On Sun, 2006-Aug-06 20:26:45 -0300, Duane Whitty wrote: >I'm somewhat interested in this as well. Towards that end would it be >correct for me to believe the process is easier if the architecture >already has a UNIX(tm) style operating system installed and operating and >if there are GNU tools available for this architecture already (assembler, >compiler, linker)?
The Un*x-like OS is not especially critical but a working toolchain is. >What I am thinking about is HP's HP 9000 PA-II RISC architecture. Seems to >me that if the above mentioned tools are available that about 85% of the >work is already done, but then I only have limited experience in this area. Maybe 85% of the total effort (starting from nothing). The other critical item is access to the system programming information for the system - and I'm not sure how readily available this is for the HP-PA. Currently FreeBSD has support for 6 platforms. There is a total of 10MB or 304KLOC in the MD-trees - this is about 50KLOC per CPU. Whilst you may be able to leverage off existing MD code (especially the NetBSD HP-PA port), a port to a new architecture is a non-trivial undertaking. >I really like the HP 9000 platform and I would love to see FreeBSD on it. The HP 9000 is really more NetBSD territory than FreeBSD territory. Unless you can get a critical mass of developers who are interested, the port is a non-starter. The Alpha port died because there wasn't sufficient interest to keep it going - and much of the loss of interest was a result of Compaq killing the Alpha. I suspect that a HP 9000 port would be starting from a much smaller base. -- Peter Jeremy
pgpIIZBCpQODq.pgp
Description: PGP signature
