[PATCH 7/8] arm64: mm: Implement 4 levels of translation tables

[Jungseok Lee]

On Tuesday, April 15, 2014 12:14 AM, Steve Capper wrote:
> On Mon, Apr 14, 2014 at 04:41:07PM +0900, Jungseok Lee wrote:
> > This patch implements 4 levels of translation tables since 3 levels of
> > page tables with 4KB pages cannot support 40-bit physical address
> > space described in [1] due to the following issue.
> >
> > It is a restriction that kernel logical memory map with 4KB + 3 levels
> > (0xffffffc000000000-0xffffffffffffffff) cannot cover RAM region from
> > 544GB to 1024GB in [1]. Specifically, ARM64 kernel fails to create
> > mapping for this region in map_mem function since __phys_to_virt for
> > this region reaches to address overflow.
> >
> > If SoC design follows the document, [1], over 32GB RAM would be placed
> > from 544GB. Even 64GB system is supposed to use the region from 544GB
> > to 576GB for only 32GB RAM. Naturally, it would reach to enable 4
> > levels of page tables to avoid hacking __virt_to_phys and __phys_to_virt.
> >
> > However, it is recommended 4 levels of page table should be only
> > enabled if memory map is too sparse or there is about 512GB RAM.
> >
> > References
> > ----------
> > [1]: Principle of ARM Memory Maps, White Paper, Issue C
> >

[ ... ]

> > diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c index
> > 6b7e895..321f569 100644
> > --- a/arch/arm64/mm/mmu.c
> > +++ b/arch/arm64/mm/mmu.c
> > @@ -222,9 +222,17 @@ static void __init alloc_init_pmd(pud_t *pud,
> > unsigned long addr,  static void __init alloc_init_pud(pgd_t *pgd, unsigned long addr,
> >  				  unsigned long end, unsigned long phys)  {
> > -	pud_t *pud = pud_offset(pgd, addr);
> > +	pud_t *pud;
> >  	unsigned long next;
> >
> > +#ifdef CONFIG_ARM64_4_LEVELS
> > +	if (pgd_none(*pgd) || pgd_bad(*pgd)) {
> > +		pud = early_alloc(PTRS_PER_PUD * sizeof(pud_t));
> > +		pgd_populate(&init_mm, pgd, pud);
> > +	}
> > +#endif
> 
> We don't need this #ifdef block, as pgd_none and pgd_bad should be zero when we have fewer than 4
> levels.

This block is needed to cover the third RAM region from 544GB to 1024GB
described in the document [1].

A single PGD can cover only up to 512GB with 4KB+4Level. In other words,
kernel would reach to panic if a system has RAM over 512GB memory map space.
That is why pgd_none should be handled.

Best Regards
Jungseok Lee