Chimera Linux enables proprietary NVIDIA drivers on musl with GLVND switch

Chimera Linux, a musl-based distribution, has developed a working approach to run proprietary NVIDIA graphics drivers despite musl's historical incompatibility with glibc-linked binaries. The project opened pull requests for both the legacy 580 and current 620 proprietary driver branches, targeting x86_64 and aarch64 architectures. The solution required three major technical changes: switching from Mesa's builtin dispatch to the GLVND vendor-neutral OpenGL dispatch library, leveraging the gcompat compatibility shim at its latest Git revision to resolve missing symbols, and using patchelf to modify the proprietary libraries so they can load under musl's dynamic linker.

The effort addresses a significant gap for musl distributions, where NVIDIA hardware owners previously faced a choice between poorly performing Nouveau drivers or no proprietary driver support at all. This particularly affects Maxwell and Pascal generation cards (GTX 700 through GTX 1000 series), which cannot reclock from boot frequencies without proprietary firmware and therefore perform poorly on open-source drivers. Newer Turing and later cards with GSP firmware work adequately with Nouveau/NVK but still lose 20-60 percent of proprietary driver performance.

How Chimera bridges musl and NVIDIA's glibc-linked drivers

• GLVND adoption: Chimera switched to GLVND on November 5 2026, replacing Mesa's builtin OpenGL dispatch; this is mandatory for proprietary NVIDIA drivers and also benefits hybrid graphics laptops where Mesa handles the integrated GPU while NVIDIA handles discrete graphics.
• gcompat shim at latest Git: The stock gcompat release lacks symbols required by NVIDIA's proprietary libraries, so Chimera bumped to the latest Git revision which provides mallinfo, __strftime_l, dladdr1, gnu_get_libc_version, and other missing glibc symbols via libgcompat.so.0 loaded through LD_PRELOAD.
• Patchelf for interpreter-less libraries: Most NVIDIA driver libraries lack a PT_INTERpreter header and cannot directly use gcompat's loader shim; patchelf modifies these binaries so musl's dynamic linker can process them, resolving the ld-linux-x86-64.so.2 dependency failure.
• Dual branch support: Pull requests cover both the legacy 580 branch (for Maxwell/Pascal and newer) and current 620 branch (for Turing and newer), with NVIDIA supplying binaries for both x86_64 and aarch64 architectures.
• Symbol resolution chain: The core OpenGL library libnvidia-glcore.so and compute library libnvidia-gpucomp.so both require glibc-specific symbols that gcompat intercepts, while libnvidia-tls and other helper libraries load normally once the interpreter path is corrected.

Why it matters

This development removes one of the most significant practical barriers to musl-based desktop Linux adoption, as NVIDIA remains the dominant discrete GPU vendor and previous musl distributions forced users toward either severely degraded performance or unsupported hardware configurations. For system administrators and users of Void Linux musl, Alpine Linux, and other musl distributions, Chimera's approach provides a reference implementation that could be ported or adapted, particularly the GLVND packaging pattern and gcompat revision requirements.