Linux Storage News Archives

Kent Overstreet announced the release today of Bcachefs-Tools 1.38.6 as the user-space tools built around the Bcachefs copy-on-write file-system. There are a few new features and a lot of performance work in v1.38.6 without bringing any on-disk format breakage.

Linux 7.2 continues seeing a fair amount of storage-related changes from file-systems to the block device code itself, software RAID, the wonderful IO_uring interface, and more. Here is some of the latest feature work that has been merged for Linux 7.2.

There is no shortage of different file-systems available for Linux. New file-systems continue to come about in the open-source world but ultimately many of them end up not being well maintained or having very limited users and not necessarily innovating enough to make them worthwhile over other alternatives. Given the continued increase in file-systems looking to get into the Linux kernel, such as FTRFS and VMUFAT being some of the most recent and then even having multiple NTFS drivers for Linux, there is now documentation in place to formally lay out criteria for new file-systems to be accepted.

The Btrfs file-system feature updates have been merged for the Linux 7.2 kernel with a few noteworthy changes for this copy-on-write file-system.

A patch series merged for the Linux 7.2 kernel addresses a race condition that can occur when a container is exiting yielding "VFS: Busy inodes after unmount" messages and a possible user-after-free condition. But the patch series also goes further and delivers a very nice optimization to lower the container unmounting latency for environments with heavy I/O load.

Merged as part of the many VFS changes for Linux 7.2 is the new OPENAT2_REGULAR flag for the openat2 system call. This can be used to limit programs to only open regular file-systems and avoid accidentally or intentionally opening up device files or other non-conventional data files on the file-system.

The XFS file-system updates for the Linux 7.2 kernel aren't too notable with the exception of its zone allocator being promoted from behind its previously-experimental flag.

In addition to the surprising impact of /proc/filesystems read optimizations for Linux 7.2, another one of the VFS pull requests for this next kernel version is delivering some nice improvements for EXT4 and XFS around IOmap, the framework that maps file data offsets in memory to their physical locations on storage.

Reading /proc/filesystems for obtaining a list of file-systems supported by the running kernel is done frequently on Linux. Namely due to being read by the SELinux library (libselinux), reading of /proc/filesystems is done more often than one would typically expect and now the Linux 7.2 kernel is optimizing for it to yield much better performance.

Newly-merged code for the in-development Linux 7.2 kernel will now expose the case-folding (case insensitive) behavior of local file-systems so that Linux file servers and others can properly report the actual behavior rather than guessing if case-folding is actually used/supported.

A few days back I wrote about Google's Eric Biggers spearheading an AVX-512 implementation of xor_gen() as the Linux kernel function used for generating and validating parity blocks such as for RAID5/RAID6. That initial implementation was yielding up to 41% better performance while a new implementation has now been posted for scoring some additional victories.

OpenZFS 2.4.3 is out today as the newest stable point release to this open-source ZFS file-system implementation as well as point releases for the OpenZFS 2.3 and 2.2 series too.

One of the exciting additions to the Linux 7.1 kernel is the introduction of the new NTFS file-system kernel driver. While in good shape already and proving advantageous over other NTFS open-source driver options, one of the initial limitations on it is around Windows native symbolic link handling but that is now in the process of being resolved.

Linux cryptography subsystem expert Eric Biggers Eric Biggers of Google worked on some pretty nice Intel/AMD x86_64 optimizations over the years. Especially around AVX-512 optimizations within the Linux kernel's crypto code has been one of his many nice improvements to the kernel in recent times. Today he's out with another enticing AVX-512 optimization and this time it's for the software RAID code.

An interesting quandary has arose on the Linux kernel mailing list over maintainership of old, unmaintained code within the Linux kernel. Someone has stepped up to maintain an old, very rare file-system driver but admittedly doesn't even use it and just submitted basic fixes. Or is it just better removing that old code?

A change coming on the way for the upcoming Linux 7.2 kernel cycle is yielding a significant improvement to the direct I/O write performance. While a big gain, technically it's a regression fix after a change mistakenly dropped the behavior several years ago.

The past few Linux kernel cycles there has been experimental support for large folios with Btrfs while for Linux 7.2 it looks like this modern file-system will be taking things further with huge folios.

Among the VFS patches queued into "-next" branches ahead of next month's Linux 7.2 merge window is the code for introducing the new OPENAT2_REGULAR flag for the openat2 system call.

For those making use of OpenZFS on Linux or FreeBSD, OpenZFS 2.4.2 is out today as the newest stable release of this ZFS file-system implementation.

Introduced in Linux 7.0 was FSERROR as generic I/O error reporting infrastructure. Linux to that point had no standardized mechanism for reporting metadata corruption or file I/O errors to user-space with each file-system doing its own thing. The Flash-Friendly File-System (F2FS) is now the latest Linux file-system preparing for FSERROR usage.

Following a presentation at last week's Linux storage, file-system, memory management and BPF summit (LSFMM) in Croatia where Linux I/O overhead compared to the Storage Performance Development Kit (SPDK) was presented, Jens Axboe was motivated to pursue some new Linux kernel optimizations for greater per-core I/O performance. This lead IO_uring developer and Linux block maintainer has managed to achieve around a 60% increase to the per-core I/O performance with his latest patches.

Queued for merging as part of the DeviceMapper changes for the upcoming Linux 7.2 kernel cycle is the new dm-inlinecrypt target for leveraging inline block device encryption.

The growing number of file-systems within the Linux kernel source tree is causing an ongoing burden for upstream developers maintaining the virtual file-system (VFS) code around it and associated code. As a result of the continuing rise of new file-systems being proposed for the Linux kernel, documentation is being introduced to establish clear guidelines for getting new file-systems accepted into the mainline kernel.

One of the most prominent changes with the upcoming Linux 7.1 kernel release is the introduction of the new NTFS driver in the Linux 7.1 kernel. This new driver provides more features and better performance than the Paragon NTFS3 driver that's been in the kernel the past few years and far better off than the original NTFS read-only driver that previously was in the kernel and for which this new driver is based. Needless to say it's also a big improvement over the NTFS-3G user-space FUSE driver too.

Miklos Szeredi of Red Hat has been developing the FUSEX file-system as an extended/experimental area for File-System in User-Space "FUSE" development.

It's crazy to realize it has been ten years already since Red Hat abandoned their Btrfs plans for Red Hat Enterprise Linux and dropped it, which was a technology preview feature since RHEL6. In its place Red Hat engineers began developing Stratis for next-gen Linux storage with ZFS/Btrfs-like features but instead building atop XFS, LUKS, Device Mapper, and Clevis. After a while since the last major release, Stratis Storage 3.9 released today.

With the Linux 7.1-rc1 kernel release due out tomorrow to cap off the Linux 7.1 merge window, one of the most notable additions this cycle is the introduction of the new NTFS driver that aims to provide better performance and more modern features than the existing NTFS3 in-kernel driver that was originally contributed by Paragon Software.

The Flash Friendly File-System (F2FS) updates have been merged for the ongoing Linux 7.1 merge window that will wrap up on Sunday. This follows earlier merges for the XFS and EXT4 drivers too.

Coming today as a big surprise -- one week after the new NTFS file-system driver was merged for Linux 7.1 and separately the existing NTFS3 kernel driver seeing some fixes -- is a new release of the NTFS-3G driver providing a FUSE-based user-space driver for NTFS on Linux and other platforms.

Last week saw the "NTFS resurrection" as Linux Torvalds put it with the new/overhauled NTFS driver having been merged for Linux 7.1. Even still, the NTFS3 driver that was contributed a few years ago by Paragon Software remains in the mainline kernel and today were some fixes/improvements merged for that existing driver.

Redb is one of the open-source, embed-friendly key-value databases written in the Rust programming language. Redb is ACID-compliant while known for being high performance and with its new Redb 4.1 release is even faster thanks to some improvements authored by Claude (AI).

It's pretty rare nowadays seeing any real changes to the JFS file-system on Linux when there are multiple far superior solutions available. But in any event, the JFS file-system driver has seen a few fixes in Linux 7.1.

The block subsystem and IO_uring changes were merged this week for Linux 7.1 in continuing to enhance Linux storage capabilities.

As a very exciting follow-up to the recent article around the new NTFS driver being submitted for Linux 7.1 to address the shortcomings of the current Paragon NTFS3 driver and the prior read-only NTFS kernel driver, that work has been merged!

Making today very exciting in Linux 7.1 merge window land was a pull request being sent out for introducing the new, modern NTFS file-system driver. Linus Torvalds has yet to comment if he's going to merge the new driver but it looks like it's ready for providing a better Linux NTFS experience over the current NTFS3 driver that was upstreamed by Paragon Software a few years ago and hasn't seen too much feature progress.

The exFAT file-system changes have landed for the in-development Linux 7.1 kernel.

Merged yesterday for the Linux 7.1 kernel is overhauling the T10 PI code for generating and verifying data integrity information. In turn the new code is cleaner while also allowing for better read storage performance.

Sent out today was an initial patch series for comment on introducing the FTRFS file-system. The FTRFS proposal is more interesting than last week's VMUFAT file-system proposal.

Among the early pull requests sent out to Linus Torvalds even before the Linux 7.0 kernel officially released on Sunday were the Btrfs file-system updates. This feature-packed CoW file-system is seeing more performance optimizations for Linux 7.1 as well as its shutdown ioctl feature no longer being experimental and a variety of fixes.

Among the new VFS features expected to land for the upcoming Linux 7.1 merge window is FSMOUNT_NAMESPACE.

The newest Linux file-system driver proposed for the kernel is... VMUFAT.

Ahead of the Linux 7.0-rc6 kernel due to be released later today, quite a number of EXT4 file-system fixes were sent out this morning.

Developers behind the Distributed Replicated Block Device "DRBD" for mirroring block devices between multiple host systems are working to resync the upstream Linux kernel DRBD support with the out-of-tree DRBD code they have been maintaining for the past ~15 years out-of-sync. It's a big undertaking but they have begun staging patches for review and testing to get this massive set of changes up to par for mainline.

With sched_ext there is support for BPF-based CPU scheduling policies for the Linux kernel while now a new initiative is working on BPF-based I/O schedulers.

The mdadm utility for managing software RAID on Linux systems is out with a new release that adds new features while addressing some recent boot failure issues that were reported.

Kent Overstreet today released Bcachefs 1.37 as the newest feature release to this out-of-tree file-system driver and user-space tooling for this next-gen, copy-on-write file-system.

The Linux ZRAM module for creating compressed block devices in RAM could be on the edge of a nice I/O compression performance boost.

For those making use of Microsoft's exFAT file-system under Linux, tagged today was exfatprogs 1.3.2 as the newest update to these open-source user-space programs for going along with the Linux kernel's exFAT file-system driver.

Following the big OpenZFS 2.4 release back in December, OpenZFS 2.4.1 was released overnight to ship support for the latest Linux 6.19 stable kernel plus a variety of different bug fixes.

We haven't heard much about eCryptfs in recent years for that stackable in-tree Linux file-system providing per-directory encryption support. The FSCRYPT framework has shown its strong capabilities in recent years with various file-systems, Canonical hasn't been pursuing its user home directory encryption like it did years ago for the Ubuntu desktop, and full disk encryption is the most secure approach for ensuring data security on your system. But to some surprise with Linux 7.0 there are the most patches to eCryptfs that we have seen in a while.