A narrative account of OpenBSD’s ARM port origins in 2003-2004, from vestigial NetBSD/arm32 code to CATS board bringup as a foundation for the Zaurus SL port.
Key Takeaways
The Sharp Zaurus SL series (2002) was the target: real keyboard, CF/SD slots, Linux-based, ARM processor, and a debug serial port that proved critical.
Before tackling Zaurus, Dale Rahn chose the Chalice Technologies CATS board (233MHz SA-110 StrongArm, ATX form factor, ISA/PCI) as a stable dev foundation, since no OpenBSD ARM port existed.
The ABLE firmware from Simtec replaced the original Cyclone firmware but introduced a cascade of bugs: broken ELF loading, broken filesystem support, and untested version 184 that broke BSD booting entirely.
Theo de Raadt negotiated CATS boards at 125 GBP instead of 199 GBP and distributed six units to core developers to parallelize the port work.
OpenBSD’s ARM32 history traces to NetBSD/arm32 (formerly RiscBSD, led by Mark Brinicombe), synced into OpenBSD then removed in early 2001 for lack of interested developers.
