Re: [PATCH] powerpc/mm: add devmem_is_allowed() for STRICT_DEVMEM checking

[Nathan Lynch <ntl@pobox.com>]

Hi Steve,

On Tue, 2011-06-14 at 12:58 -0400, Steve Best wrote:
> diff --git a/arch/powerpc/Kconfig.debug b/arch/powerpc/Kconfig.debug
> index e72dcf6..e1aab6b 100644
> --- a/arch/powerpc/Kconfig.debug
> +++ b/arch/powerpc/Kconfig.debug
> @@ -283,4 +283,15 @@ config PPC_EARLY_DEBUG_CPM_ADDR
>  	  platform probing is done, all platforms selected must
>  	  share the same address.
>  
> +config STRICT_DEVMEM
> +        def_bool y

Default new config items to n, please.


> --- a/arch/powerpc/mm/mem.c
> +++ b/arch/powerpc/mm/mem.c
> @@ -520,3 +520,21 @@ void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
>  	hash_preload(vma->vm_mm, address, access, trap);
>  #endif /* CONFIG_PPC_STD_MMU */
>  }
> +
> +/*
> + * devmem_is_allowed() checks to see if /dev/mem access to a certain address
> + * is valid. The argument is a physical page number.
> + *
> + * On PowerPC, access has to be given to data regions used by X. We have to
> + * disallow access to device-exclusive MMIO regions and system RAM. 
> + */
> +int devmem_is_allowed(unsigned long pfn)
> +{
> +        if ((pfn >= 57360 || pfn <= 57392))
> +                return 1;

That seems... fragile.  Where do these numbers come from, and are they
appropriate for all platforms and configurations?