[Suggestion] Latest randconfig build errors for CONFIG_MMU=no

[Will Deacon]

On Thu, Mar 14, 2013 at 01:25:56PM +0000, Jonathan Austin wrote:
> On 14/03/13 03:44, Chen Gang wrote:
> > Hello Maintainers:
> >
> >    when use randconfig:
> >      make V=1 EXTRA_CFLAGS=-W ARCH=arm randconfig
> >        (and "grep CONFIG_MMU .config" outputs "# CONFIG_MMU is not set")
> >
> >    the error information:
> >      arch/arm/kernel/suspend.c: 在函数‘cpu_suspend’中:
> >      arch/arm/kernel/suspend.c:56:32: 错误：提领指向不完全类型的指针
> >      arch/arm/kernel/suspend.c:70:3: 错误：隐式声明函数‘cpu_switch_mm’ [-Werror=implicit-function-declaration]
> >      arch/arm/kernel/suspend.c:71:3: 错误：隐式声明函数‘local_flush_bp_all’ [-Werror=implicit-function-declaration]
> >      arch/arm/kernel/suspend.c:72:3: 错误：隐式声明函数‘local_flush_tlb_all’ [-Werror=implicit-function-declaration]
> >
> >    I want to be sure whether we can bear it ?
> >
> 
> You're right that at the moment that suspend code is broken for NOMMU.
> 
> However, Will Deacon has a patch to fix it in his nommu branch on 
> kernel.org:
> 
> https://git.kernel.org/cgit/linux/kernel/git/will/linux.git/commit/?h=nommu
> 
> I'll let Will comment on how/when he might look at proposing that code 
> to be merged.

I can probably bring that up to speed with -rc2 if people are interested.
The tricky part is testing -- I used a software model that had a magic
control block for killing off CPUs when I wrote this originally.

Will