Messages in this thread | | | Subject | Re: Porting Linux to a new architecture | Date | Sun, 24 Mar 1996 21:31:06 +0200 (MET DST) | From | (Michael Hohmuth) |
| |
First of all, thanks to all who replied, either privately or in public!
Linus Torvalds wrote:
> I don't think any documentation exists, but the best porting model to use > is probably to read the header files in include/asm-xxx/, because those > cover the differences reasonably well. Also, just look at one of the > other architecture arch/xxx/kernel and arch/xxx/mm subdirectories.
I already went through documenting most the include/asm-* interfaces the arch-independant parts of the kernel expect; that was the easiest part, I think.
As far as I can tell from looking at the source, the architecture support is required to implement certain include/linux interfaces, too. Also, Linux makes several assumptions about kernel routines being called by the arch support at certain events, e.g. the page fault handlers. This information seems to be much more difficult to extract; how did other porters set about this?
Christopher J. Shaulis wrote:
> Well, my curiosity is perked. What new architechure are ya porting > too?? If its not a secret. =)
I'd hoped you wouldn't ask that. ;)
We consider porting the Linux kernel to the L4 microkernel, and perhaps to the L3 microkernel, too. The rationale is that we need a decent development environment running on these microkernels, and we found the successful Linux port to OSF Mach encouraging.
In fact, it should be a lot easier to port Linux to L4 than to Mach, because L4 resembles a ``plain'' i486 much closer than Mach. We even plan to directly use all the Linux device drivers for i386 (L4 doesn't have microkernel-internal device drivers) and the Linux paging strategy.
(Anyone interested in more information about L3/L4 should start browsing at <URL:http://www.inf.tu-dresden.de/~mh1/l3/>.)
Thanks again, Michael -- Email: hohmuth@inf.tu-dresden.de WWW: http://www.inf.tu-dresden.de/~mh1/
| |