Re: [PATCH mptcp-net 3/3] mptcp: fix memcg accounting for passive sockets

[Matthieu Baerts <matttbe@kernel.org>]

Hi Geliang,

On 11/11/2025 09:37, Geliang Tang wrote:
> Hi Paolo,
> 
> Thanks for this fix.
> 
> On Fri, 2025-11-07 at 22:55 +0100, Paolo Abeni wrote:
>> The passive sockets never got proper memcg accounting: the msk
>> socket is associated with the memcg at accept time, but the
>> passive subflows never got it right.
>>
>> At accept time, traverse the subflows list and associate each of them
>> with the msk memcg, and try to do the same at join completion time,
>> if
>> the msk has been already accepted.
>>
>> Fixes: cf7da0d66cc1 ("mptcp: Create SUBFLOW socket for incoming
>> connections")
>> Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/298
>> Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/597
>> Signed-off-by: Paolo Abeni <pabeni@redhat.com>
>> ---
>>  net/mptcp/protocol.c | 36 ++++++++++++++++++++++++++----------
>>  net/mptcp/protocol.h |  1 +
>>  net/mptcp/subflow.c  | 10 ++++++++++
>>  3 files changed, 37 insertions(+), 10 deletions(-)
>>
>> diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
>> index 34e6bc731085..5e9325c7ea9c 100644
>> --- a/net/mptcp/protocol.c
>> +++ b/net/mptcp/protocol.c
>> @@ -959,6 +959,8 @@ static bool __mptcp_finish_join(struct mptcp_sock
>> *msk, struct sock *ssk)
>>  	mptcp_subflow_ctx(ssk)->subflow_id = msk->subflow_id++;
>>  	mptcp_sockopt_sync_locked(msk, ssk);
>>  	mptcp_stop_tout_timer(sk);
>> +	__mptcp_inherit_cgrp_data(sk, ssk);
>> +	__mptcp_inherit_memcg(sk, ssk, GFP_ATOMIC);
> 
> In mptcp_graph_subflows(), these two lines already have the
> mem_cgroup_from_sk() check (in the squash-to patch). It's unclear
> whether mem_cgroup_from_sk() check needs to be added here too.

From what I understood, no need to check mem_cgroup_from_sk() here:

- __mptcp_inherit_cgrp_data() should not depend on that (I think)
- __mptcp_inherit_memcg() is doing that via mem_cgroup_sockets_enabled()

I *think* the squash-to patch needs some small further modifications,
but this series looks good to me.

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.