Re: [PATCH] ARM: small correction to early_ioremap support

[Doug Berger]

On 06/26/2017 05:57 PM, Doug Berger wrote:
> The fixmap pages need to be on the same pmd page for the current
> implementation of arm early_fixmap.  A build time bug check is
> used to ensure that.  However, the worst case fixmap range is
> better represented by __end_of_fixed_addresses than by the value
> __end_of_early_ioremap_region.
I see now that only the early_ioremap portion of the fixmap needs to be
on the same pmd page and not the entire collection of fixmap pages.

Therefore the check against __end_of_early_ioremap_region is the correct
choice for this build time check.

Please disregard this patch.

Sorry for the confusion,
    Doug

> 
> Fixes: 2937367b8a4b ("ARM: add support for generic early_ioremap/early_memremap")
> Signed-off-by: Doug Berger <opendmb@xxxxxxxxx>
> ---
>  arch/arm/mm/mmu.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/arm/mm/mmu.c b/arch/arm/mm/mmu.c
> index 31af3cb59a60..74c0ed5c3b08 100644
> --- a/arch/arm/mm/mmu.c
> +++ b/arch/arm/mm/mmu.c
> @@ -392,7 +392,7 @@ void __init early_fixmap_init(void)
>  	 * The early fixmap range spans multiple pmds, for which
>  	 * we are not prepared:
>  	 */
> -	BUILD_BUG_ON((__fix_to_virt(__end_of_early_ioremap_region) >> PMD_SHIFT)
> +	BUILD_BUG_ON((__fix_to_virt(__end_of_fixed_addresses) >> PMD_SHIFT)
>  		     != FIXADDR_TOP >> PMD_SHIFT);
>  
>  	pmd = fixmap_pmd(FIXADDR_TOP);
>