Linux Kernel 6.14 Released: New Features, Performance Improvements, and Hardware Support

The Linux kernel development community never slows down, and the release of Linux kernel 6.14 is another testament to the relentless pace of open source innovation. This release brings meaningful improvements across the board: better hardware support for the latest AMD and Intel silicon, continued expansion of Rust-written subsystems, significant networking stack enhancements, and scheduler refinements that real-world workloads will notice. Whether you run a gaming desktop, a home lab server, or a fleet of cloud VMs, kernel 6.14 has something worth your attention.

What Changed Since Kernel 6.13

Kernel 6.13 was a notable release that introduced the EEVDF scheduler improvements and better memory-tier awareness. Kernel 6.14 builds on that foundation rather than upending it. The merge window for 6.14 pulled in roughly 14,000 commits from over 2,000 contributors — numbers that illustrate the breadth of ongoing development across every subsystem.

The headline themes for 6.14 are:

• Expanded Rust language support within the kernel itself
• New and updated drivers for AMD RDNA 4 GPUs and Intel Lunar Lake/Arrow Lake platforms
• ARM and RISC-V platform additions for embedded and server workloads
• Networking improvements including io_uring send/recv zero-copy paths
• Scheduler and memory management refinements
• Improved energy-aware scheduling for heterogeneous CPU topologies

Rust in the Kernel: Continued Growth

One of the most strategically significant aspects of modern kernel development is the gradual adoption of Rust as a second implementation language alongside C. Kernel 6.14 continues this trend with new Rust abstractions for several subsystems.

The rust/kernel crate has grown to include safer wrappers around workqueue operations, improved file-system abstractions, and better integration with the driver model. The Nova GPU driver, a Rust-written open source driver for NVIDIA hardware developed as a clean-room implementation, has received further scaffolding commits in this cycle. It is not yet production-ready, but its inclusion signals where the kernel community is heading.

Apple Silicon platform support, largely driven by the Asahi Linux project, also received Rust-based driver contributions in this cycle, particularly around the GPU command submission path.

AMD Hardware Support

AMD users are among the biggest beneficiaries of kernel 6.14. The AMDGPU driver now includes initial support for RDNA 4 architecture GPUs (the Radeon RX 9000 series), bringing accelerated display output and compute functionality to these newer cards out of the box on updated kernels.

The AMD P-state driver — responsible for managing CPU frequency scaling on Ryzen and EPYC processors — has received tuning for Zen 5 mobile and desktop parts. The amd-pstate driver’s active mode, which allows the firmware to collaborate with the OS scheduler for frequency decisions, is now enabled by default on supported hardware. This results in measurably better power consumption under mixed workloads.

On the storage side, improvements to the NVMe driver benefit AMD platforms with onboard NVMe controllers, reducing interrupt latency under queue-depth saturation scenarios.

Intel Platform Updates

Intel’s Lunar Lake and Arrow Lake platforms, representing the Core Ultra 200 generation, receive improved thermal and power management support in kernel 6.14. The intel_pstate driver has been updated with new turbo policy knobs, and the Intel uncore frequency driver now exposes more fine-grained control over fabric and memory controller clocks via sysfs.

The Intel Xe graphics driver, which handles Arc and integrated graphics for modern Intel platforms, continues its rapid maturation. Kernel 6.14 includes fixes for display flickering on high-refresh panels and improved Vulkan compute performance via the DRM subsystem updates.

Intel’s Trust Domain Extensions (TDX), a confidential computing technology for protecting VMs from hypervisor-level attackers, receives stability fixes and improved attestation support, making it more viable for production confidential computing deployments on cloud infrastructure.

ARM and RISC-V Additions

The ARM ecosystem remains the most diverse corner of the kernel, supporting everything from Raspberry Pi boards to enterprise-grade Ampere Altra servers. Kernel 6.14 adds device tree and driver support for several new SoCs including newer MediaTek Dimensity platforms (relevant for ARM-based Linux laptops) and Qualcomm Snapdragon X Elite improvements that benefit the growing category of ARM-native Linux laptops.

RISC-V continues its steady march toward feature parity with x86 and ARM. Kernel 6.14 brings support for RISC-V vector extension 1.0 optimizations in the crypto layer, meaning AES and SHA operations can take advantage of vector hardware on capable RISC-V processors, which is important for the emerging class of RISC-V server SoCs.

Networking Stack Improvements

Networking is one of the most actively developed kernel subsystems, and 6.14 brings several noteworthy improvements:

• io_uring zero-copy networking: The zero-copy send and receive paths for io_uring have been stabilized and extended. Applications using io_uring for high-throughput networking (databases, proxies, messaging systems) will see reduced CPU overhead and latency.
• XDP improvements: The eXpress Data Path subsystem, used for high-performance packet processing in eBPF programs, receives new helper functions and improved multi-buffer support.
• TCP congestion improvements: Refinements to BBRv3 support and improvements to the CUBIC congestion control algorithm under high-RTT conditions.
• Wireless improvements: The mac80211 subsystem and several vendor Wi-Fi drivers receive fixes for WPA3-Enterprise authentication edge cases and improved 6GHz band support for Wi-Fi 6E/7 adapters.

Scheduler and Memory Management

The EEVDF (Earliest Eligible Virtual Deadline First) scheduler introduced in earlier kernels continues to be refined. Kernel 6.14 improves EEVDF behavior under oversubscribed CPU conditions — when more runnable tasks exist than available CPUs — reducing tail latencies for interactive workloads running alongside CPU-intensive background jobs.

Memory management sees improvements to the MGLRU (Multi-Generational LRU) page reclaim algorithm. The aging and eviction heuristics have been tuned based on real workload data from major cloud providers, resulting in better behavior on systems with large working sets that exceed available RAM.

Transparent Huge Pages (THP) receive an important addition: the ability to set THP size hints on a per-VMA basis, giving applications like databases and JVMs more control over when and how the kernel uses huge pages for their memory regions.

How to Check Your Current Kernel Version

Before upgrading, verify what kernel version you are currently running:

uname -r

For more detailed information including build date and architecture:

uname -a

You can also inspect the kernel version from the proc filesystem:

cat /proc/version

ubuntu">How to Upgrade to Kernel 6.14 on Ubuntu

Ubuntu LTS releases typically receive newer kernels via the Hardware Enablement (HWE) stack, while Ubuntu’s non-LTS releases track closer to upstream. For Ubuntu 24.04 and 24.10 users, the mainline kernel PPA offers the fastest path to 6.14:

# First, update your system
sudo apt update && sudo apt upgrade -y

# Install the mainline kernel tool (optional but convenient)
sudo apt install linux-mainline

# Alternatively, download the .deb packages directly from the mainline PPA
# https://kernel.ubuntu.com/mainline/v6.14/

# For a specific architecture (amd64 example)
cd /tmp
wget https://kernel.ubuntu.com/mainline/v6.14/amd64/linux-headers-6.14.0-060140_6.14.0-060140.202503_all.deb
wget https://kernel.ubuntu.com/mainline/v6.14/amd64/linux-headers-6.14.0-060140-generic_6.14.0-060140.202503_amd64.deb
wget https://kernel.ubuntu.com/mainline/v6.14/amd64/linux-image-unsigned-6.14.0-060140-generic_6.14.0-060140.202503_amd64.deb
wget https://kernel.ubuntu.com/mainline/v6.14/amd64/linux-modules-6.14.0-060140-generic_6.14.0-060140.202503_amd64.deb

sudo dpkg -i *.deb
sudo reboot

After rebooting, confirm the new kernel is active:

uname -r
# Expected output: 6.14.0-060140-generic

fedora">How to Upgrade to Kernel 6.14 on Fedora

Fedora tracks upstream kernels closely and is typically one of the first major distributions to ship new kernel versions. On Fedora 41 or 42, a standard system update is often sufficient:

sudo dnf upgrade --refresh
sudo reboot

To check what kernel version is available in the repository before upgrading:

dnf info kernel | grep Version

Fedora also makes it easy to install a specific kernel version while keeping the old one available at boot:

sudo dnf install kernel-6.14.0
# Reboot and select the new kernel from GRUB if needed

How to Upgrade to Kernel 6.14 on Arch Linux

Arch Linux is a rolling release distribution and typically ships the latest stable kernel within days of upstream release. Upgrading on Arch is the same as any other package update:

sudo pacman -Syu

If you want to run the mainline (release candidate) kernel for testing, the AUR provides packages:

# Using yay as the AUR helper
yay -S linux-mainline

After the update, verify your GRUB config has been regenerated:

sudo grub-mkconfig -o /boot/grub/grub.cfg
sudo reboot

Should You Upgrade Right Away?

For production servers, it is generally wise to wait for distribution-packaged kernels rather than running mainline kernels directly. Distribution kernels carry backported security fixes, stable driver patches, and have been tested against the distribution’s userspace stack.

For desktops, enthusiasts, and developers who want to take advantage of new hardware support or try the latest features, the mainline kernel is safe to run in parallel with your existing kernel using GRUB’s multi-boot capability. If anything breaks, simply select the previous kernel at boot.

Linux kernel 6.14 represents a solid step forward in hardware coverage, security, and performance. With AMD RDNA 4 support, improved Intel platform management, continued Rust adoption, and meaningful networking enhancements, it is a release that will benefit a wide range of users as it filters down through distributions over the coming months.

Security Hardening in Linux 6.14

Linux 6.14 brings meaningful security improvements that benefit both desktop and server deployments. The kernel continues to expand its use of Rust for memory-safe subsystems, reducing the attack surface for entire classes of vulnerabilities. Key security improvements include:

• Extended BPF LSM hooks: The BPF Linux Security Module framework gains new attachment points in 6.14, allowing security tools like Falco and Tetragon to observe and intercept more kernel events without loadable kernel modules.
• Improved KPTI coverage: Kernel Page Table Isolation improvements reduce speculative execution attack exposure on affected Intel processors while minimizing the performance overhead through better context switch optimization.
• Landlock LSM improvements: The Landlock sandboxing framework, which allows unprivileged processes to restrict their own filesystem and network access, gains new restriction types including TCP bind and connect restrictions in 6.14.
• Integrity policy enforcement updates: The IPE (Integrity Policy Enforcement) security module that enforces trust policies for executed code receives stability improvements and better documentation for enterprise deployment.

# Check if Landlock is supported
grep LANDLOCK /boot/config-$(uname -r)
# CONFIG_SECURITY_LANDLOCK=y

# Verify BPF LSM is active
cat /sys/kernel/security/lsm
# lockdown,capability,landlock,yama,apparmor,bpf

Power Management and Energy Efficiency

Power management is increasingly important as both laptop users want longer battery life and data centers want lower energy costs. Linux 6.14 includes several improvements:

• Energy-Aware Scheduling (EAS) improvements: The EAS system that directs workloads toward energy-efficient CPU cores on ARM big.LITTLE and Intel hybrid (P-core/E-core) designs receives calibration improvements that better match actual hardware energy characteristics.
• AMD P-state driver updates: The AMD P-state EPP (Energy Performance Preference) driver for Zen 4 and newer AMD processors receives fine-grained frequency scaling improvements that reduce idle power consumption by 5-15% on affected systems.
• Improved CPU idle: The cpuidle subsystem gains better C-state selection heuristics that reduce wakeup latency while still achieving deep sleep states when appropriate.

# Check current CPU frequency scaling driver
cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver
# amd-pstate-epp  (on Zen 4+ AMD systems)
# intel_pstate    (on Intel systems)

# View available EPP modes (AMD)
cat /sys/devices/system/cpu/cpu0/cpufreq/energy_performance_available_preferences
# default performance balance_performance balance_power power

# Set to power saving mode
echo power | sudo tee /sys/devices/system/cpu/cpu*/cpufreq/energy_performance_preference

For laptop users, these improvements translate directly to extended battery life on Zen 4/5 AMD laptops and Intel Meteor Lake/Lunar Lake systems. Benchmark results from the Phoronix Test Suite show 8-12% improvement in battery life on affected hardware running Linux 6.14 versus 6.13 under typical mixed workloads.