[Patch] Keep bfd section lma in sync with elf data structures

[Patch] Keep bfd section lma in sync with elf data structures

[Sterling Augustine]

[-- Attachment #1: Type: text/plain, Size: 439 bytes --]

I have committed the attached patch which updates the lma of a section 
inside the bfd data structure when it is changed inside the elf 
structure. The updated value is now printed in the map file, rather than 
the formerly stale value. This fixes bugzilla #11219.

Sterling

2010-02-05  Sterling Augustine  <sterling@tensilica.com>

         * elf.c (assign_file_positions_for_load_sections) Update lma of
         section if necessary.


[-- Attachment #2: patch --]
[-- Type: text/plain, Size: 1059 bytes --]

Index: elf.c
===================================================================
RCS file: /cvs/src/src/bfd/elf.c,v
retrieving revision 1.500
diff -u -p -r1.500 elf.c
--- elf.c	4 Feb 2010 09:16:38 -0000	1.500
+++ elf.c	5 Feb 2010 17:02:01 -0000
@@ -4462,6 +4462,22 @@ assign_file_positions_for_load_sections 
 		}
 	      p->p_memsz += adjust;
 
+	      if (p->p_paddr + p->p_memsz != sec->lma)
+		{
+		  /* This behavior is a compromise--ld has long
+		     silently changed the lma of sections when
+		     lma - vma is not equal for every section in a
+		     pheader--but only in the internal elf structures.
+		     Silently changing the lma is probably a bug, but
+		     changing it would have subtle and unknown
+		     consequences for existing scripts.
+
+		     Instead modify the bfd data structure to reflect
+		     what happened.  This at least fixes the values
+		     for the lma in the mapfile.  */
+		  sec->lma = p->p_paddr + p->p_memsz;
+		}
+
 	      if (this_hdr->sh_type != SHT_NOBITS)
 		{
 		  if (p->p_filesz + adjust < p->p_memsz)

Re: [Patch] Keep bfd section lma in sync with elf data structures

[H.J. Lu]

Can put PR 11219 in ChangeLog entry?

Thanks.


H.J.

On Fri, Feb 5, 2010 at 9:13 AM, Sterling Augustine
<sterling@tensilica.com> wrote:
> I have committed the attached patch which updates the lma of a section
> inside the bfd data structure when it is changed inside the elf structure.
> The updated value is now printed in the map file, rather than the formerly
> stale value. This fixes bugzilla #11219.
>
> Sterling
>
> 2010-02-05  Sterling Augustine  <sterling@tensilica.com>
>
>        * elf.c (assign_file_positions_for_load_sections) Update lma of
>        section if necessary.
>
>
> Index: elf.c
> ===================================================================
> RCS file: /cvs/src/src/bfd/elf.c,v
> retrieving revision 1.500
> diff -u -p -r1.500 elf.c
> --- elf.c       4 Feb 2010 09:16:38 -0000       1.500
> +++ elf.c       5 Feb 2010 17:02:01 -0000
> @@ -4462,6 +4462,22 @@ assign_file_positions_for_load_sections
>                }
>              p->p_memsz += adjust;
>
> +             if (p->p_paddr + p->p_memsz != sec->lma)
> +               {
> +                 /* This behavior is a compromise--ld has long
> +                    silently changed the lma of sections when
> +                    lma - vma is not equal for every section in a
> +                    pheader--but only in the internal elf structures.
> +                    Silently changing the lma is probably a bug, but
> +                    changing it would have subtle and unknown
> +                    consequences for existing scripts.
> +
> +                    Instead modify the bfd data structure to reflect
> +                    what happened.  This at least fixes the values
> +                    for the lma in the mapfile.  */
> +                 sec->lma = p->p_paddr + p->p_memsz;
> +               }
> +
>              if (this_hdr->sh_type != SHT_NOBITS)
>                {
>                  if (p->p_filesz + adjust < p->p_memsz)
>
>



-- 
H.J.