Linux Kernel 7.1 Sends Intel 486 Support To Silicon Heaven
Arthur T Knackerbracket writes:
https://www.zdnet.com/article/say-goodbye-to-486-processors-linux-7-1-lands/
On his way to Mumbai for the Open Source Summit India, Linus Torvalds announced thelatest Linux kernel: 7.1. This new version comes with a brandnew inkernel,Microsoft's New Technology File System (NTFS) implementation, Intel's Flexible Return and Event Delivery (FRED) enabled by default, and a purge of aging code and hardware support, including the end of the road for 486 support.
The new kernel arrives just months after Linux 7.0 debuted with major networking and filesystem changes. Version 7.1 continues the trend of tightening Linux's hardware focus while improving performance and security.
The headline change for most people in Linux 7.1 is its new native NTFS driver. Many Linux users, whether they like it or not, must deal with Microsoft file systems, and that's where this driver comes in. It replaces both the old, dusty NTFS3G FUSE driver in many setups and the Paragoncontributed NTFS3 kernel driver that has had a bumpy history with datacorruption reports. Indeed, Torvalds himself calls this new driver an "NTFS resurrection."
The new code is built on Linux's contemporary filesystem infrastructure, using iomap and folios instead of older buffer_headcentric paths. The new and improved NTFS support was designed from the outset for robust readwrite support, better error handling, and more predictable behavior under heavy parallel I/O.
According to the new NTFS developer, Namjae Jeon, while the new NTFS driver shows only modest singlethreaded write gains, its improvements are far more impressive when multiple threads pound on the same NTFS volume. There, multithreaded writes can be 35-110% faster than earlier drivers, while mounting a 4TB NTFS volume is reported to be roughly four times faster.
For users who regularly shuttle data between Windows and Linux via external drives or dualboot setups, this change should make NTFS feel much more like a firstclass citizen on Linux desktopsand laptops.
On the CPU side, Linux 7.1 flips the switch on Intel's FRED, enabling it by default on supported Intel platforms. FRED is a new hardware mechanism for handling entries and exits in privileged modes, such as interrupts, exceptions, and system calls. FRED's goal is to simplify controlflow transitions and reduce the reliance on legacy entry stacks that have grown increasingly complex and insecure over the years.
By moving to FRED, the kernel gets a cleaner separation between user and kernel control flows. This approach not only shrinks some overhead in highfrequency event handling but also promises a tighter security story around those transitions.
For now, the benefits will mostly be felt on recent Intel client and server hardware. However, as more vendors select similar mechanisms, 7.1's work positions Linux to take advantage of future architectural features in the space.
Linux 7.1 also brings a series of security improvements. This work starts with support for Intel's Linear Address Space Separation (LASS). LASS constrains how code can access different regions of the linear address space, making certain classes of memorycorruption and controlflow attacks harder to exploit by enforcing stronger separation between code and data regions. In tandem with FRED, LASS underlines an architectural shift toward hardening the boundary between userland and kernel, as well as between different types of inkernel objects.
Under the hood, the crypto subsystem gets a rework that enables more optimizations by default, which should pay off everywhere encryption and hashing are hot paths: TLS stacks, VPNs, encrypted filesystems, and distributed storage.
On the CPU front, there is further enablement for AMD's upcoming Zen 6 processors, with new IDs, errata workarounds, and tuning hooks landing in 7.1, ensuring the kernel is ready for the next generation of EPYC and Ryzen parts when they hit the market.
While 7.1 delivers clear wins for modern systems, it is also unambiguous about where Linux is no longer willing to go. This release continuesthe process of ripping out support for Intel's 486class processors and other early x86 variants. Kernel maintainers have been telegraphing for some time. Mainstream distributions had long since moved their baselines to at least i586 or x8664. To my knowledge, no major Linux distro currently supports 486 processors.
Beyond CPUs, the codebase sheds over 140,000 lines of legacy code. This includes a grab bag of obsolete network and PCMCIA drivers, as well as the removal of Baikal CPU support. The rationale is twofold: Reduce the attack surface exposed by ancient, barely tested code paths, and ease longterm maintenance by freeing developers from the need to preserve behaviors for hardware that has vanished from production. Retrocomputing enthusiasts will still be able to run older kernels, but the line between "supported" and "museum piece" is getting sharper.
At first glance, Linux 7.1 is just a point release, but its mix of a new NTFS driver, FRED and LASS enablement, aggressive legacy cleanup, and continued hardware support will be important for both desktops and data centers.
For everyday users, the most obvious impact will likely be smoother, safer interaction with NTFSformatted drives. For operators and OEMs, the story is more about security posture, platform support, and getting ready for the next wave of Intel and AMD silicon.
Read more of this story at SoylentNews.