[PATCH] m68k/sun3: Fix virtual addresses of clock and interrupt register

[PATCH] m68k/sun3: Fix virtual addresses of clock and interrupt register

[Alexander Kuleshov]

The MMU base is 32 bits size - 0xfe000000, seems that we missed
one zero in the definition of the clock and interrupt register
addresses.

Signed-off-by: Alexander Kuleshov <kuleshovmail@gmail.com>
---
 arch/m68k/sun3/config.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/m68k/sun3/config.c b/arch/m68k/sun3/config.c
index a8b942b..c9c388f 100644
--- a/arch/m68k/sun3/config.c
+++ b/arch/m68k/sun3/config.c
@@ -61,8 +61,8 @@ void __init sun3_init(void)
 	m68k_cputype = CPU_68020;
 	m68k_fputype = FPU_68881; /* mc68881 actually */
 	m68k_mmutype = MMU_SUN3;
-	clock_va    =          (char *) 0xfe06000;	/* dark  */
-	sun3_intreg = (unsigned char *) 0xfe0a000;	/* magic */
+	clock_va    =          (char *) 0xfe006000;	/* dark  */
+	sun3_intreg = (unsigned char *) 0xfe00a000;	/* magic */
 	sun3_disable_interrupts();
 
 	prom_init((void *)LINUX_OPPROM_BEGVM);
-- 
2.5.0

Re: [PATCH] m68k/sun3: Fix virtual addresses of clock and interrupt register

[Geert Uytterhoeven]

Hi Alexander,

On Mon, Sep 14, 2015 at 11:40 AM, Alexander Kuleshov
<kuleshovmail@gmail.com> wrote:
> The MMU base is 32 bits size - 0xfe000000, seems that we missed
> one zero in the definition of the clock and interrupt register
> addresses.

Thanks for your patch!

Have you tested this?

This means enabling/disabling interrupts never worked on Sun 3, which
I find a bit difficult to believe.
There are other 0x0f?????? addresses in arch/m68k/sun3/mmu_emu.c.

> Signed-off-by: Alexander Kuleshov <kuleshovmail@gmail.com>
> ---
>  arch/m68k/sun3/config.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/arch/m68k/sun3/config.c b/arch/m68k/sun3/config.c
> index a8b942b..c9c388f 100644
> --- a/arch/m68k/sun3/config.c
> +++ b/arch/m68k/sun3/config.c
> @@ -61,8 +61,8 @@ void __init sun3_init(void)
>         m68k_cputype = CPU_68020;
>         m68k_fputype = FPU_68881; /* mc68881 actually */
>         m68k_mmutype = MMU_SUN3;
> -       clock_va    =          (char *) 0xfe06000;      /* dark  */
> -       sun3_intreg = (unsigned char *) 0xfe0a000;      /* magic */
> +       clock_va    =          (char *) 0xfe006000;     /* dark  */
> +       sun3_intreg = (unsigned char *) 0xfe00a000;     /* magic */
>         sun3_disable_interrupts();

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH] m68k/sun3: Fix virtual addresses of clock and interrupt register

[Alexander Kuleshov]

Hello Geert

On Fri, Sep 25, 2015 at 1:21 PM, Geert Uytterhoeven
<geert@linux-m68k.org> wrote:
> Have you tested this?
>
> This means enabling/disabling interrupts never worked on Sun 3, which
> I find a bit difficult to believe.
> There are other 0x0f?????? addresses in arch/m68k/sun3/mmu_emu.c.
>

Nope, just trying to do it in virtual machine, I'm learning sun3
architecture now
and have some cleanups patches for it. I'll send these patches and provide more
feedback when I'm able to test it.

Re: [PATCH] m68k/sun3: Fix virtual addresses of clock and interrupt register

[Tom Bogendoerfer]

On Fri, Sep 25, 2015 at 09:21:34AM +0200, Geert Uytterhoeven wrote:
> On Mon, Sep 14, 2015 at 11:40 AM, Alexander Kuleshov
> <kuleshovmail@gmail.com> wrote:
> > The MMU base is 32 bits size - 0xfe000000, seems that we missed
> > one zero in the definition of the clock and interrupt register
> > addresses.
> 
> Thanks for your patch!
> 
> Have you tested this?
> 
> This means enabling/disabling interrupts never worked on Sun 3, which
> I find a bit difficult to believe.
> There are other 0x0f?????? addresses in arch/m68k/sun3/mmu_emu.c.

which are the correct one. Sun3 MMU doesn't even handle full 32bit
addresses, it's only capable of 28 bits. That's the reason why normal
m68k linux user space binaries can't work.

So this patch is broken. And I'm sure the unpatched version works.

Thomas.

-- 
Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]