Linux kernel 7.1 wipes Intel 486 support, and it’s not the only legacy purge

Linux kernel 7.1 is out, and it pulls the plug on Intel 486 support for good. Not in theory. Not “eventually.” More than 140,000 lines of code have been chopped, with more slated for further deletion as the old architecture heads toward silicon heaven.

This is the exact change that previously almost got cancelled. Back in May 2025, The Register reported that kernel 6.15 would drop 486 support, but the removal was called off at the last minute. Now it has landed. In April, Penguin Emperor Linus Torvalds merged the big change described back then, and the result is a cleaner kernel with fewer ancient pathways to maintain.

Why should business and infrastructure folks care? Because Linux updates are not just developer toys. The 486 era is not just a fun historical artifact. In real deployments, legacy CPU support can show up in niche appliances, long-lived industrial systems, and “we forgot that this server is older than the company reorg” environments. When kernel compatibility assumptions get smaller, operational risk shifts. You either refresh hardware, validate that the remaining supported combinations cover your fleet, or you keep an old kernel line running longer than you planned.

This release does not stop with 486. Linux support for the Russian Baikal family of CPUs has been removed as well, and ancient bus mouse ports are gone. The Register previously reported on Baikal in 2014 and again in 2021, which gives you a timeline clue: maintainers were taking repeated looks over the years, and now the support is officially being retired. The bus mouse removal follows a similar pattern. These aren’t one-off deletions, they are the kernel doing its annual cleanup on the systems that benefit least from ongoing maintenance.

At the same time, kernel 7.1 delivers upgrades that point in a practical direction: filesystem tooling that supports recovery and maintenance. The release includes a new NTFS driver called NTFSplus. For now it’s optional, but South Korean filesystem developer Namjae Jeon has revived and rewritten the original read-only NTFS driver from the 1990s. The big deal is that it can now write to NTFS volumes as well as read them. It’s also been modernized to align with current kernel filesystem methods.

The ecosystem angle matters. NTFSplus comes alongside a new and improved version of additional ntfsprogs utilities, called ntfsprogs-plus. The utilities can repair some forms of NTFS corruption and errors, which is exactly the kind of capability that tends to get baked into Linux live rescue media. The Register specifically suspects quick adoption by SystemRescue, GParted Live, and Grml. If you run or rely on these tools in operational recovery plans, kernel 7.1 is now a candidate you should track, not ignore.

NTFSplus also threatens to reshuffle the NTFS driver landscape. It stands to completely replace the driver that Paragon Software donated back in 2020. It also seems likely that the old read-only NTFS driver will be removed too, since NTFSplus is based on that code. That combination means a potential future where one driver and one toolchain becomes the default path for NTFS on Linux, simplifying some maintenance while shifting what “known good” looks like for rescue and repair workflows.

Kernel 7.1 doesn’t just modernize NTFS. exFAT support has been improved as well, with pre-allocation of contiguous file space without zeroing blocks first. That can speed up file operations and reduce fragmentation, keeping storage media faster for longer. There are also improvements in ext4 and Btrfs handling.

Beyond storage, the kernel touches performance and security knobs that matter to platform teams. The swap memory subsystem has been overhauled, and should be faster. With RAM prices still high, there’s renewed interest in memory and cache compression tools, and the source suggests there’s “much more to do here,” which is a soft way of saying performance work on memory pathways is not finished. In networking and I/O, eBPF code can now handle io_uring scheduling, marrying the io_uring API introduced in 2022 with earlier eBPF functionality. There’s also a new sched feature: an extensible kernel scheduler has been merged, previously mentioned in Oracle Linux’s UEK-next kernel.

Security and virtualization updates land too. KVM virtualization security on Arm has been tightened, and access to PIDs in the /proc virtual filesystem has been improved. The CIFS network filesystem, aka SMB, now has explicit support for creating temporary files.

Intel’s FRED support is also now on by default. Intel FRED debuted way back in kernel 6.9, but kernel 7.1 turns it on, and it helps performance on AMD processors as well. Power management gets improved for both AMD and Intel chips, plus better battery-status reporting on Apple M1 and M2 laptops. Kernel Rust support now needs Rust 1.85.

So the story of kernel 7.1 is two stories at once: an aggressive cleanup that removes old CPU and legacy I/O support, and a set of practical upgrades that strengthen modern storage and scheduling workflows. For executives and boards, the strategic stake is clear: operating systems keep moving, and support churn is real. The risk is not the deletion by itself. The risk is surprise. If your infrastructure relies on the “we’ll deal with it later” class of legacy support, kernel 7.1 is a reminder that “later” can arrive faster than procurement cycles. Meanwhile, the filesystem improvements show that investing in Linux recovery capabilities can pay off operationally, especially when disks and filesystems eventually fail. Kernel 7.1 is the kind of release where both sides of the spreadsheet matter: cost of refresh now, versus cost of downtime later.