Linus Torvalds Lashes Out At RISC-V Big Endian Plans - Phoronix

Abstract

Linus Torvalds has vehemently criticized recent proposals to formally integrate Big Endian support within the RISC-V architecture specifications and the Linux kernel. Torvalds argues that adopting Big Endian introduces needless complexity, technical debt, and fragmentation to an architecture that should prioritize Little Endian consistency. His strong opposition signals a critical architectural preference debate concerning the future standardization and maintenance burden of RISC-V within the mainstream software ecosystem.

Report

Key Highlights

• Linus Torvalds, the creator of the Linux kernel, publicly expressed severe disapproval of adding formal Big Endian (BE) support to the RISC-V architecture.
• Torvalds views the incorporation of BE as unnecessary technical debt and a significant long-term maintenance burden for kernel developers.
• The controversy centers on whether RISC-V should maintain strict Little Endian (LE) consistency (like x86) or accommodate specialized users who prefer or require Big Endian byte ordering.
• Torvalds asserts that if vendors want BE support, they should handle the complexity outside of the main kernel tree, emphasizing that RISC-V is inherently an LE architecture.

Technical Details

• Endianness: The core technical issue is byte ordering, determining how multi-byte data (like integers) are stored in memory (Most Significant Byte first for Big Endian, Least Significant Byte first for Little Endian).
• RISC-V Default: RISC-V's primary Application Binary Interface (ABI) defaults to Little Endian, aligning it with high-volume architectures like x86 and modern ARM.
• The Proposal: The debated proposal involves adding specific features or standardized methods within the RISC-V specification to handle optional Big Endian modes, often desired for networking systems or compatibility with specific legacy embedded hardware.
• Kernel Complexity: Supporting both Big and Little Endian modes simultaneously necessitates complex and often bug-prone code paths within the operating system kernel for memory handling and device interactions.

Implications

• Standardization Impact: Torvalds' powerful rejection significantly increases the political and technical difficulty of pushing formal Big Endian standards through the Linux kernel development process.
• Ecosystem Consistency: If the BE proposals are blocked, it reinforces the dominance of Little Endian within the RISC-V ecosystem, ensuring greater architectural consistency and reducing fragmentation for software developers and toolchain providers.
• Development Philosophy: This incident underscores a philosophical divide: optimizing for simplicity and modern adoption (Torvalds' view) versus accommodating niche or historical requirements (the BE proponent's view).
• Long-Term Maintenance: By blocking BE support, the Linux kernel aims to avoid years of potential maintenance headaches associated with debugging dual-endian systems, saving developer time and resources.