Re: [PATCH mptcp-next v5 6/7] mptcp: track app-limited state in mptcp_sendmsg

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Lockdep recursive locking warning due to missing nested lock annotation in mptcp_sock_rate_check_app_limited().
--

--- Patch [6]: [PATCH mptcp-next v5 6/7] mptcp: track app-limited state in mptcp_sendmsg ---
commit 991a38880507b6d69cccc4a195bee5ef68452bf9
Author: Geliang Tang <tanggeliang@kylinos.cn>

mptcp: track app-limited state in mptcp_sendmsg

This commit splits the TCP app-limited rate checking logic into a helper
that takes a tcp_sock, and adds an MPTCP-specific helper to apply this
rate checking to each subflow individually during sendmsg.

[ ... ]

> diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
> index 648eba89278f6..735cd39846462 100644
> --- a/net/mptcp/protocol.c
> +++ b/net/mptcp/protocol.c
> @@ -2045,6 +2045,21 @@ static void mptcp_rps_record_subflows(const struct mptcp_sock *msk)
>  	}
>  }
>  
> +static void mptcp_sock_rate_check_app_limited(struct sock *sk)
> +{
> +	struct mptcp_sock *msk = mptcp_sk(sk);
> +	struct mptcp_subflow_context *subflow;
> +
> +	mptcp_for_each_subflow(msk, subflow) {
> +		struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
> +		bool slow;
> +
> +		slow = lock_sock_fast(ssk);

[Severity: Medium]
Does acquiring the subflow socket lock here trigger a lockdep recursive
locking warning?

mptcp_sendmsg() acquires the parent MPTCP socket lock using lock_sock(sk),
and then calls mptcp_sock_rate_check_app_limited(sk) right after. This
function iterates over the connection subflows and acquires each subflow's
socket lock via lock_sock_fast(ssk).

Because both locks share the same lock class, lockdep might flag this
nested acquisition as a possible recursive lock splat, which can disable
further lock debugging.

Is it possible to use lock_sock_fast_nested(ssk) or a similar nested lock
annotation here to suppress the false positive?

> +		tcp_sock_rate_check_app_limited(tcp_sk(ssk));
> +		unlock_sock_fast(ssk, slow);
> +	}
> +}
> +
>  static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
>  {
>  	struct mptcp_sock *msk = mptcp_sk(sk);

[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/cover.1787368526.git.tanggeliang@kylinos.cn?part=6