DRM colorops, Emacs compatibility, and Rust GPU work

DRM developers advanced color pipeline and Rust driver work while Emacs maintainers debated hash table equality and long-term binary compatibility. Blender sought Android tablet GPU data and an AMD engineer floated a dedicated ROCm path in QEMU.

YUV colorop lands for amdgpu and VKMS

A sixth revision of a ten-patch series adds a fixed-matrix YUV-to-RGB colorop to the DRM plane color pipeline. Implementations for amdgpu and VKMS have been reviewed and tested, giving drivers a shared path for YUV conversion inside the atomic color pipeline rather than ad-hoc plane logic.

Emacs hash tables gain structural equal

A patch proposes implementing structural equal for hash tables in Emacs Lisp. Discussion among roughly ten participants weighs the utility of deep comparison against backward-compatibility risks for existing code that treats hash tables by identity.

RMS thread on upgrades breaking old binaries

Richard Stallman opened a lengthy emacs-devel thread on why GNU/Linux upgrades break older programs, citing concrete GCC and Glibc changes that break historical Emacs builds. Participants examine possible adjustments in distro packaging and Emacs practices to improve long-term compatibility.

Blender seeks Android tablet Vulkan reports

Blender developers asked for Vulkan capability reports from Android tablets to judge support for mid-range and older devices. The thread collected dozens of replies aimed at mapping which GPUs can realistically run Blender workloads.

Rust DRM panel bindings hit soundness issues

A five-patch series introduces Rust abstractions for DRM panels. Automated review flagged multiple critical reference-counting and soundness problems that must be resolved before the bindings can land.

Nova-core PRAMIN window support

Version two of a twelve-patch series adds PRAMIN window handling and related Rust abstractions to the nova-core NVIDIA driver. Maintainer review is under way on the new memory-window interface.

Dedicated ROCm compute backend proposed for QEMU

An AMD engineer suggested a ROCm compute backend separate from the virglrenderer display path. The goal is reduced queue contention and proper support for asynchronous events in GPU compute virtualization.

Rust ID range reservation for nova-core

The fifth revision of a patch set adds Rust bitmap bindings for reserving contiguous ID ranges, intended for nova-core. Review discussion focused on NonZero ergonomics and safe allocation of ID pools.