bpf.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jakub Sitnicki <jakub@cloudflare.com>
To: Martin Lau <kafai@fb.com>
Cc: "bpf\@vger.kernel.org" <bpf@vger.kernel.org>,
	"netdev\@vger.kernel.org" <netdev@vger.kernel.org>,
	"kernel-team\@cloudflare.com" <kernel-team@cloudflare.com>,
	John Fastabend <john.fastabend@gmail.com>
Subject: Re: [PATCH bpf-next 4/8] bpf, sockmap: Don't let child socket inherit psock or its ops on copy
Date: Tue, 26 Nov 2019 19:36:09 +0100	[thread overview]
Message-ID: <87d0deo57q.fsf@cloudflare.com> (raw)
In-Reply-To: <20191126171607.pzrg5qhbavh7enwh@kafai-mbp.dhcp.thefacebook.com>

On Tue, Nov 26, 2019 at 06:16 PM CET, Martin Lau wrote:
> On Tue, Nov 26, 2019 at 04:54:33PM +0100, Jakub Sitnicki wrote:
>> On Mon, Nov 25, 2019 at 11:38 PM CET, Martin Lau wrote:
>> > On Sat, Nov 23, 2019 at 12:07:47PM +0100, Jakub Sitnicki wrote:
>> > [ ... ]
>> >
>> >> @@ -370,6 +378,11 @@ static inline void sk_psock_restore_proto(struct sock *sk,
>> >>  			sk->sk_prot = psock->sk_proto;
>> >>  		psock->sk_proto = NULL;
>> >>  	}
>> >> +
>> >> +	if (psock->icsk_af_ops) {
>> >> +		icsk->icsk_af_ops = psock->icsk_af_ops;
>> >> +		psock->icsk_af_ops = NULL;
>> >> +	}
>> >>  }
>> >
>> > [ ... ]
>> >
>> >> +static struct sock *tcp_bpf_syn_recv_sock(const struct sock *sk,
>> >> +					  struct sk_buff *skb,
>> >> +					  struct request_sock *req,
>> >> +					  struct dst_entry *dst,
>> >> +					  struct request_sock *req_unhash,
>> >> +					  bool *own_req)
>> >> +{
>> >> +	const struct inet_connection_sock_af_ops *ops;
>> >> +	void (*write_space)(struct sock *sk);
>> >> +	struct sk_psock *psock;
>> >> +	struct proto *proto;
>> >> +	struct sock *child;
>> >> +
>> >> +	rcu_read_lock();
>> >> +	psock = sk_psock(sk);
>> >> +	if (likely(psock)) {
>> >> +		proto = psock->sk_proto;
>> >> +		write_space = psock->saved_write_space;
>> >> +		ops = psock->icsk_af_ops;
>> > It is not immediately clear to me what ensure
>> > ops is not NULL here.
>> >
>> > It is likely I missed something.  A short comment would
>> > be very useful here.
>>
>> I can see the readability problem. Looking at it now, perhaps it should
>> be rewritten, to the same effect, as:
>>
>> static struct sock *tcp_bpf_syn_recv_sock(...)
>> {
>> 	const struct inet_connection_sock_af_ops *ops = NULL;
>>         ...
>>
>>         rcu_read_lock();
>> 	psock = sk_psock(sk);
>> 	if (likely(psock)) {
>> 		proto = psock->sk_proto;
>> 		write_space = psock->saved_write_space;
>> 		ops = psock->icsk_af_ops;
>> 	}
>> 	rcu_read_unlock();
>>
>>         if (!ops)
>> 		ops = inet_csk(sk)->icsk_af_ops;
>>         child = ops->syn_recv_sock(sk, skb, req, dst, req_unhash, own_req);
>>
>> If psock->icsk_af_ops were NULL, it would mean we haven't initialized it
>> properly. To double check what happens here:
> I did not mean the init path.  The init path is fine since it init
> eveything on psock before publishing the sk to the sock_map.
>
> I was thinking the delete path (e.g. sock_map_delete_elem).  It is not clear
> to me what prevent the earlier pasted sk_psock_restore_proto() which sets
> psock->icsk_af_ops to NULL from running in parallel with
> tcp_bpf_syn_recv_sock()?  An explanation would be useful.

Ah, I misunderstood. Nothing prevents the race, AFAIK.

Setting psock->icsk_af_ops to null on restore and not checking for it
here was a bad move on my side.  Also I need to revisit what to do about
psock->sk_proto so the child socket doesn't end up with null sk_proto.

This race should be easy enough to trigger. Will give it a shot.

Thank you for bringing this up,
Jakub

>
>>
>> In sock_map_link we do a setup dance where we first create the psock and
>> later initialize the socket callbacks (tcp_bpf_init).
>>
>> static int sock_map_link(struct bpf_map *map, struct sk_psock_progs *progs,
>> 			 struct sock *sk)
>> {
>>         ...
>> 	if (psock) {
>>                 ...
>> 	} else {
>> 		psock = sk_psock_init(sk, map->numa_node);
>> 		if (!psock) {
>> 			ret = -ENOMEM;
>> 			goto out_progs;
>> 		}
>> 		sk_psock_is_new = true;
>> 	}
>>         ...
>>         if (sk_psock_is_new) {
>> 		ret = tcp_bpf_init(sk);
>> 		if (ret < 0)
>> 			goto out_drop;
>> 	} else {
>> 		tcp_bpf_reinit(sk);
>> 	}
>>
>> The "if (sk_psock_new)" branch triggers the call chain that leads to
>> saving & overriding socket callbacks.
>>
>> tcp_bpf_init -> tcp_bpf_update_sk_prot -> sk_psock_update_proto
>>
>> Among them, icsk_af_ops.
>>
>> static inline void sk_psock_update_proto(...)
>> {
>>         ...
>> 	psock->icsk_af_ops = icsk->icsk_af_ops;
>> 	icsk->icsk_af_ops = af_ops;
>> }
>>
>> Goes without saying that a comment is needed.
>>
>> Thanks for the feedback,
>> Jakub

  reply	other threads:[~2019-11-26 18:36 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-11-23 11:07 [PATCH bpf-next 0/8] Extend SOCKMAP to store listening sockets Jakub Sitnicki
2019-11-23 11:07 ` [PATCH bpf-next 1/8] bpf, sockmap: Return socket cookie on lookup from syscall Jakub Sitnicki
2019-11-24  5:32   ` John Fastabend
2019-11-23 11:07 ` [PATCH bpf-next 2/8] bpf, sockmap: Let all kernel-land lookup values in SOCKMAP Jakub Sitnicki
2019-11-24  5:35   ` John Fastabend
2019-11-23 11:07 ` [PATCH bpf-next 3/8] bpf, sockmap: Allow inserting listening TCP sockets into SOCKMAP Jakub Sitnicki
2019-11-24  5:38   ` John Fastabend
2019-11-23 11:07 ` [PATCH bpf-next 4/8] bpf, sockmap: Don't let child socket inherit psock or its ops on copy Jakub Sitnicki
2019-11-24  5:56   ` John Fastabend
2019-11-25 22:38   ` Martin Lau
2019-11-26 15:54     ` Jakub Sitnicki
2019-11-26 17:16       ` Martin Lau
2019-11-26 18:36         ` Jakub Sitnicki [this message]
     [not found]           ` <87sglsfdda.fsf@cloudflare.com>
2019-12-11 17:20             ` Martin Lau
2019-12-12 11:27               ` Jakub Sitnicki
2019-12-12 19:23                 ` Martin Lau
2019-12-17 15:06                   ` Jakub Sitnicki
2019-11-26 18:43         ` John Fastabend
2019-11-27 22:18           ` Jakub Sitnicki
2019-11-23 11:07 ` [PATCH bpf-next 5/8] bpf: Allow selecting reuseport socket from a SOCKMAP Jakub Sitnicki
2019-11-24  5:57   ` John Fastabend
2019-11-25  1:24   ` Alexei Starovoitov
2019-11-25  4:17     ` John Fastabend
2019-11-25 10:40       ` Jakub Sitnicki
2019-11-25 22:07         ` Martin Lau
2019-11-26 14:30           ` Jakub Sitnicki
2019-11-26 19:03             ` Martin Lau
2019-11-27 21:34               ` Jakub Sitnicki
2019-11-23 11:07 ` [PATCH bpf-next 6/8] libbpf: Recognize SK_REUSEPORT programs from section name Jakub Sitnicki
2019-11-24  5:57   ` John Fastabend
2019-11-23 11:07 ` [PATCH bpf-next 7/8] selftests/bpf: Extend SK_REUSEPORT tests to cover SOCKMAP Jakub Sitnicki
2019-11-24  6:00   ` John Fastabend
2019-11-25 22:30   ` Martin Lau
2019-11-26 14:32     ` Jakub Sitnicki
2019-12-12 10:30     ` Jakub Sitnicki
2019-11-23 11:07 ` [PATCH bpf-next 8/8] selftests/bpf: Tests for SOCKMAP holding listening sockets Jakub Sitnicki
2019-11-24  6:04   ` John Fastabend
2019-11-24  6:10 ` [PATCH bpf-next 0/8] Extend SOCKMAP to store " John Fastabend
2019-11-25  9:22   ` Jakub Sitnicki

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87d0deo57q.fsf@cloudflare.com \
    --to=jakub@cloudflare.com \
    --cc=bpf@vger.kernel.org \
    --cc=john.fastabend@gmail.com \
    --cc=kafai@fb.com \
    --cc=kernel-team@cloudflare.com \
    --cc=netdev@vger.kernel.org \
    --subject='Re: [PATCH bpf-next 4/8] bpf, sockmap: Don'\''t let child socket inherit psock or its ops on copy' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).