Linus Torvalds announced The recent release of the new version of «Linux Kernel 6.19″, a release that stands out for its massiveness (more than 15,000 corrections) and that serves as an aesthetic and technical prelude to what will come later.
Torvalds He has already announced that the next version will jump to version 7.0. not because of a radical break in the code, but because of the accumulation of versions in the current series and, more symbolically, because support for Rust is finally leaving its experimental phase to become integrated as a core feature.
In this new release, more than 2.200 developers have collaborated to introduce 15.657 corrections, resulting in a 52 MB patch that alters more than 13.000 files.
Main new features of Linux 6.19
Among the innovations presented in this new version of the Kernel, the crown jewel is the Live Update Orchestrator (LUO). This subsystem It represents a historic advance for cloud computing and mission-critical data centers. Basado en the Kexec HandOver mechanism (WAREHOUSE), the LUO allows the administrator to restart the kernel and apply updates without interruption. the functioning of the processes nor lose the state of the system.
A Difference of the solutions of livepatching previous ones, which were limited to minor security patches, LUO addresses total continuityIt preserves device state, interrupt handling, and, most importantly, keeps direct memory access (DMA) operations running. This means a server can migrate to a completely new kernel while user applications continue to operate as if nothing has happened.
Another new feature in Linux 6.19 is in the file subsystem, which is especially beneficial for mass storage servers. Ext4, which now supports data blocks larger than the traditional 4 KB in x86 systems. This ability to handle large files improves buffer write performance by 50%, optimizing memory usage on modern high-capacity disks. Meanwhile, Btrfs improves its interaction with the hardware, This allows resource-intensive processes like disk scrubbing or device replacement to no longer block the system's sleep mode. Now, the kernel saves the state of these operations before entering sleep mode and resumes them transparently upon waking.
On the other hand, the protocol NFS receives implementation of revocable directory delegation. This function It allows the server to delegate full control of a directory to the client. who can monitor changes locally without overloading the network with constant requests. If another client attempts to modify the same directory, the server dynamically revokes the delegation, maintaining data integrity with minimal latency.
Regarding security, in Linux 6.19 receives Intel LASS integration improvements (Linear Address Space Separation), the kernel now uses the processor hardware to create a barrier between user space and system space. By dividing the address space using the highest-order bit, The processor can instantly verify if an access is legitimate even before speculative execution begins. This technique is a definitive defense against side-channel vulnerabilities such as Spectre and Meltdown.
Complementing this shielding, the kernel introduces support for PCIe Link Encryption and Device AuthenticationIn protected virtualization environments such as Intel TDX or AMD SEV-SNP, DMA traffic traveling between physical devices (such as network cards or AI accelerators) and virtual machines can now be encrypted and authenticated.
The network subsystem It also received improvements for the next generation of data infrastructures, adding support for links up to 1600 Gbps (1.6T). To achieve these speeds, a thorough cleaning of internal bottlenecks has been carried out: the elimination of spinlocks in the transmission functions and the use of lock-free structures have quadrupled performance under extreme loads. Furthermore, the system of Asynchronous I/O io_uring debuts zero copy receive mechanism (zcrx), which allows receiving network data directly into the application's memory without going through intermediate copies in the kernel, drastically reducing latency and CPU load.
The industrial and automotive sectors also stand out arrival of the CAN XL protocol, since by increasing the data field to 2048 bytes, It allows Ethernet frame tunneling and native integration with TCP/IP networks within the vehicleThis, combined with the ability to reach speeds of 20 Mbps, prepares Linux to be the heart of next-generation autonomous driving and telemetry systems.
Finally, The graphics section has received an update. which will delight many, for the DRM subsystem (Direct Rendering Manager) It now has an API that allows for direct HDR color transformations in the display hardware, before and after layer merging. This eliminates the need to process color using GPU shaders or the CPU, which not only guarantees absolute color accuracy in professional editors, but also drastically reduces power consumption in laptops.
Last but not least, it is also worth highlighting that the controller AMDGPU has finally absorbed support for graphics cards from the Radeon HD 7000 and 8000 families (GCN 1.0 and 1.1 architectures). This migration is not just a name change; it allows these cards, which are over a decade old, to increase their performance by an average of 24% and gain compatibility with the modern Vulkan 1.3 API, demonstrating that in the Linux world, well-designed hardware never dies, but evolves.
Finally, if you are interested in knowing more about it, you can consult the details in the following link