Re: [PATCH v5 11/19] arm64: rsi: Map unprotected MMIO as decrypted

[Suzuki K Poulose <suzuki.poulose@arm.com>]

Hi Steven

On 19/08/2024 14:19, Steven Price wrote:
> From: Suzuki K Poulose <suzuki.poulose@arm.com>
> 
> Instead of marking every MMIO as shared, check if the given region is
> "Protected" and apply the permissions accordingly.
> 
> Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
> Signed-off-by: Steven Price <steven.price@arm.com>
> ---
> New patch for v5
> ---
>   arch/arm64/kernel/rsi.c | 15 +++++++++++++++
>   arch/arm64/mm/mmu.c     |  2 +-
>   2 files changed, 16 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/arm64/kernel/rsi.c b/arch/arm64/kernel/rsi.c
> index 381a5b9a5333..672dd6862298 100644
> --- a/arch/arm64/kernel/rsi.c
> +++ b/arch/arm64/kernel/rsi.c
> @@ -6,6 +6,8 @@
>   #include <linux/jump_label.h>
>   #include <linux/memblock.h>
>   #include <linux/psci.h>
> +
> +#include <asm/io.h>
>   #include <asm/rsi.h>
>   
>   struct realm_config config;
> @@ -93,6 +95,16 @@ bool arm64_rsi_is_protected_mmio(phys_addr_t base, size_t size)
>   }
>   EXPORT_SYMBOL(arm64_rsi_is_protected_mmio);
>   
> +static int realm_ioremap_hook(phys_addr_t phys, size_t size, pgprot_t *prot)
> +{
> +	if (arm64_rsi_is_protected_mmio(phys, size))
> +		*prot = pgprot_encrypted(*prot);
> +	else
> +		*prot = pgprot_decrypted(*prot);
> +
> +	return 0;
> +}
> +
>   void __init arm64_rsi_init(void)
>   {
>   	/*
> @@ -107,6 +119,9 @@ void __init arm64_rsi_init(void)
>   		return;
>   	prot_ns_shared = BIT(config.ipa_bits - 1);
>   
> +	if (arm64_ioremap_prot_hook_register(realm_ioremap_hook))
> +		return;
> +
>   	static_branch_enable(&rsi_present);
>   }
>   
> diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c
> index 06b66c23c124..0c2fa35beca0 100644
> --- a/arch/arm64/mm/mmu.c
> +++ b/arch/arm64/mm/mmu.c
> @@ -1207,7 +1207,7 @@ void set_fixmap_io(enum fixed_addresses idx, phys_addr_t phys)
>   	else
>   		prot = pgprot_encrypted(prot);
>   
> -	__set_fixmap(idx, phys, prot);
> +	__set_fixmap(idx, phys & PAGE_MASK, prot);

This looks like it should be part of the previous patch ? Otherwise 
looks good to me.

Suzuki