Linux Storage News Archives

OpenZFS 2.4-rc1 was released a few days ago with faster encryption performance using AVX2 and other enhancements. For those just looking for bug fixes and expanded Linux kernel compatibility, OpenZFS 2.3.4 is out today as the newest stable point release.

Linux IO_uring and block subsystem maintainer Jens Axboe this week queued the patch series for multi-shot support with provided buffers. Thus the feature should go in for Linux 6.18 as another nice enhancement to the wonderful IO_uring kernel innovation.

The first release candidate of OpenZFS 2.4 is now available for testing of this ZFS file-system implementation for Linux and FreeBSD systems.

Barring any issues from coming up in the near future, it looks like DM-PCACHE will be submitted for the Linux 6.18 cycle later in the year as a high throughput, low latency cache for DAX-capable persistent memory devices.

It was nearly one year ago in the Linux 6.13 kernel that the ReiserFS file-system was dropped from the mainline kernel after having been deprecated in 2022. That dropped 32.8k lines of code from the Linux kernel but some documentation remnants of ReiserFS were mistakenly left in but now in the process of dropping those remnants for the defunct file-system.

For those making use of the Network File System (NFS), the in-development Linux 6.17 kernel is bringing some nice improvements both for the NFS server and client code.

Amid the ongoing discussion over what will happen too Bcachefs in the mainline Linux kernel, an interesting anecdote around Btrfs was mentioned.

We still don't know what's going to happen for Bcachefs in the Linux 6.17 kernel even with the merge window set to end on Sunday with the Linux 6.17-rc1 release. Linus Torvalds commented over one month ago that they would be parting ways for Linux 6.17. At the start of the Linux 6.17 merge window a Bcachefs pull request was submitted but nearly two weeks later it's still not been pulled and Linus Torvalds hasn't commented on the matter.

On Linux 6.15.3+ there have been increased reports of log tree corruption being hit by users of the Btrfs file-system. Fortunately, a fix has now been submitted for Linux 6.17 Git and then for back-porting to the recent stable kernel versions.

Following yesterday's F2FS pull request, the exFAT file-system updates were sent out and since merged for the ongoing Linux 6.17 kernel merge window.

Sent out today for the Linux 6.17 merge window and already merged hours later were the Flash-Friendly File-System (F2FS) feature updates.

Earlier this year there was talk of Linux potentially dropping its Apple HFS and HFS+ file-system drivers as they had been orphaned for a decade and proving to be a maintenance burden. Following that some developers stepped up to better maintain the code for HFS and HFS+ file-system support. In Linux 6.17 we are seeing some of the fruits of that work.

The EXT4 file-system enhancements for Linux 6.17 were merged on Thursday and bring better scalability to the block allocation code as well as fixing the file-system's large folios support. The scalability work can show some wild gains in select areas.

Merged already for the in-development Linux 6.17 kernel were the many block subsystem and IO_uring changes for enhancing I/O on Linux as we roll toward the H2'2025 Linux distribution releases.

Along with the better handling of multi-device file-systems such as Btrfs' native RAID capabilities and now allowing more efficient writing of zeroes to modern storage devices, the number of VFS pull requests for Linux 6.17 also added some other extra goodies.

In addition to the VFS changes merrged yesterday for allowing multi-device file-systems to better cope with losing a disk, another notable change as part of the VFS pull requests for Linux 6.17 allows more efficiently zeroing out a range on modern NVMe SSDs or SCSI drives.

Merged on Monday were the EROFS file-system updates for Linux 6.17. EROFS continues to be a common read-only file-system choice for some mobile/embedded devices as well as container use-cases.

The VFS changes were merged a short time ago to the in-development Linux 6.17 kernel. Among the notable changes there is a patch that will allow file-systems like Btrfs and Bcachefs to better handle losing a disk in their built-in RAID/multi-device capabilities.

With the Linux 6.17 merge window now open, Bcachefs file-system lead developer Kent Overstreet has submitted his planned changes for this next kernel version. But we await to see how Linus Torvalds will respond.

One of the nice Linux kernel accomplishments during the pandemic was getting the NTFS3 driver upstreamed for that modern NTFS file-system read/write driver developed by Paragon Software. In recent times that NTFS3 driver has been seeing occasional fixes and for the Linux 6.17 kernel -- and perhaps then back-ported to existing kernels -- are some notable fixes for those relying on drives formatted with this Microsoft file-system.

Ahead of the Linux 6.16 stable kernel expected to be released on Sunday, some last minute Bcachefs file-system fixes join various other kernel regression/bug fixes landing today in Git.

With the Linux 6.16 stable kernel potentially being released this coming Sunday, the Linux 6.17 merge window is quickly approaching. Btrfs maintainer and SUSE engineer David Sterba sent out the Btrfs pull request already of new feature code for this copy-on-write file-system in Linux 6.17.

The EROFS read-only file-system ahead of the upcoming Linux 6.17 merge window has been working on metadata compression support to allow for even smaller container image sizes but at the cost of higher I/O latency.

As we await to see what Linus Torvalds will end up doing about the Bcachefs file-system come Linux 6.17, for the ongoing Linux 6.16 cycle he continues to honor the Bcachefs pull requests containing fixes.

New feature patches posted for review today on the Linux kernel mailing list are working to implement the NVMe specification's Controller Data Queue (CDQ) functionality within the NVMe storage driver.

Ahead of tomorrow's Linux 6.16-rc6 kernel release a number of Bcachefs file-system fixes were merged.

Linux block maintainer Jens Axboe queued up a patch this week to drop the pktcdvd driver from the mainline kernel, which is expected to be submitted for the upcoming Linux 6.17 cycle. The pktcdvd driver has been in the kernel for over two decades since the Linux 2.6 days for packet-writing CD/DVD support albeit is hardly useful in today's world.

A handful of Bcachefs file-system fixes were submitted yesterday to Linus Torvalds for merging ahead of the Linux 6.16-rc5 release due out on Sunday. Today they were merged.

It looks like Linux 6.17 could end up enabling experimental support for large data folios that could help with bringing some performance improvements under real-world workloads for this copy-on-write file-system.

Last week was a Bcachefs pull request consisting of fixes and a new "journal_rewind" feature to aide as a disaster recovery tool for the file-system. But with that code being submitted as part of the ongoing Linux 6.16 release candidates, it drew criticism from Linus Torvalds and other kernel developers. However, one week later and that discussion having subsided a few days ago, Linus Torvalds ended up merging all of the code. But there is a concerning warning for the future of Bcachefs in the mainline Linux kernel.

Today marks the debut of mmc-utils 1.0 as the inaugural release for this de facto solution for managing MMC storage devices under Linux.

There is some tension on the Linux kernel mailing list with some late Bcachefs feature work sent in as part of "fixes" for the ongoing Linux 6.16 kernel cycle. Established rules aim for only new feature code to be introduced during the kernel merge windows, which ended nearly two weeks ago for Linux 6.16, but Bcachefs wanting to be exempt to continue to allow new feature code to still land for the cycle in the name of data safety.

Following last week's release of OpenZFS 2.2.8, OpenZFS 2.3.3 is now available as the newest point release of this current stable series for this open-source ZFS file-system implementation on Linux and FreeBSD systems.

Another round of Bcachefs file-system fixes were submitted and merged this week for the ongoing Linux 6.16 cycle, including the ability to auto-fix more file-system check "fsck" errors.

While OpenZFS 2.3 has been stable for several months, for those still relying on the OpenZFS 2.2 series there is a new stable point release.

In addition to the NFS server changes for Linux 6.16 with now supporting larger I/O block sizes, the Network File-System (NFS) client changes were merged this week for this next kernel version.

An initial patch series sent back out in April proposed PCACHE as a persistent memory cache for block devices. PCACHE was born out of the CXL block device driver and brings some benefits over the likes of BCache and dm-writecache.

Last week many Bcachefs performance optimizations, recovery work, and enhanced error messages were merged at the start of the Linux 6.16 merge window. Now ahead of the Linux 6.16-rc1 release coming on Sunday to cap off the merge window, a second round of Bcachefs enhancements and fixes were merged.

SquashFS-Tools 4.7 is out today as a big feature update to the user-space utilities for creating/modifying/extracting SquashFS read-only file-system images. SquashFS 4.7 delivers some big performance improvements and other nice enhancements.

Merged last week for the Linux 6.16 kernel coming later this summer were the many block subsystem updates as well as the IO_uring feature updates.

The FUSE improvements have been merged for the Linux 6.16 in enhancing the capabilities for file-systems implemented in user-space.

For those running an Network File System "NFS" server, the upcoming Linux 6.16 kernel will allow optionally enabling a larger payload size that may yield better performance. Eventually the default payload size may be increased if all goes well from user feedback and testing.

Following the exciting EXT4 performance work, XFS atomic writes, and other exciting file-system pull requests submitted for the ongoing Linux 6.16 merge window, the Flash-Friendly File-System (F2FS) changes have been submitted and merged for this next kernel version.

The MMC subsystem feature changes have been merged for the Linux 6.16 kernel. Interestingly and surprisingly, it's not until now that the Linux kernel has properly supported the graceful host removal for eMMC and SD cards.

Ted Ts'o sent out the EXT4 file-system changes today for the Linux 6.16 kernel. While EXT4 may not see as much code churn these days given its mature state compared to say Btrfs and Bcachefs, with Linux 6.16 are some tantalizing performance improvements.

Google engineer Eric Biggers took time away from all his impressive crypto performance optimizations to the Linux kernel for modern Intel and AMD CPUs to spend time getting support for hardware-wrapped inline encryption keys into the mainline kernel. Google's Android kernel has been carrying this functionality for several years to help enhance security and will now be found in the mainline kernel too for capable platforms.

Building off the atomic write preparations that landed in prior Linux kernel versions, merged for the Linux 6.16 cycle is atomic write support for the XFS file-system.

Among the many VFS pull requests merged today by Linus Torvalds for the now-open Linux 6.16 merge window is support by the power subsystem for being able to initiate file-system freeze/thaw events as part of the system suspend and resume cycle.

Bcachefs lead developer Kent Overstreet has already sent out the plethora of feature updates destined for the now-open Linux 6.16 kernel cycle. There has been a lot of work going on for Bcachefs that is ready for merging in Linux 6.16 with several performance improvements, continued self-healing / recovery work, improved error messages, and more for this copy-on-write file-system.

ConfigFS is seeing a new maintainer step up to manage this RAM-based file-system. With Andreas Hindborg taking over maintainership of ConfigFS he's also landing patches authored by him for adding Rust programming language support to this file-system.