- devices de-duplication (e.g. lsblk --dedup WWN); this is useful for example on systems with multi-path devices where the same device is accessible by more ways
- M:N relationships without repeating devices (e.g. lsblk --merge); this feature is implemented by additional tree graph.
$ lsblk --merge NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT loop0 7:0 0 955.7M 0 loop ┌┈▶ ├─test-thin-metadata 253:0 0 2M 0 dm └┬▶ └─test-thin-data 253:1 0 953.7M 0 dm ┌┈▶ └┈┈test-thin-pool 253:2 0 953.7M 0 dm ┆ └─test-thin 253:3 0 190.8M 0 dm └┬▶ loop1 7:1 0 190.8M 0 loop └┈┈┈┈┈test-thin-extsnap 253:4 0 190.8M 0 dm
The same situation with classic lsblk output:
$ lsblk NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT loop0 7:0 0 955.7M 0 loop ├─test-thin-metadata 253:0 0 2M 0 dm │ └─test-thin-pool 253:2 0 953.7M 0 dm │ ├─test-thin 253:3 0 190.8M 0 dm │ └─test-thin-extsnap 253:4 0 190.8M 0 dm └─test-thin-data 253:1 0 953.7M 0 dm └─test-thin-pool 253:2 0 953.7M 0 dm ├─test-thin 253:3 0 190.8M 0 dm └─test-thin-extsnap 253:4 0 190.8M 0 dm loop1 7:1 0 190.8M 0 loop └─test-thin-extsnap 253:4 0 190.8M 0 dm
You can see that lsblk repeating test-thin and test-thin-extsnap, and loop1 seems unrelated for the setup at first glance.
The repeating of the block devices in the output is more annoying for multi-path devices. See another example in my previous blog: https://karelzak.blogspot.com/2018/11/lsblk-merge.html
I guess that FUSE user will be very happy with v2.34; finally the command umount(8) supports non-root user unmount for FUSE mounts. The requirement is FUSE specific user_id=
The command mount(8) now allows to use "--all -o remount" to remount all filesystems with specified or fstab options. It's possible to use filters (-t and -O).
The new command hardlink has been merged to util-linux. The command consolidates duplicate files via hardlinks. The current implementation is from Fedora, but in the next versions we will reuse also code from Debian fork to create one widely usable implementation.
The command lscpu(1) now prints more information about your CPU, for example 'Frequency boost' and 'Vulnerability' fields. The caches calculation has been modified to print summary from all system caches rather than per code numbers; and new command line option --caches lists details about CPU caches.
... and many another small changes :-) See complete release notes at kernel.org: https://mirrors.edge.kernel.org/pub/linux/utils/util-linux/v2.34/v2.34-ReleaseNotes
Thanks to Sami Kerola, Stanislav Brabec, Ruediger Meier and others!