On Thu, 30 May 2002, Preston L. Bannister wrote: > I'm curious - who is adapting Plan9 for new PC hardware?
A cast of literally 10 or 20 peoples. Thousands of millipeople! That said, a Plan 9 driver is MUCH easier to write than a Linux driver, for example, so new drivers are pretty quick. And the kernel interface doesn't change much if at all, so you don't need to keep tweaking them. > It seems like the downfall of LinuxBIOS is the limited space available > in the stock EPROMs, the growth in the Linux kernel, and the continual > appearance of new PC chipsets. Yes, this does cause trouble with the basic concept. However, now that various DOE labs are planning to buy thousands of linuxbios-capable nodes we are suddenly seeing more interest in putting big flash on some motherboards -- the motherboard vendors are contacting us. So the idea is not hopeless. Motherboard vendors in at least two cases are supporting ports as well. I don't actually see the new PC chipset issue as a problem. I'm a bystander on some of the new ports. Also at least one chip vendor is doing their own ports now -- you'll see more on this later. I estimated at least 3 years until we started to get critical mass with vendors. We started 2 1/2 years ago and we have gone from "you're an idiot" to "here's hardware, do a port" to "we've done the port for you" in that time. I've just had one of the Big Guys contact me to see about a port. This was always a risky venture, but it does seem to be succeeding, due to the efforts of a lot of good people all over the world. Week by week, it seems slow, but where we are now is pretty amazing compared to a year ago. I am even hearing from companies who are losing sales because they don't support linuxbios -- they come to us and see what they need to do. > So how is it that Plan9 is working with current hardware? because Bell Labs is still supporting Plan 9, to the tune of (I estimate) $10M/year or more (just look at how many guys are working on it and you get a rough idea). ron
