答复: [PATCH] x86/mm: Directly return instead of using local ret variable

[白浩文 <baihaowen@meizu.com>]

Thank you for pointing my mistaken.
________________________________________
发件人: Dave Hansen <dave.hansen@intel.com>
发送时间: 2022年3月28日 10:49:19
收件人: 白浩文; dave.hansen@linux.intel.com; luto@kernel.org; peterz@infradead.org
抄送: linux-kernel@vger.kernel.org
主题: Re: [PATCH] x86/mm: Directly return instead of using local ret variable

On 3/27/22 19:12, Haowen Bai wrote:
>
> diff --git a/arch/x86/mm/pf_in.c b/arch/x86/mm/pf_in.c
> index 3f83e31..4eff3ee 100644
> --- a/arch/x86/mm/pf_in.c
> +++ b/arch/x86/mm/pf_in.c
> @@ -130,7 +130,6 @@ enum reason_type get_ins_type(unsigned long ins_addr)
>       unsigned char *p;
>       struct prefix_bits prf;
>       int i;
> -     enum reason_type rv = OTHERS;
>
>       p = (unsigned char *)ins_addr;
>       p += skip_prefix(p, &prf);
> @@ -141,7 +140,7 @@ enum reason_type get_ins_type(unsigned long ins_addr)
>       CHECK_OP_TYPE(opcode, imm_wop, IMM_WRITE);
>
>  exit:
> -     return rv;
> +     return OTHERS;
>  }
>  #undef CHECK_OP_TYPE
>

Does this even compile?  Take a look at this macro:

> #define CHECK_OP_TYPE(opcode, array, type) \
>         for (i = 0; i < ARRAY_SIZE(array); i++) { \
>                 if (array[i] == opcode) { \
>                         rv = type; \
>                         goto exit; \
>                 } \

It uses 'rv'.