Re: [PATCH v3 05/13] m68k: mm: use pgtable-nopXd instead of 4level-fixup

[Geert Uytterhoeven <geert@linux-m68k.org>]

Hi Mike,

On Mon, Nov 4, 2019 at 7:57 AM Mike Rapoport <rppt@kernel.org> wrote:
> From: Mike Rapoport <rppt@linux.ibm.com>
> m68k has two or three levels of page tables and can use appropriate
> pgtable-nopXd and folding of the upper layers.
>
> Replace usage of include/asm-generic/4level-fixup.h and explicit
> definitions of __PAGETABLE_PxD_FOLDED in m68k with
> include/asm-generic/pgtable-nopmd.h for two-level configurations and with
> include/asm-generic/pgtable-nopud.h for three-lelve configurations and
> adjust page table manipulation macros and functions accordingly.
>
> Signed-off-by: Mike Rapoport <rppt@linux.ibm.com>
> Acked-by: Greg Ungerer <gerg@linux-m68k.org>

Thanks for your patch!

The build error reported for v1 by kbuild test robot when building for
sun3x is still there (m68k defconfig or sun3x_defconfig):

    arch/m68k/sun3x/dvma.c: In function ‘dvma_map_cpu’:
    arch/m68k/sun3x/dvma.c:98:33: error: passing argument 2 of
‘pmd_alloc’ from incompatible pointer type
[-Werror=incompatible-pointer-types]
       if((pmd = pmd_alloc(&init_mm, pgd, vaddr)) == NULL) {
                                     ^~~
    In file included from arch/m68k/sun3x/dvma.c:17:
    include/linux/mm.h:1875:61: note: expected ‘pud_t *’ {aka ‘struct
<anonymous> *’} but argument is of type ‘pgd_t *’ {aka ‘struct
<anonymous> *’}
     static inline pmd_t *pmd_alloc(struct mm_struct *mm, pud_t *pud,
unsigned long address)
                                                          ~~~~~~~^~~

This indeed boots fine on ARAnyM, which emulates on 68040.
It would be good to have some boot testing on '020/030, too.

> --- a/arch/m68k/mm/kmap.c
> +++ b/arch/m68k/mm/kmap.c

> @@ -196,17 +198,21 @@ void __iomem *__ioremap(unsigned long physaddr, unsigned long size, int cachefla
>                         printk ("\npa=%#lx va=%#lx ", physaddr, virtaddr);
>  #endif
>                 pgd_dir = pgd_offset_k(virtaddr);
> -               pmd_dir = pmd_alloc(&init_mm, pgd_dir, virtaddr);
> +               p4d_dir = p4d_offset(pgd_dir, virtaddr);
> +               pud_dir = pud_offset(p4d_dir, virtaddr);
> +               pmd_dir = pmd_alloc(&init_mm, pud_dir, virtaddr);
>                 if (!pmd_dir) {
>                         printk("ioremap: no mem for pmd_dir\n");
>                         return NULL;
>                 }
>
>                 if (CPU_IS_020_OR_030) {
> +#if CONFIG_PGTABLE_LEVELS == 3

This check puzzled me a bit: when we get here, CONFIG_PGTABLE_LEVELS is
always true.
However, the check cannot be removed, as the code it protects fails to compile
when building for Coldfire.

Perhaps this can be made more clear by reverting the order?
I.e.

    #if CONFIG_PGTABLE_LEVELS == 3
            if (CPU_IS_020_OR_030) {
                    ...
            } else
    #endif
            {

Or is there some better way?

>                         pmd_dir->pmd[(virtaddr/PTRTREESIZE) & 15] = physaddr;
>                         physaddr += PTRTREESIZE;
>                         virtaddr += PTRTREESIZE;
>                         size -= PTRTREESIZE;
> +#endif
>                 } else {
>                         pte_dir = pte_alloc_kernel(pmd_dir, virtaddr);
>                         if (!pte_dir) {
> @@ -258,19 +264,24 @@ void __iounmap(void *addr, unsigned long size)
>  {
>         unsigned long virtaddr = (unsigned long)addr;
>         pgd_t *pgd_dir;
> +       p4d_t *p4d_dir;
> +       pud_t *pud_dir;
>         pmd_t *pmd_dir;
>         pte_t *pte_dir;
>
>         while ((long)size > 0) {
>                 pgd_dir = pgd_offset_k(virtaddr);
> -               if (pgd_bad(*pgd_dir)) {
> -                       printk("iounmap: bad pgd(%08lx)\n", pgd_val(*pgd_dir));
> -                       pgd_clear(pgd_dir);
> +               p4d_dir = p4d_offset(pgd_dir, virtaddr);
> +               pud_dir = pud_offset(p4d_dir, virtaddr);
> +               if (pud_bad(*pud_dir)) {
> +                       printk("iounmap: bad pgd(%08lx)\n", pud_val(*pud_dir));
> +                       pud_clear(pud_dir);
>                         return;
>                 }
> -               pmd_dir = pmd_offset(pgd_dir, virtaddr);
> +               pmd_dir = pmd_offset(pud_dir, virtaddr);
>
>                 if (CPU_IS_020_OR_030) {
> +#if CONFIG_PGTABLE_LEVELS == 3

Likewise.

>                         int pmd_off = (virtaddr/PTRTREESIZE) & 15;
>                         int pmd_type = pmd_dir->pmd[pmd_off] & _DESCTYPE_MASK;
>

> @@ -341,14 +355,17 @@ void kernel_set_cachemode(void *addr, unsigned long size, int cmode)
>
>         while ((long)size > 0) {
>                 pgd_dir = pgd_offset_k(virtaddr);
> -               if (pgd_bad(*pgd_dir)) {
> -                       printk("iocachemode: bad pgd(%08lx)\n", pgd_val(*pgd_dir));
> -                       pgd_clear(pgd_dir);
> +               p4d_dir = p4d_offset(pgd_dir, virtaddr);
> +               pud_dir = pud_offset(p4d_dir, virtaddr);
> +               if (pud_bad(*pud_dir)) {
> +                       printk("iocachemode: bad pud(%08lx)\n", pud_val(*pud_dir));
> +                       pud_clear(pud_dir);
>                         return;
>                 }
> -               pmd_dir = pmd_offset(pgd_dir, virtaddr);
> +               pmd_dir = pmd_offset(pud_dir, virtaddr);
>
>                 if (CPU_IS_020_OR_030) {
> +#if CONFIG_PGTABLE_LEVELS == 3

Likewise

>                         int pmd_off = (virtaddr/PTRTREESIZE) & 15;
>
>                         if ((pmd_dir->pmd[pmd_off] & _DESCTYPE_MASK) == _PAGE_PRESENT) {

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

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