Re: [PATCH v2 07/17] kvm-arm: arm32: Introduce stage2 page table helpers

From: Marc Zyngier
Date: Thu Apr 14 2016 - 10:23:07 EST


On 14/04/16 14:20, Suzuki K Poulose wrote:
> Define the page table helpers for walking the stage2 pagetable
> for arm. Since both hyp and stage2 have the same number of levels,
> as that of the host we reuse the host helpers.
>
> The exceptions are the p.d_addr_end routines which have to deal
> with IPA > 32bit, hence we use the open coded version of their host helpers
> which supports 64bit.
>
> Cc: Marc Zyngier <marc.zyngier@xxxxxxx>
> Reviewed-by: Christoffer Dall <christoffer.dall@xxxxxxxxxx>
> Signed-off-by: Suzuki K Poulose <suzuki.poulose@xxxxxxx>
> ---
> Changes since v1:
> - Drop mm (which is always NULL) parameter to stage2_p.d_xxx. Pass NULL
> explicitly to the host helpers.
> ---
> arch/arm/include/asm/kvm_mmu.h | 1 +
> arch/arm/include/asm/stage2_pgtable.h | 59 +++++++++++++++++++++++++++++++++
> 2 files changed, 60 insertions(+)
> create mode 100644 arch/arm/include/asm/stage2_pgtable.h
>
> diff --git a/arch/arm/include/asm/kvm_mmu.h b/arch/arm/include/asm/kvm_mmu.h
> index c2b2b27..7d207b4 100644
> --- a/arch/arm/include/asm/kvm_mmu.h
> +++ b/arch/arm/include/asm/kvm_mmu.h
> @@ -47,6 +47,7 @@
> #include <linux/highmem.h>
> #include <asm/cacheflush.h>
> #include <asm/pgalloc.h>
> +#include <asm/stage2_pgtable.h>
>
> int create_hyp_mappings(void *from, void *to);
> int create_hyp_io_mappings(void *from, void *to, phys_addr_t);
> diff --git a/arch/arm/include/asm/stage2_pgtable.h b/arch/arm/include/asm/stage2_pgtable.h
> new file mode 100644
> index 0000000..5b61196
> --- /dev/null
> +++ b/arch/arm/include/asm/stage2_pgtable.h
> @@ -0,0 +1,59 @@
> +/*
> + * Copyright (C) 2016 - ARM Ltd
> + *
> + * stage2 page table helpers
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License version 2 as
> + * published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program. If not, see <http://www.gnu.org/licenses/>.
> + */
> +
> +#ifndef __ARM_S2_PGTABLE_H_
> +#define __ARM_S2_PGTABLE_H_
> +
> +#define stage2_pgd_none(pgd) pgd_none(pgd)
> +#define stage2_pgd_clear(pgd) pgd_clear(pgd)
> +#define stage2_pgd_present(pgd) pgd_present(pgd)
> +#define stage2_pgd_populate(pgd, pud) pgd_populate(NULL, pgd, pud)
> +#define stage2_pud_offset(pgd, address) pud_offset(pgd, address)
> +#define stage2_pud_free(pud) pud_free(NULL, pud)
> +
> +#define stage2_pud_none(pud) pud_none(pud)
> +#define stage2_pud_clear(pud) pud_clear(pud)
> +#define stage2_pud_present(pud) pud_present(pud)
> +#define stage2_pud_populate(pud, pmd) pud_populate(NULL, pud, pmd)
> +#define stage2_pmd_offset(pud, address) pmd_offset(pud, address)
> +#define stage2_pmd_free(pmd) pmd_free(NULL, pmd)
> +
> +#define stage2_pud_huge(pud) pud_huge(pud)
> +
> +/* Open coded p*d_addr_end that can deal with 64bit addresses */
> +static inline phys_addr_t stage2_pgd_addr_end(phys_addr_t addr, phys_addr_t end)
> +{
> + phys_addr_t boundary = (addr + PGDIR_SIZE) & PGDIR_MASK;
> + return (boundary - 1 < end - 1) ? boundary : end;
> +}
> +
> +#define stage2_pud_addr_end(addr, end) (end)
> +
> +static inline phys_addr_t stage2_pmd_addr_end(phys_addr_t addr, phys_addr_t end)
> +{
> + phys_addr_t boundary = (addr + PMD_SIZE) & PMD_MASK;
> + return (boundary - 1 < end - 1) ? boundary : end;
> +}
> +
> +#define stage2_pgd_index(addr) pgd_index(addr)
> +
> +#define stage2_pte_table_empty(ptep) kvm_page_empty(ptep)
> +#define stage2_pmd_table_empty(pmdp) kvm_page_empty(pmdp)
> +#define stage2_pud_table_empty(pudp) 0

nit: given that the kvm_page_empty returns a boolean, I'd have defined
this to 'false' instead of zero. No need to respin for this though.

Another concern is that while stage2_* looks pretty generic, the
implementation is very KVM specific. As long as KVM is the only
in-kernel hypervisor, it doesn't matter.

> +
> +#endif /* __ARM_S2_PGTABLE_H_ */
>

Irrespective of the above:

Acked-by: Marc Zyngier <marc.zyngier@xxxxxxx>

M.
--
Jazz is not dead. It just smells funny...