Native and Hosted kernel design
Posted: Fri Dec 14, 2012 7:53 pm
I wrote a small X86 monolithic kernel with a few command line commands about 3 years ago.
It got deleted as well as 1 TB of data. no I have decided to return and try again. This time with a new design.
I am trying to do a micro kernel this time (written in C and Assembly(gas syntax)) with two different modes of running it.
one as a native micro kernel.
Two hosted on top of a host kernel like Linux, *BSDs, Mac OS X, etc.
I would like to know what would be required to run on top of these other kernels and be able to both use the host's drivers and allow the host to use my drivers. While at the same time keeping the hosted programs that are running on my kernel from talking to the host system or using the host APIs.
I has thinking virtual drivers to access the host hardware and using PCI pass through to allow my hosted kernel to access the hardware for which the Host OS my not have a driver for.
I don't have any Idea how I could allow the host to use my drivers though
It got deleted as well as 1 TB of data. no I have decided to return and try again. This time with a new design.
I am trying to do a micro kernel this time (written in C and Assembly(gas syntax)) with two different modes of running it.
one as a native micro kernel.
Two hosted on top of a host kernel like Linux, *BSDs, Mac OS X, etc.
I would like to know what would be required to run on top of these other kernels and be able to both use the host's drivers and allow the host to use my drivers. While at the same time keeping the hosted programs that are running on my kernel from talking to the host system or using the host APIs.
I has thinking virtual drivers to access the host hardware and using PCI pass through to allow my hosted kernel to access the hardware for which the Host OS my not have a driver for.
I don't have any Idea how I could allow the host to use my drivers though