[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [f-cpu] F-CPU gcc CVS



> > Eh? :
>
> My fault; I got it now.  But it is "slightly broken":  When I turn off
> optimizations, gcc uses loadaddrid for all labels (if fcpu_reorg isn't
> run, the `call' member is never set, and print_operand assumes a data
> operand).

yes I know - but it should still work even with penalty on each
address use.
IMHO now comes the emulator - you could handle syscalls by means
of mapping them directly or thru some xlat table to linux syscalls.
Yes I know it is linux dependent (so should be option) but it is
the fastest way.
Then we can take uClibc and write syscall fn in assembler. Then we
should be able to compile any application (bash f.e.) and RUN it !

It allows us to test compiled output and measure some things before
digging blindly into compiler ...

devik

*************************************************************
To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe f-cpu       in the body. http://f-cpu.seul.org/