Re: CVS commit: src/sys/arch

[yamt%mwd.biglobe.ne.jp@localhost (YAMAMOTO Takashi)]

> > > Any idea why i386 has the different code fragment
> > > 'cii = 0, ci = cpu_info_list; ci != NULL; ci = ci->ci_next'
> > > for its own copy of CPU_INFO_FOREACH?
> > 
> > the sys/cpu.h version is for UP-only ports.  i386 has SMP.
> 
> Two additional questions:
> 
> acorn has CPU_INFO_ITERATOR being cpu_t; is there any reason to not
> make this the default everywhere? (I don't see much value in it
> being per-arch, unless there's something I'm missing.)

i don't know.

i think it makes sense to maintain an MI list in kern_cpu.c.

> There also seems to be various things per-arch that are conditionally
> defined on whether MULTIPROCESSOR is set; would it be reasonable
> to make the MP/UP code global, the version selected by MULTIPROCESSOR?

making it depend on MULTIPROCESSOR is not desirable.
eg. consider kernel modules.

YAMAMOTO Takashi