[PATCH] arm64: errata: Fix exec handling in erratum 1418040 workaround

[Marc Zyngier]

On Wed, 15 Dec 2021 23:02:40 +0000,
D Scott Phillips <scott at os.amperecomputing.com> wrote:
> 
> The erratum 1418040 workaround changes vct access trapping when switching
> between compat and non-compat threads. The workaround logic assumes that
> the hardware vct trapping state matches the previous task's compat-ness.
> However, when a non-compat task execs a compat binary or vice versa, the
> cntkctl state and task compat-ness get out of sync. Keep the hardware
> trapping state in sync with the task personality.
> 
> Fixes: d49f7d7376d0 ("arm64: Move handling of erratum 1418040 into C code")
> Signed-off-by: D Scott Phillips <scott at os.amperecomputing.com>
> Cc: <stable at vger.kernel.org> # 5.4.x
> ---
>  arch/arm64/include/asm/elf.h | 20 ++++++++++++++++++++
>  1 file changed, 20 insertions(+)
> 
> diff --git a/arch/arm64/include/asm/elf.h b/arch/arm64/include/asm/elf.h
> index 97932fbf973d..463ff8f00aa7 100644
> --- a/arch/arm64/include/asm/elf.h
> +++ b/arch/arm64/include/asm/elf.h
> @@ -160,8 +160,18 @@ typedef struct user_fpsimd_state elf_fpregset_t;
>  
>  #define SET_PERSONALITY(ex)						\
>  ({									\
> +	if (IS_ENABLED(CONFIG_ARM64_ERRATUM_1418040))			\
> +		preempt_disable();					\
>  	clear_thread_flag(TIF_32BIT);					\
>  	current->personality &= ~READ_IMPLIES_EXEC;			\
> +	if (this_cpu_has_cap(ARM64_WORKAROUND_1418040)) {		\
> +		u64 val;						\
> +		val = read_sysreg(cntkctl_el1);				\
> +		val |= ARCH_TIMER_USR_VCT_ACCESS_EN;			\
> +		write_sysreg(val, cntkctl_el1);				\
> +	}								\
> +	if (IS_ENABLED(CONFIG_ARM64_ERRATUM_1418040))			\
> +		preempt_enable();					\
>  })
>  
>  /* update AT_VECTOR_SIZE_ARCH if the number of NEW_AUX_ENT entries changes */
> @@ -223,7 +233,17 @@ int compat_elf_check_arch(const struct elf32_hdr *);
>   */
>  #define COMPAT_SET_PERSONALITY(ex)					\
>  ({									\
> +	if (IS_ENABLED(CONFIG_ARM64_ERRATUM_1418040))			\
> +		preempt_disable();					\
>  	set_thread_flag(TIF_32BIT);					\
> +	if (this_cpu_has_cap(ARM64_WORKAROUND_1418040)) {		\
> +		u64 val;						\
> +		val = read_sysreg(cntkctl_el1);				\
> +		val &= ~ARCH_TIMER_USR_VCT_ACCESS_EN;			\
> +		write_sysreg(val, cntkctl_el1);				\
> +	}								\
> +	if (IS_ENABLED(CONFIG_ARM64_ERRATUM_1418040))			\
> +		preempt_enable();					\
>   })
>  #ifdef CONFIG_COMPAT_VDSO
>  #define COMPAT_ARCH_DLINFO						\

I'd have opted for the use of sysreg_clear_set() instead of the
open-coded RMW, as it would elide the write-back in most cases.

This looks otherwise OK to me. With that fixed:

Reviewed-by: Marc Zyngier <maz at kernel.org>

	M.

-- 
Without deviation from the norm, progress is not possible.