EKP++ - Experimental Kernel Project++ / News: Recent posts

The latest-and-greatest code has been imported under the ekp-05 tag. This is the first update in the CVS since two years ago. That's pretty impressive for a dead project :-).

So what are going to be the changes?
1) Mulitple architecture support. IA-32 and PPC.
2) OOP design from the ground up... instead of using C++ as a 'neater' C.

EKP is alive and well, although the CVS certainly does not reflect that. A large overhaul has been going on, and it will be just easier to have SF staff replace the current CVS repository with my new tarball (after I finish the large structural changes).

EKP++ is not dead. There has been recent work - cleaning up interrupts.cpp and vsprintf.cpp, re-organizing the source a bit, and an implementation of a page allocator. After much delibiration, and disagreement with some of the ideas behind MIT's exokernel, I have decided EKP++ to continue forth as a simple message-passing microkernel.

Yes - the source in the CVS is ancient, mostly due to the flakiness of SF's CVS (can't commit???) and due to CVS's backwardness (must locally delete things.... argh). When I have the time, I will upload the latest CVS, but I am mostly thinking about releasing tarballs of source under the "Files" section.

No, this project is not dead. I plan to work on it some during and after Thanksgiving, as I barely have any time to work on it now.

Yesterday, September 21 ~6:00 PM, the EKP++ codebase was finally imported into the CVS. As of 10:00 AM September 22, the Backup CVS (read: WebCVS) has not been synced yet.

The EKP++ project is finally re-locating itself to SourceForge.net. I say re-locating, as opposite to re-located, since IMSA's policy of blocking all ports but 80 prevents me from uploading the source tree into the CVS. Expect that to be done by the end of Sunday. (or earlier)

The current state of affairs is quite simple - we can (and have) set up a basic GDT, we can (and have) set up a simple IDT (filled to the brim with unsophisticated CPU Exception handlers and such). We can also print debug information to the Bochs (and x86 emulator used in the development) console (as long as Bochs was compiled with the port 0xe9 hack). Since EKP++ is an exokernel - there is really no justification for having a video driver being part of the kernel. (Note: the video driver sourcecode has been extracted from the kernel sources and will be used to make a "servce" that will run on top of the kernel - it will be in the CVS, probably in the services directory - its just not part of the kernel)... read more