{zfs,spl}-module-parameters.5: fix nonexistent parameters

[nabijaczleweli]

Motivation and Context

#12125 (comment), I hope for this to be pretty uncontroversial

Description

Commit messages outline history for all of these

How Has This Been Tested?

Looked at it, I guess? I made the selection through the very scientific process of getting all the names in the zfs-m-p page and contrasting that to a modinfo listing from zfs-2.0.3-6~bpo10+1, which yielded a list short enough for manual extraxion, like spl-m-p.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Performance enhancement (non-breaking change which improves efficiency)
• Code cleanup (non-breaking change which makes code smaller or more readable)
• Breaking change (fix or feature that would cause existing functionality to change)
• Library ABI change (libzfs, libzfs_core, libnvpair, libuutil and libzfsbootenv)
• Documentation (a change to man pages or other documentation)

Checklist:

• My code follows the OpenZFS code style requirements.
• I have updated the documentation accordingly.
• I have read the contributing document.
• I have added tests to cover my changes.
• I have run the ZFS Test Suite with this change applied.
• All commit messages are properly formatted and contain Signed-off-by.

[allanjude]

zio_decompress_fail_fraction still exists, although it is not exposed as a module parameter, it is used by zdb to simulate failures.
Not sure if it makes sense to remove the documentation for it, but maybe it should explain that it is not a full module parameter or something.

Although there is a similar variable for decryption failure, that doesn't appear to be documented.

[nabijaczleweli]

It seems that I have to quote the messages

For spl-m-p:

Both were removed in 4fbdb10 ("remove kmem_cache module parameter KMC_EXPIRE_AGE")

For zfs-m-p:

zfs_arc_overflow_shift was never a parameter: ca0bf58 ("Illumos 5497 - lock contention on arcs_mtx") is the only result in git log -Soverflow_shift, and it wasn't exposed then, nor is it now

zfs_read_chunk_size was renamed to zfs_vnops_read_chunk_size in e53d678 ("Share zfs_fsync, zfs_read, zfs_write, et al between Linux and FreeBSD")

zio_decompress_fail_fraction was never a parameter: it was added in c3bd3fb ("OpenZFS 9403 - assertion failed in arc_buf_destroy()") as a developer aid for setting in zdb, but it's a dangerous test tunable and has no place in public documentation, (not to mention that it obviously doesn't work):

Although this did uncover a few low priority issues, this unfortuantely also causes ztest to ASSERT in many locations where the code is working correctly since it is designed to fail on IO errors. Developers can manually set this variable with the '-o' option to find and debug issues.

[behlendorf]

Since zio_decompress_fail_fraction was added to the documentation it makes me wonder if the original intent was to make it available as a tunable option. However, since it clearly never was and is very developer focused I agree let's drop from from the user visible docs.

[rlaager]

Before anyone requests a review from me (like with the other man page changes), I figured I'd take a look.

The only question I have is about spl_kmem_cache_obj_per_slab. That one seems to still exist:
https://github.com/openzfs/zfs/blob/master/module/os/linux/spl/spl-kmem-cache.c#L93

4fbdb10 says it removes spl_kmem_cache_obj_per_slab**_min**

This comment is the limit of my expertise in this area.

[nabijaczleweli]

How fortunate that this doesn't touch spl_kmem_cache_obj_per_slab, then

[rlaager]

How fortunate that this doesn't touch spl_kmem_cache_obj_per_slab, then

Hah, yep, you're right. My bad!