This package is now a split package, installing both userspace tools and DKMS module sources, similar to extra/bcachefs-{tools,dkms}.
Additionally, starting with v1.37.0, FUSE support is unconditionally enabled in bcachefs-tools.

Any issues with the package can be discussed on #bcachefs or #bcachefs:matrix.bcachefs.org.

doesnt currently build tons of undefined symbol errors. example)

  = note: rust-lld: error: undefined symbol: rust_fuse_rcu_unregister
          >>> referenced by fusemount.rs:37 (src/commands/fusemount.rs:37)

Now that bcachefs-kernel-dkms-git is being deprecated, I had to rollback to 68f3f723779a for things to work properly.

@ticpu, the current package uses the same split between tools and DKMS as the package in [extra] (which is also tne general practice in Arch, see e.g. nvidia module). Does it not work for you?

If you need any help integrating the other changes such as the -debug split, I volunteer as Co-Maintainer as well.

I'm not aware of any issues building this package with options=(debug) either. The -debug packages are well-formed. What is it that you suggest to integrate?

Now that bcachefs-kernel-dkms-git is being deprecated, I had to rollback to 68f3f723779a for things to work properly.

This commit removed DKMS with a dubious commit message:

commit 598ab6dd10f7ba0948c2d65eda692bd69340aceb
Author: neycrol <330578697@qq.com>
Date:   Tue Feb 3 21:04:41 2026 +0800

 chore: normalize pkgver and cleanup

If you need any help integrating the other changes such as the -debug split, I volunteer as Co-Maintainer as well.

This package is now a split package, installing both userspace tools and DKMS module sources, similar to extra/bcachefs-{tools,dkms}.
Additionally, starting with v1.37.0, FUSE support is unconditionally enabled in bcachefs-tools.

Any issues with the package can be discussed on #bcachefs or #bcachefs:matrix.bcachefs.org.

Yes, I'm an active user of bcachefs and I'd be glad to co-maintain.

@intelfx Thanks for the patch series — I've looked through them and they make sense. I'm a bit short on time for personal reasons at the moment, so I haven't been able to test them properly. Would you be interested in co-maintaining this package? That way you could apply and iterate on these changes directly. Let me know.

Hi, given the situation with bcachefs-kernel-dkms-git, would you please consider applying this patch to provide DKMS kernel module code in this package as well?

https://0x0.st/PLks.mbox

The last patch in the series creates a split pkgbase for -tools and -dkms, mirroring the package in [extra].

Feel free to apply or ignore any of the previous patches; those are cleanups I have locally.

Thank you. (In hindsight, I realize my message was worded way too aggressively. My apologies for that.)

Thanks for the feedback, @intelfx.

After carefully reviewing the VCS packaging guidelines, I agree with most of your points. I’ve reworked and polished the package accordingly, and the updated version is now published.

Appreciate you taking the time to point these out.

@neycrol

Is there any reason you dropped soname dependencies, and changed pkgver() to use current date instead of a git tag (which is expressly discouraged by the packaging guidelines for VCS packages)? Both of these changes are wrong and should be reverted. Please keep your package in sync with extra/bcachefs-tools unless you have a good reason to deviate.

Continuing, RUSTFLAGS="... -C link-arg=$LDFLAGS ... is bogus and will break the build as soon as the user's LDFLAGS happen to contain a space.

Please use cargo fetch --locked --target "$(rustc --print host-tuple)" instead of deriving a tuple by hand.

There are other things as well but these are most visible.