netdev - Re: [PATCH bpf v7 2/5] bpf: fix wrong copied

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <87a5bliyiz.fsf@cloudflare.com>
Date: Mon, 20 Jan 2025 11:13:08 +0100
From: Jakub Sitnicki <jakub@...udflare.com>
To: Jiayuan Chen <mrpre@....com>
Cc: bpf@...r.kernel.org,  john.fastabend@...il.com,  netdev@...r.kernel.org,
  martin.lau@...ux.dev,  ast@...nel.org,  edumazet@...gle.com,
  davem@...emloft.net,  dsahern@...nel.org,  kuba@...nel.org,
  pabeni@...hat.com,  linux-kernel@...r.kernel.org,  song@...nel.org,
  andrii@...nel.org,  mhal@...x.co,  yonghong.song@...ux.dev,
  daniel@...earbox.net,  xiyou.wangcong@...il.com,  horms@...nel.org,
  corbet@....net,  eddyz87@...il.com,  cong.wang@...edance.com,
  shuah@...nel.org,  mykolal@...com,  jolsa@...nel.org,  haoluo@...gle.com,
  sdf@...ichev.me,  kpsingh@...nel.org,  linux-doc@...r.kernel.org
Subject: Re: [PATCH bpf v7 2/5] bpf: fix wrong copied_seq calculation

On Mon, Jan 20, 2025 at 11:35 AM +08, Jiayuan Chen wrote:
> On Sat, Jan 18, 2025 at 11:29:04PM +0800, Jiayuan Chen wrote:
>> On Sat, Jan 18, 2025 at 03:50:22PM +0100, Jakub Sitnicki wrote:
>> > On Thu, Jan 16, 2025 at 10:05 PM +08, Jiayuan Chen wrote:
>> > > '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,
>> > > +}
>> > > +#endif /* CONFIG_BPF_STREAM_PARSER */
>> > > +
>> > >  int tcp_bpf_update_proto(struct sock *sk, struct sk_psock *psock, bool restore)
>> > >  {
>> > >  	int family = sk->sk_family == AF_INET6 ? TCP_BPF_IPV6 : TCP_BPF_IPV4;
>> > > @@ -681,6 +722,12 @@ int tcp_bpf_update_proto(struct sock *sk, struct sk_psock *psock, bool restore)
>> > >  
>> > >  	/* Pairs with lockless read in sk_clone_lock() */
>> > >  	sock_replace_proto(sk, &tcp_bpf_prots[family][config]);
>> > > +#if IS_ENABLED(CONFIG_BPF_STREAM_PARSER)
>> > > +	if (psock->progs.stream_parser && psock->progs.stream_verdict) {
>> > > +		psock->copied_seq = tcp_sk(sk)->copied_seq;
>> > > +		psock->read_sock = tcp_bpf_strp_read_sock;
>> > 
>> > Just directly set psock->strp.cb.read_sock to tcp_bpf_strp_read_sock.
>> > Then we don't need this intermediate psock->read_sock callback, which
>> > doesn't do anything useful.
>> >
>> Ok, I will do this.
>> (BTW, I intended to avoid bringing "struct strparser" into tcp_bpf.c so I
>> added a wrapper function instead in skmsg.c without calling it directly) 
>> 
> I find that tcp_bpf_update_proto is called before sk_psock_init_strp. Any
> assignment of psock->cb.strp will be overwritten in sk_psock_init_strp.

Or just don't set ->read_sock in strp_init.
It's being reset only because you made it so in patch 1 :-)