Hi Greg,
On Fri, Sep 10, 2010 at 06:36, Greg Ungerer <[email protected]> wrote:
> Geert Uytterhoeven wrote:
>> On Wed, Sep 8, 2010 at 13:43, Greg Ungerer <[email protected]> wrote:
>>> On 08/09/10 06:35, Geert Uytterhoeven wrote:
>>>>
>>>> Now I see:
>>>>
>>>> | arch/m68k/kernel/entry.S: Assembler messages:
>>>> | arch/m68k/kernel/entry.S:56: Error: can't resolve `0' {*ABS* section}
>>>> - `THREAD_SIZE' {*UND* section}
>>>> | arch/m68k/kernel/entry.S:64: Error: can't resolve `0' {*UND* section}
>>>> - `THREAD_SIZE' {*UND* section}
>>>> | arch/m68k/kernel/entry.S:119: Error: can't resolve `0' {*UND* section}
>>>> - `THREAD_SIZE' {*UND* section}
>>>> | arch/m68k/kernel/entry.S:203: Error: can't resolve `0' {*UND* section}
>>>> - `THREAD_SIZE' {*UND* section}
>>>> | arch/m68k/kernel/entry.S:236: Error: can't resolve `0' {*UND* section}
>>>> - `THREAD_SIZE' {*UND* section}
>>>> | arch/m68k/kernel/entry.S:257: Error: can't resolve `0' {*UND* section}
>>>> - `THREAD_SIZE' {*UND* section}
>>>>
>>>> this is expanded from `GET_CURRENT'.
>>>
>>> Just a thought, is it enough to include thread_info.h inside entry_mm.h?
>>
>> No, because <asm/thread_info.h> is no longer ASSEMBLY-clean, as it
>> includes
>> <asm/current.h>, which is C-only.
>> The #include <asm/current.h> in there used to be protected by #ifndef
>> ASSEMBLY.
>> If I move that include up just above the definition of struct thread_info,
>> <asm/thread_info.h> is ASSEMBLY-clean again.
>>
>> However, just including <asm/thread_info.h> from <asm/entry_mm.h> causes
>> other
>> problems for the C part, as <asm/thread_info.h> needs the definition of
>> struct
>> restart_block in <linux/thread_info.h>. And <linux/thread_info.h> is
>> definitely
>> not ASSEMBLY-clean.
>>
>> So I came up with the patch (white-space damaged due to the GMail web
>> interface)
>> below. Note that this changes the behavior for m68knommu, which didn't
>> include
>> <asm/current.h> from <asm/thread_info.h> before. I hope it doesn't hurt...
>
> Unfortunately that breaks the m68knommu build:
>
> /home/gerg/new-wave.2635/linux-2.6.x.git/arch/m68k/include/asm/current.h: In
> function ‘get_current’:
> /home/gerg/new-wave.2635/linux-2.6.x.git/arch/m68k/include/asm/current.h:21:
> error: implicit declaration of function ‘current_thread_info’
> ...
>
> If we just wrap the include of asm/current.h with its own "#ifndef
> __ASSEMBLY__", instead of moving it from above struct thread_info,
> is that better for you?
Builds fine, too.
>> diff --git a/arch/m68k/include/asm/entry_mm.h
>> b/arch/m68k/include/asm/entry_mm.h
>> index 6f70823..73b8c8f 100644
>> --- a/arch/m68k/include/asm/entry_mm.h
>> +++ b/arch/m68k/include/asm/entry_mm.h
>> @@ -3,6 +3,9 @@
>>
>> #include <asm/setup.h>
>> #include <asm/page.h>
>> +#ifdef __ASSEMBLY__
>> +#include <asm/thread_info.h>
>> +#endif
>>
>> /*
>> * Stack layout in 'ret_from_exception':
>> diff --git a/arch/m68k/include/asm/thread_info.h
>> b/arch/m68k/include/asm/thread_info.h
>> index f28775e..aec5eff 100644
>> --- a/arch/m68k/include/asm/thread_info.h
>> +++ b/arch/m68k/include/asm/thread_info.h
>> @@ -22,6 +22,8 @@
>>
>> #ifndef __ASSEMBLY__
>>
>> +#include <asm/current.h>
>> +
>> struct thread_info {
>> struct task_struct *task; /* main task structure */
>> unsigned long flags;
>> @@ -49,8 +51,6 @@ struct thread_info {
>>
>> #ifdef CONFIG_MMU
>>
>> -#include <asm/current.h>
>> -
>> #ifdef ASM_OFFSETS_C
>> #define task_thread_info(tsk) ((struct thread_info *) NULL)
>> #else
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]
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
--
To unsubscribe from this list: send the line "unsubscribe linux-m68k" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
