Jens Axboe

marcan on leaving the Asahi project: "No matter how much we did, how many impossible feats we pulled off, people always wanted more. And more."

Yep this is basically ANY open source project. Doesn't matter what you give people, they will always complain about something.

The networking side of the io_uring receive zero-copy support was finally merged, it's queued for the 6.15 kernel. See this merge for details:

git.kernel.org/pub/scm/linu...

Now we just need to stage the io_uring side on top of that! Exciting to finally see this happening.

epoll wait support for io_uring. This may seem like an odd thing, but it's handy for mixed use cases or event loop conversions. See the cover letter linked below for details:

lore.kernel.org/linux-fsdeve...

and code:

git.kernel.dk/cgit/linux/l...

The core of this work is now upstream in Linus's tree, and the per-fs bits have been rebased on top of that:

git.kernel.dk/cgit/linux/l...

xfs/btrfs/nfs should be fine to include, ext4 would ideally need an iomap conversion first.

Want to make a real impact on the 12th edition of Kernel Recipes?

Support #kr2025, the conference that brings the Linux kernel community together. Let’s keep it accessible, inspiring, and innovative—your sponsorship could make all the difference! 🤝

Join us:
kernel-recipes.org/en/2025/spon...

To be fair, it is actually pretty big. Probably not as big as he'd think looking at the map, but it's big none the less.

Only thing I've found so far is that now I'm posting on 3 separate spots rather than just 1! But yeah, the cesspool that is X is not great. Still the best engagement on the tech front though :-(

Posted v8 of the uncached buffered IO patchset. This should be fine for merging at this point, let's hope we can make the 6.14 kernel release.

lore.kernel.org/linux-fsdeve...

Separate branches exist for the core support and fs support, see the cover letter for details.

Sign in · GitLab GitLab Enterprise Edition at UW

Interesting memcached experiment from @UWaterloo showing some nice wins. Maybe 2025 will be the year of io_uring networking.

git.uwaterloo.ca/lseo/io_uring