Re: Requesting Help about submit ‘’mptcp: cope racing subflow creation in mptcp_rcv_space_adjust‘’

Re: Requesting Help about submit ‘’mptcp: cope racing subflow creation in mptcp_rcv_space_adjust‘’

[Paolo Abeni]

Please avoid direct email messages. There is an mptcp-specific ML that
you should CC for this kind of questions. Adding it.

On 8/19/25 9:14 AM, Yang, Zhi wrote:
> |I am an engineer from Wind River and I want to backport the CVE(|_|
> CVE-2024-53122 <https://nvd.nist.gov/vuln/detail/CVE-2024-53122>|_|) to
> kernel v5.10 now.|
> U|pstream submit is below:|
> /_|https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/
> commit/?id=ce7356ae35943cc6494cc692e62d51a734062b7d <https://
> git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?
> id=ce7356ae35943cc6494cc692e62d51a734062b7d>|_/
> |But there is no function |/|tcp_can_send_ack|/| in kernel v5.10 for it
> was first added from commit fd8976790a6c1f ("mptcp: be careful on MPTCP-
> level ack") in v5.11-rc1.|
> |So I change |/|tcp_can_send_ack|/|to its latest defination in kernel
> v5.10 such as:|
> /|- tcp_cleanup_rbuf(ssk, 1);|/
> */+ if (tcp_can_send_ack(ssk))/*
> /+ tcp_cleanup_rbuf(ssk, 1);/
> |->|
> /|- tcp_cleanup_rbuf(ssk, 1);|/
> */+ if (!((1 << inet_sk_state_load(ssk)) &/*
> */(TCPF_SYN_SENT | TCPF_SYN_RECV | TCPF_TIME_WAIT | TCPF_CLOSE |
> TCPF_LISTEN)))/*
> /|+ tcp_cleanup_rbuf(ssk, 1);|/
> 
> |Can you tell me if It’s necessary to backport the fix, and if
> necessary, the modification is OK?|

I can't tell if the mentioned fix is necessary for you, because it
depends on a lot of factors: do you use 5.10 in production? do you have
mptcp enabled? do you have actual services using mptcp? If all the
replies to the above questions are 'yes', than likely the mentioned fix
is needed. But in such a case you will be much better off moving to a
(much newer) kernel. Or rebase the whole mptcp stack.

In any case the code above is IMHO a reasonable backport of the
mentioned fix.

/P

> 
> |Thank you for your time and looking forward to your reply!|