> [EMAIL PROTECTED] writes:
> > > >There is one thing I dont understand: You initialise cr1 using r0, but
> > > >r0 is not initialised in order to set MMU , is it ?
> > >
> > > Yes, it's loaded by this subroutine call (which winds up at
> > > .Larmv4_flush_early)
> > >
> > > > mov pc, r10 @ Call processo
> > > >r flush
> > > >(returns ctrl reg)
> >
> > Ok, so r0 is correctly initialised.
> > Do you see any reason why I get the M printed on my terminal
> > before switching MMU on, and I dont get the m after MMU on ?
> >
> > The progs continues at .Lalready_done_mmap, doesn't it ?
>
> I've just caught up on this discussion on the armlinux mailing
> list. I have the same problem :-( I'm a bit further along than
> you now that I've written the beginnings of a debugger (I wish
> I was joking..) The problem is in console_init, specifically the
> test to see if the UART exists in serial_console_setup ...
> On my hardware, and I'm guessing on yours, this doesn't exist - I
> want to use the UART on the 21285....
>
So it means that you arrive at serial_console_setup.
My problem for now resides before realy starting the kernel, in
head-armv.S
I'm able to write things on serial by writing bytes at 0x42000160
(after been sure UART is free by testing 0x42000178) when MMU is OFF.
But When I try to write things on serial just after having switch MMU
ON ,in the beginning of .Lalready_done_mmap (So I write at
0xFE000160, which is the supposed virtual adresse of UART), nothing
happens. So I wonder what happens after having put MMU ON.
As you seem to have more experience than me on it, maybe have you an
idea of what could be wrong at that point .
Christophe
unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
