struct strparser *strp, struct sk_buff *skb)
 		kfree_skb(skb);
 		goto out;
 	}
-	skb_set_owner_r(skb, sk);
 	prog = READ_ONCE(psock->progs.skb_verdict);
 	if (likely(prog)) {
+		skb->sk = sk;
 		tcp_skb_bpf_redirect_clear(skb);
 		ret = sk_psock_bpf_run(psock, prog, skb);
 		ret = sk_psock_map_verd(ret, tcp_skb_bpf_redirect_fetch(skb));
+		skb->sk = NULL;
 	}
 	sk_psock_verdict_apply(psock, skb, ret);
 out:
@@ -956,12 +953,13 @@ static int sk_psock_verdict_recv(read_descriptor_t *desc, struct sk_buff *skb,
 		kfree_skb(skb);
 		goto out;
 	}
-	skb_set_owner_r(skb, sk);
 	prog = READ_ONCE(psock->progs.skb_verdict);
 	if (likely(prog)) {
+		skb->sk = sk;
 		tcp_skb_bpf_redirect_clear(skb);
 		ret = sk_psock_bpf_run(psock, prog, skb);
 		ret = sk_psock_map_verd(ret, tcp_skb_bpf_redirect_fetch(skb));
+		skb->sk = NULL;
 	}
 	sk_psock_verdict_apply(psock, skb, ret);
 out:

Re: [PATCH bpf v2 0/2] bpf, sockmap fixes

[patchwork-bot+netdevbpf]

Hello:

This series was applied to bpf/bpf.git (refs/heads/master):

On Thu, 01 Apr 2021 14:59:58 -0700 you wrote:
> This addresses an issue found while reviewing latest round of sock
> map patches and an issue reported from CI via Andrii. After this
> CI ./test_maps is stable for me.
> 
> The CI discovered issue was introduced by over correcting our
> previously broken memory accounting. After the fix, "bpf, sockmap:
> Avoid returning unneeded EAGAIN when redirecting to self" we fixed
> a dropped packet and a missing fwd_alloc calculations, but pushed
> it too far back into the packet pipeline creating an issue in the
> unlikely case socket tear down happens with an enqueued skb. See
> patch for details.
> 
> [...]

Here is the summary with links:
  - [bpf,v2,1/2] bpf, sockmap: fix sk->prot unhash op reset
    https://git.kernel.org/bpf/bpf/c/1c84b33101c8
  - [bpf,v2,2/2] bpf, sockmap: fix incorrect fwd_alloc accounting
    https://git.kernel.org/bpf/bpf/c/144748eb0c44

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

Re: [PATCH bpf v2 0/2] bpf, sockmap fixes

[Andrii Nakryiko]

On Thu, Apr 1, 2021 at 3:00 PM John Fastabend <john.fastabend@gmail.com> wrote:
>
> This addresses an issue found while reviewing latest round of sock
> map patches and an issue reported from CI via Andrii. After this
> CI ./test_maps is stable for me.
>
> The CI discovered issue was introduced by over correcting our
> previously broken memory accounting. After the fix, "bpf, sockmap:
> Avoid returning unneeded EAGAIN when redirecting to self" we fixed
> a dropped packet and a missing fwd_alloc calculations, but pushed
> it too far back into the packet pipeline creating an issue in the
> unlikely case socket tear down happens with an enqueued skb. See
> patch for details.
>
> Tested with usual suspects: test_sockmap, test_maps, test_progs
> and test_progs-no_alu32.
>
> v2: drop skb_orphan its not necessary and use sk directly instead
>     of using psock->sk both suggested by Cong
>
> ---

It might be that this didn't fix all the issues. We just got another
sockmap timeout in test_maps ([0]).

  [0] https://travis-ci.com/github/kernel-patches/bpf/builds/224971212

>
> John Fastabend (2):
>       bpf, sockmap: fix sk->prot unhash op reset
>       bpf, sockmap: fix incorrect fwd_alloc accounting
>
>
>  net/core/skmsg.c | 12 +++++-------
>  1 file changed, 5 insertions(+), 7 deletions(-)
>
> --
>