SGI O2+ RM7000 set_uncached_handler error during kernel built

SGI O2+ RM7000 set_uncached_handler error during kernel built

[Tom Weustink]

Hello,

I finally got a round building a toolchain for creating kernels for my 
SGI O2+ RM7000 machine.

I did a ip32_defconfig and used menuconfig to edit some small options.
One was to set it too RM7000 obviously.

When I build it, I get the following error:

  CC      arch/mips/kernel/traps.o
arch/mips/kernel/traps.c: In function `set_uncached_handler':
arch/mips/kernel/traps.c:1360: error: `TO_PHYS_MASK' undeclared (first 
use in this function)

When I build it for R5000 (only that changed in menuconfig) the compile 
runs without a problem.

Checking arch/mips/kernel/traps.c:1360 shows that installs the uncached 
CPU exception handler to UNCAC.
Now I know the RM7000 is kinda weird, but shouldn't it just work on it 
since it does work on R5000 since it's MIPS64?

I also heard from `Kumba that the kernel won't boot at all due to a bug 
in IP32 so that it hangs extremely early, but having it to just built 
would be nice for me atm.

Regards,

Tom

PS: if I need to supply any additional information, let me know!

Re: SGI O2+ RM7000 set_uncached_handler error during kernel built

[Ralf Baechle]

On Sat, May 27, 2006 at 04:26:17AM +0200, Tom Weustink wrote:

> I finally got a round building a toolchain for creating kernels for my 
> SGI O2+ RM7000 machine.
> 
> I did a ip32_defconfig and used menuconfig to edit some small options.
> One was to set it too RM7000 obviously.
> 
> When I build it, I get the following error:
> 
>  CC      arch/mips/kernel/traps.o
> arch/mips/kernel/traps.c: In function `set_uncached_handler':
> arch/mips/kernel/traps.c:1360: error: `TO_PHYS_MASK' undeclared (first 
> use in this function)
> 
> When I build it for R5000 (only that changed in menuconfig) the compile 
> runs without a problem.
> 
> Checking arch/mips/kernel/traps.c:1360 shows that installs the uncached 
> CPU exception handler to UNCAC.
> Now I know the RM7000 is kinda weird, but shouldn't it just work on it 
> since it does work on R5000 since it's MIPS64?
> 
> I also heard from `Kumba that the kernel won't boot at all due to a bug 
> in IP32 so that it hangs extremely early, but having it to just built 
> would be nice for me atm.

The build error part was trivial, fix below.

  Ralf

commit 948a25c865f2c486f3f2a6034c359c2104c777df
Author: Ralf Baechle <ralf@linux-mips.org>
Date:   Tue May 30 15:55:05 2006 +0100

    [MIPS] Fix 64-bit build for RM7000.
    
    RM7000 has 40-bit virtual / 36-bit physical address space.
    
    Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

diff --git a/include/asm-mips/addrspace.h b/include/asm-mips/addrspace.h
index 42520cc..1386af1 100644
--- a/include/asm-mips/addrspace.h
+++ b/include/asm-mips/addrspace.h
@@ -129,6 +129,7 @@ #define PHYS_TO_XKPHYS(cm,a)		(_LLCONST_
 #if defined (CONFIG_CPU_R4300)						\
     || defined (CONFIG_CPU_R4X00)					\
     || defined (CONFIG_CPU_R5000)					\
+    || defined (CONFIG_CPU_RM7000)					\
     || defined (CONFIG_CPU_NEVADA)					\
     || defined (CONFIG_CPU_TX49XX)					\
     || defined (CONFIG_CPU_MIPS64)

Re: Re: SGI O2+ RM7000 set_uncached_handler error during kernel built

[Tom Weustink]

> 
> > I finally got a round building a toolchain for creating kernels for my 
> > SGI O2+ RM7000 machine.
> > 
> > I did a ip32_defconfig and used menuconfig to edit some small options.
> > One was to set it too RM7000 obviously.
> > 
> > When I build it, I get the following error:
> > 
> >  CC      arch/mips/kernel/traps.o
> > arch/mips/kernel/traps.c: In function `set_uncached_handler':
> > arch/mips/kernel/traps.c:1360: error: `TO_PHYS_MASK' undeclared (first 
> > use in this function)
> > 
> > When I build it for R5000 (only that changed in menuconfig) the compile 
> > runs without a problem.
> > 
> > Checking arch/mips/kernel/traps.c:1360 shows that installs the uncached 
> > CPU exception handler to UNCAC.
> > Now I know the RM7000 is kinda weird, but shouldn't it just work on it 
> > since it does work on R5000 since it's MIPS64?
> > 
> > I also heard from `Kumba that the kernel won't boot at all due to a bug 
> > in IP32 so that it hangs extremely early, but having it to just built 
> > would be nice for me atm.
> 
> The build error part was trivial, fix below.
> 
>   Ralf
> 
Great! I will try it out tonight.
I kind of had the idea something pointing to RM7000 was missing, obviously I had no idea where though.

Tom


> commit 948a25c865f2c486f3f2a6034c359c2104c777df
> Author: Ralf Baechle <ralf@linux-mips.org>
> Date:   Tue May 30 15:55:05 2006 +0100
> 
>     [MIPS] Fix 64-bit build for RM7000.
>     
>     RM7000 has 40-bit virtual / 36-bit physical address space.
>     
>     Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
> 
> diff --git a/include/asm-mips/addrspace.h b/include/asm-mips/addrspace.h
> index 42520cc..1386af1 100644
> --- a/include/asm-mips/addrspace.h
> +++ b/include/asm-mips/addrspace.h
> @@ -129,6 +129,7 @@ #define PHYS_TO_XKPHYS(cm,a)		(_LLCONST_
>  #if defined (CONFIG_CPU_R4300)						\
>      || defined (CONFIG_CPU_R4X00)					\
>      || defined (CONFIG_CPU_R5000)					\
> +    || defined (CONFIG_CPU_RM7000)					\
>      || defined (CONFIG_CPU_NEVADA)					\
>      || defined (CONFIG_CPU_TX49XX)					\
>      || defined (CONFIG_CPU_MIPS64)

-- 


Bis zu 70% Ihrer Onlinekosten sparen: GMX SmartSurfer!
      Kostenlos downloaden: http://www.gmx.net/de/go/smartsurfer