What happened around #MPTCP ๐ in the last couple of months? Bug-fixes ๐, improvements ๐, app support โ๏ธ, conferences ๐๏ธ, whatโs next ๐ฎ: read on to find out more about what happened recently ๐ฉโ๐ปโคต๏ธ
#MPTCP features that will be present in the future #Linux v6.12: ๐ง
blackhole_timeout
sysctlMP_JOIN
to help in case of configuration issues ๐(As always: these features have been added to Linus tree, hopefully they will still be there when this version will be released!)
Can maintainers easily ignore old bugs they discover when working on new features? ๐
Delaying these features now, or having to deal with bugs later, possibly when it is really not a good time for that? ๐
I wrote a bit about this topic in my last blog post ๐ฐ โคต๏ธ
https://blog.mptcp.dev/2024/07/21/maintainers-responsibilities.html
I don't know if I did the good choice at the end, but don't hesitate to share your opinions! ๐
mptcp.dev website ๐ has been updated last month. This new website was really needed to help supporting new users who are interested by #MPTCP. I wrote a blog post about that ๐ฐ:
https://blog.mptcp.dev/2024/05/03/supporting-new-users.html
This has been made possible thanks to @NGIZero support, making this MPTCP project more community oriented, and that's great ๐
Our new ๐ website is now online ๐
โก๏ธ https://www.mptcp.dev ๐
It is now for a wider audience:
Feel free to have a look, and suggest modifications!
Thanks to everybody who was involved in this upgrade!
To celebrate that, we have a new logo! ๐ฅณ
Backporting bug ๐ fixes to stable Linux kernel versions: I recently published a new blog post showing what kernel maintainers usually have to do:
https://blog.mptcp.dev/2024/03/04/backports.html
And some news about #MPTCP in the #Linux kernel as usual! ๐ฐ
Recently, I stopped my work at #Tessares and started a new challenge: Improving #MPTCP deployability, sponsored by @NGIZero (many thanks for letting me doing that!).
https://nlnet.nl/project/MPTCP-deployability/
My new blog is there if you want more details about this "Angel Project" and what I did so far: https://blog.mptcp.dev/2024/01/01/Angel-Project.html
My recent work for the #MPTCP upstream project was focused on the #CI migration to GitHub Actions ๐ค, fixing issues with the new Netdev CI (finally ๐) validating the selftests, fixing issues with external tools (QEmu, #ss from IPRoute2) ๐, and more. I'm still proud to be sponsored by @NGIZero ๐
Find out more on this new blog post: https://blog.mptcp.dev/2024/02/02/CI-CI-CI.html
#MPTCP in the future #Linux v6.5 version should have more exposed info:
ADD_ADDR
(+ echo) transmission MIB countersgetsockopt(MPTCP_INFO)
getsockopt(MPTCP_FULL_INFO)
aggregating MPTCP and subflows info (with ID)(These modifications are in Linusโ tree, hopefully everything will still be there when v6.5 will be released!)
#MPTCP in #Linux v6.4 got some improvements around the reception of connection requests:
MPTCP_INFO
and Netlink (ss -M
) no longer fill info not used by the PM in usedMany fixes have also been done and backported to stable versions. There was even a special mention in v6.3-rc7 announcement ๐
Nothing particular stands out in the rc this week, unless you count the mptcp selftest changes that are about making the tests work on stable kernels too. That accounts for almost a quarter of the patchโฆ