Re: [PATCH] Unicore architecture patch review, part 2

[Arnd Bergmann <arnd@arndb.de>]

On Wednesday 08 December 2010, Guan Xuetao wrote:

> > > @@ -0,0 +1,390 @@
> > > +/* 0 */		CALL(sys_restart_syscall)
> > > +		CALL(sys_exit)
> > > +		CALL(sys_fork_wrapper)
> > > +		CALL(sys_read)
> > > +		CALL(sys_write)
> > > +/* 5 */		CALL(sys_open)
> > > +		CALL(sys_close)
> > 
> > When you start using the generic unistd.h file, you can also replace this table
> > with something like arch/tile/kernel/sys.c.
> Well. I will use the generic unistd.h in UniCore-64 version.

I thought you had agreed to break ABI compatibility with your existing
code base and use the generic ABI everywhere.

Did I misunderstand you or did you make up your mind since then?

> > Hmm, when the architecture was being defined, why didn't you ask for a
> > cycle counter? It really improves the delay code a lot.
> No software readable cycle counter in UniCore32.
> 
> > If you have a good time base by now, you should use it. Is the OST_OSCR
> > something you could use here?
> OST_OSCR is much coarse in here, which is 14.318M Hz.

Ok, I see.

	Arnd