I had the Linux kernel and some simple user-space tools like busybox running on an embedded platform with 512 kB RAM and 2MB flash back in 1999! Those were fun times. To be honest 512 kb was possible but very on the limit, I think the product we launched with it had a few megs of RAM eventually. We had to invent a journalling flash filesystem as well in order to make it work in practice, something that didn't exist back then either. But Linux then was really a breakthrough compared to the horrible mess of embedded OSes that were needed otherwise to handle TCP/IP, filesystems and multitasking.
Yeah the 2/8 combo was probably what we went with in the product as well. The 512k was more like a shoehorned concept demo in an existing product.
The next thing we did was make a version of our CPU with an MMU, designed to work optimally with Linux (the first version was on the uClinux concept, with a kernel without MMU support and user-space programs that couldn't rely on fork() or mmap() fully). After a year or 2 with MMU-less Linux, it was like heaven to be able to run on an MMU :)
