Re: [PATCH v3 4/4] arm64: mte: Optimize mte_assign_mem_tag_range()

[Mark Rutland]

On Fri, Jan 15, 2021 at 12:00:43PM +0000, Vincenzo Frascino wrote:
> mte_assign_mem_tag_range() is called on production KASAN HW hot
> paths. It makes sense to optimize it in an attempt to reduce the
> overhead.
> 
> Optimize mte_assign_mem_tag_range() based on the indications provided at
> [1].

... what exactly is the optimization?

I /think/ you're just trying to have it inlined, but you should mention
that explicitly.

> 
> [1] https://lore.kernel.org/r/CAAeHK+wCO+J7D1_T89DG+jJrPLk3X9RsGFKxJGd0ZcUFjQT-9Q@xxxxxxxxxxxxxx/
> 
> Cc: Catalin Marinas <catalin.marinas@xxxxxxx>
> Cc: Will Deacon <will@xxxxxxxxxx>
> Signed-off-by: Vincenzo Frascino <vincenzo.frascino@xxxxxxx>
> ---
>  arch/arm64/include/asm/mte.h | 26 +++++++++++++++++++++++++-
>  arch/arm64/lib/mte.S         | 15 ---------------
>  2 files changed, 25 insertions(+), 16 deletions(-)
> 
> diff --git a/arch/arm64/include/asm/mte.h b/arch/arm64/include/asm/mte.h
> index 1a715963d909..9730f2b07b79 100644
> --- a/arch/arm64/include/asm/mte.h
> +++ b/arch/arm64/include/asm/mte.h
> @@ -49,7 +49,31 @@ long get_mte_ctrl(struct task_struct *task);
>  int mte_ptrace_copy_tags(struct task_struct *child, long request,
>  			 unsigned long addr, unsigned long data);
>  
> -void mte_assign_mem_tag_range(void *addr, size_t size);
> +static inline void mte_assign_mem_tag_range(void *addr, size_t size)
> +{
> +	u64 _addr = (u64)addr;
> +	u64 _end = _addr + size;
> +
> +	/*
> +	 * This function must be invoked from an MTE enabled context.
> +	 *
> +	 * Note: The address must be non-NULL and MTE_GRANULE_SIZE aligned and
> +	 * size must be non-zero and MTE_GRANULE_SIZE aligned.
> +	 */
> +	do {
> +		/*
> +		 * 'asm volatile' is required to prevent the compiler to move
> +		 * the statement outside of the loop.
> +		 */
> +		asm volatile(__MTE_PREAMBLE "stg %0, [%0]"
> +			     :
> +			     : "r" (_addr)
> +			     : "memory");
> +
> +		_addr += MTE_GRANULE_SIZE;
> +	} while (_addr < _end);

Is there any chance that this can be used for the last bytes of the
virtual address space? This might need to change to `_addr == _end` if
that is possible, otherwise it'll terminate early in that case.

> +}

What does the code generation look like for this, relative to the
assembly version?

Thanks,
Mark.

> +
>  
>  #else /* CONFIG_ARM64_MTE */
>  
> diff --git a/arch/arm64/lib/mte.S b/arch/arm64/lib/mte.S
> index 9e1a12e10053..a0a650451510 100644
> --- a/arch/arm64/lib/mte.S
> +++ b/arch/arm64/lib/mte.S
> @@ -150,18 +150,3 @@ SYM_FUNC_START(mte_restore_page_tags)
>  	ret
>  SYM_FUNC_END(mte_restore_page_tags)
>  
> -/*
> - * Assign allocation tags for a region of memory based on the pointer tag
> - *   x0 - source pointer
> - *   x1 - size
> - *
> - * Note: The address must be non-NULL and MTE_GRANULE_SIZE aligned and
> - * size must be non-zero and MTE_GRANULE_SIZE aligned.
> - */
> -SYM_FUNC_START(mte_assign_mem_tag_range)
> -1:	stg	x0, [x0]
> -	add	x0, x0, #MTE_GRANULE_SIZE
> -	subs	x1, x1, #MTE_GRANULE_SIZE
> -	b.gt	1b
> -	ret
> -SYM_FUNC_END(mte_assign_mem_tag_range)
> -- 
> 2.30.0
>