Intel Linux engineers have been working on enhancing the NVMe storage performance with today's high core count processors. Due to situations where multiple CPUs could end up sharing the same NVMe IRQ(s), performance penalties can arise if the IRQ affinity and the CPU's cluster do not align. There is a pending patch to address this situation. A 15% performance improvement was reported with the pending patch.

It's typically rare these days for the ATA subsystem updates in the Linux kernel to contain anything really noteworthy. But today some important fixes were merged for the ATA code to deal with a reported power management regression affecting the past number of Linux kernel releases over the last year. ATAPI devices with dummy ports weren't hitting their low-power state and in turn preventing the CPU from reaching low-power C-states but thankfully that is now resolved with this code.

Following the December launch of Pop!_OS 24.04 LTS and the first major COSMIC desktop release, System76 software engineers have continued making improvements to their Rust-based desktop environment.

Deploying vLLM for LLM inference and serving on NVIDIA hardware can be as easy as pip3 install vllm. Beautifully simple just as many of the AI/LLM Python libraries can deploy straight-away and typically "just work" on NVIDIA. Running vLLM atop AMD Radeon/Instinct hardware though has traditionally meant either compiling vLLM from source yourself or AMD's recommended approach of using Docker containers that contain pre-built versions of vLLM. Finally there is now a blessed Python wheel for making it easier to install vLLM without Docker and leveraging ROCm.

Following recent discussions over AI contributions to the LLVM open-source compiler project, they have come to an agreement on allowing AI/tool-assisted contributions but that there must be a human involved that is first looking over the code before opening any pull request and similar. Strictly AI-driven contributions without any human vetting will not be permitted.

A lot of HID subsystem updates have been queuing up ahead of the Linux 6.20~7.0 merge window in February. There is a lot of new hardware support on the way along with quirks for some existing hardware support ranging from laptop keyboard issues to enabling support for more PS4/PS5 guitars under Linux.

Patches are now positioned to go into the upcoming Linux 6.20~7.0 kernel cycle for supporting Intel discrete GPU firmware updating on non-x86 systems.

A change proposal has been cleared by the Fedora Engineering and Steering Committee "FESCo" for providing a nice out-of-the-box experience for Windows on ARM laptops namely the recent Snapdragon X1 laptops and will also be important for the upcoming Snapdragon X2 laptops too.

Merged in time for the upcoming Mesa 26.0 release is the merging of Vulkan driver support for the Qualcomm Adreno Gen 8 graphics support that is notably used by the new Snapdragon X2 laptop SoCs as well as the Snapdragon 8 Elite Gen 5.

DragonFlyBSD's AMDGPU kernel graphics driver continues to be a port of the AMDGPU Linux kernel driver. Their latest porting effort for AMD graphics on DragonFlyBSD is now enabling optional support for the GCN 1.1 "Sea Islands (CIK) graphics processors on this modern alternative to the prior Radeon kernel driver.

A proposal has been laid out for a new X.Org Server "main" Git branch to house their development going forward and cleaning up the development lapses over the past few years. Ultimately the hope is for having a new cleaned-up X.Org Server and XWayland Git branch for shipping new releases in 2026.

Last year Valve contractor Timur Kristóf managed to improve the AMDGPU driver enough for old GCN 1.0 Southern Islands and GCN 1.1 Sea Islands GPUs that with Linux 6.19 AMDGPU is now the default for those GPUs with better performance, RADV Vulkan out-of-the-box, and other benefits. He isn't done though improving the old GCN 1.0/1.1 era GPU support on this modern AMDGPU kernel driver - a new patch series posted today brings some power management fixes.

A new feature expected to be merged for the upcoming Linux 7.0 kernel cycle is adding an OPEN_TREE_NAMESPACE flag for the open_tree() system call. This OPEN_TREE_NAMESPACE option can provide a nice performance win with added security benefits if you are dealing a lot with containerized workloads on Linux.

In late 2023 Mozilla began providing Debian packages of Firefox Nightly builds complete with an APT repository. Those on Debian/Ubuntu distributions have a much easier path for enjoying Firefox Nightly since then and now Mozilla engineers are providing similar RPM builds of Firefox nightly too.

Queued into the Linux networking subsystem's "net-next" branch ahead of the Linux 6.20~7.0 merge window next month is cake_mq as a multi-queue aware variant of the sch_cake network scheduler. The intent with cake_mq is to better scale the network traffic rate shaper across multiple CPU cores.

Earlier this month we looked at the Dell Pro Max GB10 performance up against AMD's Ryzen AI Max+ "Strix Halo" with the superior performance for the green team for performance and power efficiency. For those wondering how the Dell Pro Max GB10 performance comes up for the much talked about NVIDIA GH200, here are some comparison benchmarks.

A new feature that appears ready for introduction in the upcoming Linux 6.20~7.0 kernel cycle is revocable resource management.

Support for newer HDMI features in the open-source AMD Linux graphics driver have been limited due to being blocked by the HDMI Forum. There are though some new HDMI gaming features being enabled via new AMDGPU kernel driver patches that are coming outside of AMD and based on public knowledge and/or "trying things out until they work/break" for functionality like HDMI Variable Refresh Rate (VRR) and Auto Low Latency Mode.

There have been a number of nice RADV driver Vulkan ray-tracing performance optimizations for Mesa in recent times... Here is yet another merge request now merged for Mesa 26.0 and helping deliver some nice performance uplift for ray-traced games on Linux. And, yes, this is yet another Valve contribution to this open-source AMD Radeon Linux graphics driver.

Following last week's updated Intel LLM-Scaler-vLLM release for helping advance vLLM usage on Intel Arc Graphics, LLM Scaler Omni is out with a new release today for that LLM-Scaler environment focused on image / voice / video generation using Omni Studio and Omni Serving modes.