The arrival of Linux 7.0 marks a new milestone in the evolution of the kernelBut not so much because of the round number itself, but because of the sum of changes it incorporates. Linus Torvalds has confirmed its availability This stable version comes after an intense development cycle, with many small fixes, massive testing, and a clear focus on stability and next-generation hardware.
Although Torvalds insists that the Numbering jumps do not correspond to a specific major "mega-change"Linux 7.0 has become the de facto foundation upon which key distributions like Ubuntu 26.04 LTS and many popular rolling releases are built. Among its major strengths are a smarter task scheduler, significant improvements to memory and swap, the definitive integration of Rust into the kernel, and enhanced support for CPUs, GPUs, and NPUs that haven't even reached the market yet.
Why is it now called Linux 7.0 and not 6.20?
The decision to jump to Branch 7.x has more to do with internal organization than with marketingTorvalds continues his practice of resetting the counter when a series reaches version x.19, to avoid long and confusing numbering. In this case, after Linux 6.19, the next natural step was 7.0.
During the preceding weeks, the release candidates They showed unusually high activity. This volume of commits didn't signify a flood of last-minute updates, but rather that the community was polishing a large number of minor bugs. There were moments of unease, especially during RC2 and RC3, which Torvalds described as some of the biggest in a long time, but ultimately development stayed on schedule.
In the final week before launch, the pattern remained the same: "many small fixes" that seemed benignTorvalds also pointed out an interesting change in context: the use of artificial intelligence tools to find extreme cases and subtle errors is becoming commonplace in the development cycle, to the point that it could become the new normal.
Linux 7.0 release schedule and arrival in distributions
The Linux 7.0 development cycle has followed the usual pattern of a few ten weeks between the first release candidate (7.0-rc1) and the final versionMeanwhile, estimates pointed to a release date around April 12th, with some leeway if an extra research and confirmation period was needed. Ultimately, the stable release occurred within those projections, without significant delays despite a somewhat hectic research and confirmation period.
For those who use rolling release distributions Like Arch Linux or similar systems, the new kernel will quickly reach the official repositories after the stable label is released. At the other extreme, in more conservative environments like Debian stable or derivatives, the update to 7.0 may take considerably longer or not arrive at all, depending on the policies of each project.
Ubuntu 26.04 LTS will be released directly with Linux 7.0 As a base, while Ubuntu 24.04 LTS will receive this kernel via backport in an update planned for July, likely the last major kernel version Canonical will offer for that edition. However, users of other intermediate versions like 25.10 will not see 7.0 by default and will have to resort, if they wish, to packages from the mainline PPA, external DEBs, or manual compilation, with the support implications that entails.
Most distributions used in administrations, educational centers and companies They tend to prioritize LTS versions and kernels with extended support. Linux 7.0 is not a long-life edition, so in critical servers and production systems of public organizations, it will be common to remain stuck on 6.x branches supported until 2028, while 7.0 will gradually gain traction, especially in workstations, labs, test deployments, and environments where early support for new hardware is needed.
A more refined task scheduler: goodbye to some of the micro-stuttering
One of the changes that users can most notice on a daily basis is the kernel task scheduler revisionFor years, certain scenarios have shown small stutters (micro-stutter) when a critical task loses control of the processor at a delicate moment, for example, when compiling, playing games, or running workloads with intense peaks.
Linux 7.0 introduced the so-called Time Slice Extension (TSE)This mechanism allows tasks deemed important to have a little more CPU time before being interrupted. This extra granting of milliseconds reduces untimely interruptions without compromising overall fairness between processes, which is especially important on desktops, laptops, and workstations where interactive applications are combined with background workloads.
The planner improvement doesn't come alone: ​​the Memory management has also been significantly improvedThe kernel allocates and reclaims memory more intelligently, and bottlenecks that affected performance under pressure have been eliminated. This is noticeable both in systems with a lot of RAM, where queues are managed better, and in more modest systems, where the use of swap and zram becomes especially important.
Memory, swap, and zRAM: more performance with a full house
Linux 7.0 continues the work started in versions 6.18 and 6.19 to increase the efficiency of the swap subsystemIn the first phase, performance under memory pressure was improved; now, data reads back from swap to RAM are optimized when it is saturated.
The load tests where multiple processes share the same exchanged pagesRedis configurations with persistence have shown performance improvements of up to 20%. In desktop environments, the gains are more modest, but results tend to be the same as or better than the previous baseline, with no apparent penalties.
A significant new feature for many laptops and mid-range devices is that the kernel You can write compressed data directly from zram to disk When the memory is full, there's no need to decompress them beforehand. This change reduces extra work and improves efficiency in systems that combine zram with disk swap, which is common in distributions used extensively on older or low-cost computers.
Rust is here to stay: security and new drivers in Linux 7.0
One of the technical headlines of this release is that the Rust language ceases to be an experiment and becomes a full-fledged citizen within the kernel. What began in 2022 as a limited test is now consolidated as a stable part of the code, with the approval of Linus Torvalds and the continued work of the Rust-for-Linux project, led by developers like Miguel Ojeda.
This does not mean that C will disappear from the coreC will remain the predominant language in the vast majority of subsystems, but starting with Linux 7.0, the door is opened for new drivers and components to be written directly in Rust. The goal is to reduce vulnerabilities related to memory management, which, according to internal estimates, account for around 70% of serious security flaws.
Rust provides structural safeguards against common errors such as out-of-range access, double releases, or the use of dangling pointers. For the industries that rely on Linux in sectors such as banking, telecommunications, administration, and healthcareThis move represents a reinforcement of basic security, something especially valuable now that EU regulations are becoming increasingly demanding in terms of cybersecurity.
File systems: Self-healing XFS, faster EXT4 and NTFS3
The area of ​​storage also receives significant attention. One of the most striking additions is the XFS file system's "self-healing" capabilityThrough a new daemon, xfs_healer, managed by systemd, the system monitors metadata errors and I/O failures in real time and can initiate automatic repairs without needing to unmount the volume.
This functionality relies on a new generic framework for file system error reportingThis standard unifies how the kernel communicates metadata corruption and I/O problems to user space using fsnotify. Until now, each filesystem had its own mechanisms, if any, which complicated centralized monitoring and automated response.
EXT4, the default file system in many distributions like Ubuntu, improves the concurrent writing with direct I/OThe changes delay the splitting of unwritten extents until the operation is complete and prevent unnecessary cache invalidations, which benefits scenarios where multiple processes write concurrently, such as backup tools, build systems, or download managers.
For those who live with Windows partitions or external drives, the NTFS3 driver receives a substantial updateDeferred allocation has been added to improve performance, along with iomap-based operations and more efficient readahead for large directory scans. Multi-cluster reads have been refined in exFAT, with performance improvements especially noticeable on media with small clusters, such as certain SD cards and low-capacity USB drives.
Overall performance: processes, files, and latency
Beyond visible changes, Linux 7.0 introduces internal improvements in the creation and destruction of processesas well as in file opening and closing operations. Specific benchmarks show that PID allocation is now between 10 and 16% faster, while open/close operations can be between 4 and 16% faster on multi-core machines.
In terms of security, the following is added: BPF filtering for io_uringThis allows for the sandboxing of operations that many administrators previously preferred to disable outright as a precaution. This maintains the performance gains of io_uring, but with the ability to fine-tune what can be done and how, a valuable feature in data centers and private clouds.
The kernel also takes advantage of this leap to Remove historical features that make little sense in the current park, such as laptop_mode, an energy-saving mechanism for mechanical hard drives that dated back to the 2.6 kernel era. With the dominance of SSDs in laptops and the complexity it added to memory and write code, developers have decided that it is no longer worth maintaining.
Support for current and future hardware: Intel Nova Lake, AMD Zen 6 and more
One of the focuses of Linux 7.0 is to prepare the ground for CPU and GPU architectures that will arrive on the market in the coming yearsOn Intel's side, the kernel incorporates base support for future Nova Lake CPUs, including desktop variants and configurations with different numbers of cores, as well as additional work on Crescent Island accelerators.
In modern Intel processors (10th generation and later), the kernel activates automatic mode by default for Intel TSX (Transactional Synchronization Extensions)This technology, which was once massively disabled due to vulnerabilities like TSX Asynchronous Abort, is now being reactivated on non-vulnerable chips and remains disabled on affected ones, thanks to a self-detection logic. The result is a potential performance increase in multithreaded workloads that can take advantage of TSX, without compromising security.
On the AMD side, Linux 7.0 includes support for Performance events and metrics for the next generation Zen 6This includes counters related to branch prediction, L1 and L2 cache activity, TLB, and uncore events such as memory controller activity. While end users won't see immediate changes, this data is valuable for developers and administrators preparing software and platforms for when new processors are released.
In virtualization, KVM adds support for AMD ERAPS (Enhanced Return Address Predictor Security)A Zen 5 security feature that extends the depth of the Return Stack Buffer in virtual machine environments. This allows VMs to benefit from the same protections and return prediction capabilities as the host system.
Graphics, NPU and video: Ready GPUs and greater AI efficiency thanks to Linux 7.0
In terms of graphics, Linux 7.0 continues to expand the reach of free drivers. The amdgpu driver continues to incorporate IP blocks for GPUs based on RDNA 3.5 and potential RDNA 4 successors.This lays the groundwork for future cards that haven't yet been officially announced. A deeper integration between GPUs and NPUs is also anticipated in future generations of Radeon hardware, although no public details have been released yet.
For users of Intel Arc GPUs and integrated Xe graphics, the new kernel It exposes much more thermal telemetry through HWMONNow you can see not only the overall GPU temperature, but also shutdown limits, critical and maximum values, as well as readings from the memory controller, the PCIe link, and even individual VRAM channels. This improves temperature monitoring and diagnostics, which is especially useful for high-end desktop and laptop computers that are starting to be sold with these GPUs.
In the NVIDIA world, the Open-source NVK driver for recent GPUs restores support for large pagesThis results in performance improvements for certain 3D and compute workloads that can take advantage of that page size.
Beyond pure GPU support, Linux 7.0 introduces a revamped computational acceleration subsystem to communicate directly with NPUsThis allows AI tasks to run directly on the NPU without additional intermediaries, offering significant benefits: battery consumption is reduced by up to 80% compared to running the same tasks on the CPU, and increased efficiency means more applications can perform inferences locally without relying so heavily on the cloud. For users and organizations concerned about data sovereignty, processing AI models on the device itself is a clear advantage.
Laptops, peripherals, and new keys for the AI ​​era
On laptops, many changes may go unnoticed but make a difference in daily use. ASUS WMI driver improves control of brightness, backlighting, and RGB effects In lines like ROG and TUF, it includes support for shortcuts such as the Fn + F5 fan control key on some models. The HP WMI driver adds manual fan control to HP Victus laptops and resolves minor issues such as the audio mute LED on the Victus 16, which wasn't updating correctly.
Lenovo laptops and portable consoles, such as the Legion family and Legion Go-type devices, They expose more hardware sensors to monitoring tools Thanks to improvements in the Lenovo WMI driver, it's now easier to monitor temperatures and fan speeds from Linux. For brands like TUXEDO, the kernel adds the ability to manage cTGP (Configurable Total Graphics Power) on some InfinityBook Gen7 models with NVIDIA 3000 series GPUs, although currently through sysfs attributes and not graphical interfaces.
Among the unusual peripherals gaining support are the following: Rock Band 4 Bluetooth controllers for PS4 and PS5These now run directly on Linux, and the Logitech K980 solar-powered wireless keyboard is fully supported via Bluetooth. New HID codes related to AI agent interaction keys have also been added, anticipating laptops with dedicated buttons for smart assistants and integrated AI features.
Architectures and platforms: ARM, RISC-V, Loongson and more
Linux 7.0 continues to expand the range of supported architectures. This version strengthens support for ARM, RISC-V and Loongson platformsas well as for veteran processors like SPARC or DEC Alpha, which continue to receive timely updates thanks to a very loyal community.
In the case of RISC-V, the kernel wins support for user-space control flow integrity mechanisms (CFI), an important component for strengthening software security in this emerging architecture. Progress is also being made in integrating specific SoCs like the Spacemit K3 RVA23 and in supporting new wireless connectivity specifications such as WiFi 8 (Ultra High Reliability), which is beginning to take shape in the network stack, although its widespread deployment is still years away.
In the ARM field, in addition to the aforementioned Rockchip SoCs, efforts continue to Improve the experience on devices with Qualcomm SnapdragonThis includes new laptop-oriented chips like the Snapdragon X Elite and X2 Elite. Version 7.0 integrates new PHY elements and other support blocks, but the ecosystem itself acknowledges that there's still a way to go to achieve a fully polished experience on ARM laptops running Linux.
Kernel security and post-quantum cryptography
Security remains a central focus. In addition to the indirect benefits of Rust, Linux 7.0 introduces changes in cryptographic infrastructure and in module signature managementOne of the most relevant decisions is the withdrawal of SHA-1 as a signature algorithm for kernel modules, replacing it with schemes based on ML-DSA, considered more robust against next-generation attacks and aligned with the transition to post-quantum cryptography.
During the final phase of development, the maintainers also They have resolved specific vulnerabilities that could have delayed the launchThese included spurious hardware errors detected in AMD Zen 3 CPUs and out-of-bounds access in the X.509 certificate code that could be triggered by unprivileged users and had been present in the main kernel for three years.
In parallel, the kernel security documentation, specifically the file security-bugs.rst has been updated to better guide AI tools This includes both automated reports and human users reporting bugs. The goal is to reduce the noise and focus on reports with truly useful information, given that the volume of notifications has skyrocketed as automated tools have improved.
Linux 7.0 in the cloud and data protection
In cloud environments, where Linux remains dominant, this version reinforces the isolation of virtual machines and protection of data in transit and at restA key focus is on encrypted memory enclaves and improved isolation techniques, designed to prevent even personnel with high privileges in the infrastructure from inspecting sensitive customer data.
Large international providers such as Meta or Amazon, which maintain a significant presence in data centers, demand mechanisms that allow data to be invisible even to administratorsLinux 7.0 moves in that direction through improved isolation and encryption tools, which, combined with XFS self-healing and standardized I/O error reporting, provides a more robust foundation for financial, healthcare, or government services deployed in public and private clouds.
Overall, Linux 7.0 presents itself as a version that, without being marketed as a revolution, It consolidates many lines of work initiated in the 6.xy series and prepares the ecosystem for the next decade of hardware and services.From desktops to the cloud, including laptops, servers, and embedded devices, this kernel enhances stability, optimizes memory and storage performance, brings AI to the device with lower power consumption, and strengthens security at the language, cryptography, and isolation levels. It's not an Extended Support Release, but it serves as a clear benchmark for measuring the direction of Linux kernel development.
