Re: [RFC PATCH mptcp-next] mptcp: annotate data-races around sysctl reads

Matthieu Baerts <[email protected]>
Newsgroups dev.linux.lists.mptcp
Organization NGI0 Core
Message-ID <[email protected]>
Hi Tao, Gang,

Thank you for looking at this!

On 14/08/2026 07:26, Tao Cui wrote:
> Hi Gang,
> 
> 在 2026/8/14 11:51, [email protected] 写道:
>> August 14, 2026 at 11:27 AM, "Tao Cui" <[email protected] mailto:[email protected]?to=%22Tao%20Cui%22%20%3Ccui.tao%40linux.dev%3E > wrote:
>>
>> Hi, Tao
>>
>> Thanks for your patch, but it has been submitted by Matt before in [1].

Thank you for the review!

>> I think Matt wanted change the PM and sched sysctl from string to atomically,
>> that may need another patches (READ_ONCE is not enough), right? If no, you can
>> wait for the other maintainers' reply.

Correct.

>> If yes, I still think this patch should keep author as matt, the rest of
>> others can be yours.
>>
>> Note: Some patches about sched is only in our export branch, not mainlined,
>> so it's better to do your work based on it.
>>
> 
> Thanks for the review. I did see #626 before sending and referenced
> it, I just read its scope wrong: I assumed the patch it links was only
> about the string sysctls and missed that the numeric readers were
> already covered there.

Do you plan to work on that? If yes, that's great, thank you. Just one
thing: if you are working on addressing an issue that is on GitHub, do
you mind adding a comment there to say that you are working on it,
please? Just to avoid having multiple people doing the same thing. (Just
to be clear: I'm not working on it.)

> And thanks for the export branch tip, I'd only been looking at
> mainline and didn't know some of the sched patches are only in export.
> 
> The only bit mine adds is the WRITE_ONCE() on the pm_type store in
> proc_path_manager(), to pair with the mptcp_get_pm_type() read. Matt,
> if you want that too just grab it, otherwise I can send it as a small
> follow-up.

Good catch! If you are planning to work on the other remaining sysctl,
do you mind taking my patch, add the missing WRITE_ONCE() to it + add
your co-developed-by + signoff to it, then add your new patch(es) on top
of it and send everything (including the first patch), please?

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.