objdump -d with -j and objdump disassembly of .plt

[Maciej W. Rozycki]

On Fri, 30 Jan 2026, Alan Modra wrote:

> objdump -D disassembles non-code sections, paying no heed to symbols
> in those sections.  I think "objdump -d -j .data" should do the same,
> ie. not switch into dumping the section as data because an object
> symbol is encountered.

 I'm not convinced this is entirely uncontroversial, but in any case I 
think this asks for documentation in the `-d' option description in the 
manual.

 Also:

> diff --git a/ld/testsuite/ld-mips-elf/no-shared-1-n64.d b/ld/testsuite/ld-mips-elf/no-shared-1-n64.d
> index b5cd86e7285..f6cfc3b2930 100644
> --- a/ld/testsuite/ld-mips-elf/no-shared-1-n64.d
> +++ b/ld/testsuite/ld-mips-elf/no-shared-1-n64.d
> @@ -15,11 +15,13 @@ Disassembly of section \.text:
>  Disassembly of section \.data:
>  
>  0000000000060000 .*:
> -   60000:	00000000 	.*
> -   60004:	00068000 	.*
> +   60000:	00000000 .*
> +   60004:	00068000 .*
>  #...
>  Disassembly of section \.got:
>  
>  0000000000060010 <_GLOBAL_OFFSET_TABLE_>:
>  	\.\.\.
> -   60018:	80000000 00000000 00000000 00068000  .*
> +   60018:	80000000 .*
> +	\.\.\.
> +   60024:	00068000 .*

 This is a regression to the semantics of the test case: you've removed 
the matching of 32-bit halves of the two respective 64-bit GOT entries, so 
if LD starts writing gibberish there sometime (to `__gnu_local_gp' entry 
in particular), it will be missed.  Please bring the test case back to its 
original coverage.

  Maciej