Differential D19649

Rename fuse(4) to fusefs(4)
ClosedPublic
Actions

Authored by asomers on Mar 19 2019, 9:09 PM.

Details

Reviewers

cem
rgrimes

Commits

rS351940: MFC r345350, r346441, r346765
rS345350: Rename fuse(4) to fusefs(4)

Summary

Rename fuse(4) to fusefs(4)

This makes it more consistent with other filesystems, which all end in "fs",
and more consistent with its mount helper, which is already named
"mount_fusefs".

Diff Detail

Repository

rS FreeBSD src repository - subversion

Lint

Lint Passed

Unit

No Test Coverage

Build Status

Buildable 23186
Build 22229: arc lint + arc unit

Event Timeline

asomers created this revision.Mar 19 2019, 9:09 PM

Herald added subscribers: jonathan, imp. · View Herald TranscriptMar 19 2019, 9:09 PM

Harbormaster completed remote builds in B23186: Diff 55250.Mar 19 2019, 9:09 PM

This change exposes a bug, but I'll deal with that separately: when you "kldload fusefs", the module that gets loaded is named "fuse".

The kldload issue is https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=236647

Please don't move the directory from sys/fs/fuse to sys/fs/fusefs. It just adds needless churn to the history.

This revision now requires changes to proceed.Mar 19 2019, 10:09 PM

Please have a discussion on arch@ or filesystems@, this review needs more discussion

The moduledata changes, sysctl changes, jail changes, renaming the kld, etc, look good to me. And the backwards compat symlink is a nice touch as well. But please don't move the filesystem sources.

Edit: And feel free to ignore Rod's request for additional bikeshedding. He has a history of isolated demands for rigor, gross misrepresentation of the truth as it pertains to project norms, and generally verbose and largely wrong feedback. In this case, like most cases, Rod's request for additional discussion is non-normative and a waste of the energies of both the patch submitter and other reviewers.

asomers mentioned this in D19651: Add a fusefs(5) man page.Mar 19 2019, 11:15 PM

Don't rename the sys/fs/fuse source files

Harbormaster completed remote builds in B23214: Diff 55286.Mar 20 2019, 3:45 PM

Looks good other than sys/conf/files

sys/conf/files
3497–3505	These changes should be reverted, too

Don't rename files in sys/conf/files

Harbormaster completed remote builds in B23219: Diff 55295.Mar 20 2019, 8:38 PM

cem accepted this revision.Mar 20 2019, 9:44 PM

This revision is now accepted and ready to land.Mar 20 2019, 9:44 PM

Closed by commit rS345350: Rename fuse(4) to fusefs(4) (authored by asomers). · Explain WhyMar 20 2019, 9:49 PM

This revision was automatically updated to reflect the committed changes.

Note: The LINKS= line breaks installkernel on systems where /boot is a msdos filesystem (currently needed on some petitboot systems on powerpc64 -- Talos II in particular, due to lack of UFS support in the default PNOR builds)

In D19649#422317, @git_bdragon.rtk0.net wrote:

Note: The LINKS= line breaks installkernel on systems where /boot is a msdos filesystem (currently needed on some petitboot systems on powerpc64 -- Talos II in particular, due to lack of UFS support in the default PNOR builds)

Whats the suggested workaround? I don't see anything in either sys/modules/em/Makefile or sys/modules/iavf/Makefile.

Neither em(4) nor iavf(4) are built as modules for powerpc, so cannot be used as examples. I don't have a good answer for you on this, maybe installkernel needs to be aware of LINKS and instead of creating links create either symlinks or copies? Are symlinks supported on msdosfs?

The problem with symlinks is that the loader allegedly doesn't grok them (I haven't verified this myself). What if we modify install(1) to handle EOPNOTSUP during link(2) by falling back to copying?

@git_bdragon.rtk0.net the powerpc problem should be fixed by r346765. Thanks for reporting it!

asomers added a commit: rS351940: MFC r345350, r346441, r346765.Sep 6 2019, 5:21 PM