On Fri, 14 Jan 2000, Philip Blundell wrote:
> > I'm starting to look into it now, I'm not sure gcc is the right
> >tool for what I'm doing but I will have a look at it and help out if its
> >suitable or if I can.
>
> What is it that you're trying to do?
Probably not something you guys would be too intersted in. I'm
trying to address the problem of effectively implimenting high-performance
microkernels. A lot of the kernel is implimented in assembler and the
register allocations make the C calling convention a source or real pain
and pure overhead. So where I would like to use C for the less critical
and more structured code (also for code common to different platforms)
there is a lot of hassel on the assembly to C and C to assembly borders.
We also have some problems with the syscall stubs for similar reasons.
The way to do it seems to be to design a tool to embbed C in
assembly. What I want to do it develop a compiling tool that adds some
syntacs for specifing the a different calling convention for each group of
functions on a assember/C,C/assember bondary. Sort of a C with additional
contraint information for the back-end code generators. I'm having a look
at IDL's (Interface Descriptor Languages) to see if thats what I want.
Otherwise a redone back-end to gcc or similar compiler might do the trick.
I'm writting up a proposal for such a tool, if you have time I might send
it your way once its a little more complete for comments if thast ok?
> > >Who are the several people?
>
> Among others, Richard Earnshaw, Nick Clifton, Scott and myself. GCC's ARM
> backend is quite actively maintained.
>
Great, I'm starting to have a look at it now. I'll let you know
once I get to a helpful stage as I'm interested in a StrongARM optimised
gcc :)
Cheers Adam
unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
++ Please use [EMAIL PROTECTED] for ++
++ kernel-related discussions. ++
