Re: [PATCH v5 07/17] arm64: add basic pointer authentication support

[Suzuki K Poulose]

Hi Kristina,

On 05/10/18 09:47, Kristina Martsenko wrote:
> From: Mark Rutland <mark.rutland@xxxxxxx>
>
> This patch adds basic support for pointer authentication, allowing
> userspace to make use of APIAKey. The kernel maintains an APIAKey value
> for each process (shared by all threads within), which is initialised to
> a random value at exec() time.
>
> To describe that address authentication instructions are available, the
> ID_AA64ISAR0.{APA,API} fields are exposed to userspace. A new hwcap,
> APIA, is added to describe that the kernel manages APIAKey.
>
> Instructions using other keys (APIBKey, APDAKey, APDBKey) are disabled,
> and will behave as NOPs. These may be made use of in future patches.
>
> No support is added for the generic key (APGAKey), though this cannot be
> trapped or made to behave as a NOP. Its presence is not advertised with
> a hwcap.
>
> Signed-off-by: Mark Rutland <mark.rutland@xxxxxxx>
> [kristina: init keys in arch_bprm_mm_init; add AA64ISAR1.API HWCAP_CAP; use sysreg_clear_set]
> Signed-off-by: Kristina Martsenko <kristina.martsenko@xxxxxxx>
> Tested-by: Adam Wallis <awallis@xxxxxxxxxxxxxx>
> Cc: Catalin Marinas <catalin.marinas@xxxxxxx>
> Cc: Ramana Radhakrishnan <ramana.radhakrishnan@xxxxxxx>
> Cc: Suzuki K Poulose <suzuki.poulose@xxxxxxx>
> Cc: Will Deacon <will.deacon@xxxxxxx>

> diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
> index 0dd171c7d71e..3157685aa56a 100644
> --- a/arch/arm64/kernel/cpufeature.c
> +++ b/arch/arm64/kernel/cpufeature.c
> @@ -1040,6 +1040,11 @@ static void cpu_has_fwb(const struct arm64_cpu_capabilities *__unused)
>   }
>   
>   #ifdef CONFIG_ARM64_PTR_AUTH
> +static void cpu_enable_address_auth(struct arm64_cpu_capabilities const *cap)
> +{
> +	sysreg_clear_set(sctlr_el1, 0, SCTLR_ELx_ENIA);
> +}
> +
>   static bool has_address_auth(const struct arm64_cpu_capabilities *entry,
>   			     int __unused)
>   {
> @@ -1267,6 +1272,7 @@ static const struct arm64_cpu_capabilities arm64_features[] = {
>   		.capability = ARM64_HAS_ADDRESS_AUTH,
>   		.type = ARM64_CPUCAP_SYSTEM_FEATURE,
>   		.matches = has_address_auth,
> +		.cpu_enable = cpu_enable_address_auth,
>   	},
>   #endif /* CONFIG_ARM64_PTR_AUTH */
>   	{},
> @@ -1314,6 +1320,10 @@ static const struct arm64_cpu_capabilities arm64_elf_hwcaps[] = {
>   #ifdef CONFIG_ARM64_SVE
>   	HWCAP_CAP(SYS_ID_AA64PFR0_EL1, ID_AA64PFR0_SVE_SHIFT, FTR_UNSIGNED, ID_AA64PFR0_SVE, CAP_HWCAP, HWCAP_SVE),
>   #endif
> +#ifdef CONFIG_ARM64_PTR_AUTH
> +	HWCAP_CAP(SYS_ID_AA64ISAR1_EL1, ID_AA64ISAR1_APA_SHIFT, FTR_UNSIGNED, 1, CAP_HWCAP, HWCAP_APIA),
> +	HWCAP_CAP(SYS_ID_AA64ISAR1_EL1, ID_AA64ISAR1_API_SHIFT, FTR_UNSIGNED, 1, CAP_HWCAP, HWCAP_APIA),

This is a bit problematic. If all the CPUs have just the IMPDEF
algorithm available, we could fail to match the first entry (APA) for a
late secondary CPU and thus thus fail the CPU from booting. I guess we 
need a custom entry which reuses the has_address_auth() as the matches().

Rest looks fine to me.

Suzuki