[PATCH v1 1/6] LoongArch: Fix ld --no-relax bug

[WANG Xuerui]

On 11/16/23 14:23, mengqinggang wrote:
> When ld --no-relax, pcalau12i + ld.d still can be relaxed.

"When calling ld with --no-relax, ..."

Apart from the grammatical fix, is the "pcalau12i + ld.d pair still gets 
relaxed" behavior actually intended, or is it the broken behavior that's 
being fixed? Looking at the code it's not so evident so it's better to 
explain a bit.

> ---
>   bfd/elfnn-loongarch.c | 19 ++++++-------------
>   1 file changed, 6 insertions(+), 13 deletions(-)
>
> diff --git a/bfd/elfnn-loongarch.c b/bfd/elfnn-loongarch.c
> index 18ad3cc91ca..1162cb35cd6 100644
> --- a/bfd/elfnn-loongarch.c
> +++ b/bfd/elfnn-loongarch.c
> @@ -3996,29 +3996,22 @@ loongarch_elf_relax_section (bfd *abfd, asection *sec,
>   	    loongarch_relax_align (abfd, sec, sym_sec, info, rel, symval);
>   	  break;
>   	case R_LARCH_DELETE:
> -	  if (info->relax_pass == 1)
> +	  if (1 == info->relax_pass)
>   	    {
>   	      loongarch_relax_delete_bytes (abfd, sec, rel->r_offset, 4, info);
>   	      rel->r_info = ELFNN_R_INFO (0, R_LARCH_NONE);
>   	    }
>   	  break;
>   	case R_LARCH_PCALA_HI20:
> -	  if (info->relax_pass == 0)
> -	    {
> -	      if (i + 4 > sec->reloc_count)
> -		break;
> -	      loongarch_relax_pcala_addi (abfd, sec, rel, symval);
> -	    }
> +	  if (0 == info->relax_pass && (i + 4) <= sec->reloc_count)
> +	    loongarch_relax_pcala_addi (abfd, sec, rel, symval);
>   	  break;
>   	case R_LARCH_GOT_PC_HI20:
> -	  if (local_got)
> +	  if (local_got && 0 == info->relax_pass
> +	      && (i + 4) <= sec->reloc_count)
>   	    {
> -	      if (i + 4 > sec->reloc_count)
> -		break;
>   	      if (loongarch_relax_pcala_ld (abfd, sec, rel))
> -		{
> -		  loongarch_relax_pcala_addi (abfd, sec, rel, symval);
> -		}
> +		loongarch_relax_pcala_addi (abfd, sec, rel, symval);
>   	    }
>   	  break;
>   	default: