next up previous
Next: The Cross-Development Environment Up: Bootstrapping the Project Previous: Bootstrapping the Project

Picking a Starting Point


It is not practical to write all of the machine-dependent code from scratch. You should use as a base the NetBSD port for the machine most similar to the target machine. Ideally, there should already be a port to the CPU of the target machine; this will save you the trouble of rewriting all of the machine-language portion of the machine-dependent code. The easiest CPU to port to is the Motorola 68000-series, starting with the 68020. There are several ports already for this chip ( e.g., Macintosh, Sun 3, Amiga, and HP 300), and a good bit of code is already shared among them. Other supported CPUs are the Intel i386, the MIPS R3000 (whose port name is ``PMAX''), and the Sun Sparc chip. Send e-mail to core@NetBSD.ORG to get an up-to-date list of working and in-progress ports.

Choose a name of your port that describes as specifically as possible the machine that the port will run on. For example, our Macintosh port was originally called ``Mac'', but was later renamed to ``Mac68k'' because the PowerPC chip was not supported. Make a copy of the /usr/src/sys/arch/ name sub-tree (where name is the name of the base port you choose), rename the subdirectories and files appropriately, and you are ready to start the port.

Lawrence Kesteloot
Fri Jan 20 16:29:52 EST 1995