public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
* patch for elf32-mips.c
@ 1999-09-25  6:14 Felix Lee
  0 siblings, 0 replies; 11+ messages in thread
From: Felix Lee @ 1999-09-25  6:14 UTC (permalink / raw)
  To: binutils

symptom:

    for some mips-elf targets, the linker sometimes
    spuriously complains:

	relocation truncated to fit: R_MIPS_LITERAL no symbol

analysis:

    mips_elf_calculate_relocation () sometimes neglects to
    set gp and gp0 before using them.

fix:

1999-09-25  Felix Lee  <flee@cygnus.com>

	* elf32-mips.c (mips_elf_calculate_relocation): Typo
	  and R_MIPS_LITERAL bug.

Index: elf32-mips.c
===================================================================
RCS file: /cvs/binutils/binutils/bfd/elf32-mips.c,v
retrieving revision 1.51
diff -c -p -r1.51 elf32-mips.c
*** elf32-mips.c	1999/08/27 00:08:49	1.51
--- elf32-mips.c	1999/09/25 12:53:02
*************** mips_elf_calculate_relocation (abfd, 
*** 5895,5901 ****
        *namep = bfd_elf_string_from_elf_section (input_bfd,
  						symtab_hdr->sh_link,
  						sym->st_name);
!       if (*namep == '\0')
  	*namep = bfd_section_name (input_bfd, sec);
  
        target_is_16_bit_code_p = (sym->st_other == STO_MIPS16);
--- 5895,5901 ----
        *namep = bfd_elf_string_from_elf_section (input_bfd,
  						symtab_hdr->sh_link,
  						sym->st_name);
!       if (**namep == '\0')
  	*namep = bfd_section_name (input_bfd, sec);
  
        target_is_16_bit_code_p = (sym->st_other == STO_MIPS16);
*************** mips_elf_calculate_relocation (abfd, 
*** 6070,6075 ****
--- 6070,6077 ----
        
      case R_MIPS_HI16:
      case R_MIPS_LO16:
+     case R_MIPS_LITERAL:
+     case R_MIPS16_GPREL:
      case R_MIPS_GPREL16:
      case R_MIPS_GPREL32:
        gp0 = _bfd_get_gp_value (input_bfd);

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Patch for elf32-mips.c
@ 2001-06-06 13:45 H . J . Lu
  0 siblings, 0 replies; 11+ messages in thread
From: H . J . Lu @ 2001-06-06 13:45 UTC (permalink / raw)
  To: binutils

I checked in the following patch.


H.J.
---
2001-06-06  H.J. Lu  <hjl@gnu.org>

	* elf32-mips.c (_bfd_mips_elf_object_p): Set the bad symtab
	for SGI only.

Index: elf32-mips.c
===================================================================
RCS file: /cvs/src/src/bfd/elf32-mips.c,v
retrieving revision 1.97
diff -u -p -r1.97 elf32-mips.c
--- elf32-mips.c	2001/05/23 17:26:35	1.97
+++ elf32-mips.c	2001/06/06 20:44:39
@@ -2331,7 +2331,8 @@ _bfd_mips_elf_object_p (abfd)
   /* Irix 5 and 6 is broken.  Object file symbol tables are not always
      sorted correctly such that local symbols precede global symbols,
      and the sh_info field in the symbol table is not always right.  */
-  elf_bad_symtab (abfd) = true;
+  if (SGI_COMPAT(abfd))
+    elf_bad_symtab (abfd) = true;
 
   bfd_default_set_arch_mach (abfd, bfd_arch_mips,
 			     elf_mips_mach (elf_elfheader (abfd)->e_flags));

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: patch for elf32-mips.c
       [not found] <199909251314.GAA16811@crane.prod.itd.earthlink.net>
@ 1999-09-25  9:40 ` Mark Mitchell
  0 siblings, 0 replies; 11+ messages in thread
From: Mark Mitchell @ 1999-09-25  9:40 UTC (permalink / raw)
  To: flee; +Cc: binutils

>>>>> "Felix" == Felix Lee <flee@teleport.com> writes:

    Felix> 1999-09-25 Felix Lee <flee@cygnus.com>

    Felix> 	* elf32-mips.c (mips_elf_calculate_relocation): Typo
    Felix> and R_MIPS_LITERAL bug.

Thanks for catching these.  The fixes look right to me, for whatever
that's worth.

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com

^ permalink raw reply	[flat|nested] 11+ messages in thread

* patch for elf32-mips.c
@ 1999-09-25  6:14 Felix Lee
  0 siblings, 0 replies; 11+ messages in thread
From: Felix Lee @ 1999-09-25  6:14 UTC (permalink / raw)
  To: binutils

symptom:

    for some mips-elf targets, the linker sometimes
    spuriously complains:

	relocation truncated to fit: R_MIPS_LITERAL no symbol

analysis:

    mips_elf_calculate_relocation () sometimes neglects to
    set gp and gp0 before using them.

fix:

1999-09-25  Felix Lee  <flee@cygnus.com>

	* elf32-mips.c (mips_elf_calculate_relocation): Typo
	  and R_MIPS_LITERAL bug.

Index: elf32-mips.c
===================================================================
RCS file: /cvs/binutils/binutils/bfd/elf32-mips.c,v
retrieving revision 1.51
diff -c -p -r1.51 elf32-mips.c
*** elf32-mips.c	1999/08/27 00:08:49	1.51
--- elf32-mips.c	1999/09/25 12:53:02
*************** mips_elf_calculate_relocation (abfd, 
*** 5895,5901 ****
        *namep = bfd_elf_string_from_elf_section (input_bfd,
  						symtab_hdr->sh_link,
  						sym->st_name);
!       if (*namep == '\0')
  	*namep = bfd_section_name (input_bfd, sec);
  
        target_is_16_bit_code_p = (sym->st_other == STO_MIPS16);
--- 5895,5901 ----
        *namep = bfd_elf_string_from_elf_section (input_bfd,
  						symtab_hdr->sh_link,
  						sym->st_name);
!       if (**namep == '\0')
  	*namep = bfd_section_name (input_bfd, sec);
  
        target_is_16_bit_code_p = (sym->st_other == STO_MIPS16);
*************** mips_elf_calculate_relocation (abfd, 
*** 6070,6075 ****
--- 6070,6077 ----
        
      case R_MIPS_HI16:
      case R_MIPS_LO16:
+     case R_MIPS_LITERAL:
+     case R_MIPS16_GPREL:
      case R_MIPS_GPREL16:
      case R_MIPS_GPREL32:
        gp0 = _bfd_get_gp_value (input_bfd);

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: PATCH for elf32-mips.c
  1999-08-07  6:55 ` Koundinya.K
@ 1999-08-07 14:42   ` Ralf Baechle
  0 siblings, 0 replies; 11+ messages in thread
From: Ralf Baechle @ 1999-08-07 14:42 UTC (permalink / raw)
  To: Koundinya.K; +Cc: Mark Mitchell, binutils, ian

On Sat, Aug 07, 1999 at 07:27:51PM +0530, Koundinya.K wrote:

> mark@codesourcery.com said:
> -> This patch repairs some bain dramage.  Ralf reports much improved
> -> results with this patch.  I'm working through some other issues, with
> -> help from Ian.
> 
> -> -- Mark Mitchell                   mark@codesourcery.com CodeSourcery,
> -> LLC               http://www.codesourcery.com 
> 
> This has also solved my earlier problems. In fact after more than one month 
> I am able to use the GNU ld again !!. I am excited about that.
> Thanx to Mark and Ian for looking into the problems that I reported.
> ( I have really bugged you people a lot :-)
> 
> I am also able to generate a shared library ( I have tested by building one 
> libkk.so). But If I use the crtbegin and crtend files I am not able to 
> generate the shared library. So I decided to do off with that.
> 
> Then When I try a test program to use that shared library I get a message 
> like:
> 
> /usr/cygnus/H-mips-dde-sysv4.2MP/mips-dde-sysv4.2MP/bin/ld: ./libkk.so: 
> invalid string offset 5920 >= 239 for section `.dynstr'
> 
> But then I am able to run the binary and it works fine. Amazing !!.

Let me append myself to this status report.  Current binutils plus a few
extra hacks are now able to compile working kernels and themselves again
on Linux/MIPS.  However when I using current binutils for building
current binutils with shared libbfd, then the resulting executable will
die due to a SIGBUS.  This is very strange since the lw which according
to gdb is causing the fault uses a valid memory reference.  I can show
this by using a a gdb print command to dereference it manually.

In short, I'm quite satisfied again with the state of binutils and wish
to say thanks as well.

  Ralf

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: PATCH for elf32-mips.c
  1999-08-05 19:46 PATCH " Mark Mitchell
@ 1999-08-07  6:55 ` Koundinya.K
  1999-08-07 14:42   ` Ralf Baechle
  0 siblings, 1 reply; 11+ messages in thread
From: Koundinya.K @ 1999-08-07  6:55 UTC (permalink / raw)
  To: Mark Mitchell; +Cc: binutils, ian

Hi all !!,
	
mark@codesourcery.com said:
-> This patch repairs some bain dramage.  Ralf reports much improved
-> results with this patch.  I'm working through some other issues, with
-> help from Ian.

-> -- Mark Mitchell                   mark@codesourcery.com CodeSourcery,
-> LLC               http://www.codesourcery.com 

This has also solved my earlier problems. In fact after more than one month 
I am able to use the GNU ld again !!. I am excited about that.
Thanx to Mark and Ian for looking into the problems that I reported.
( I have really bugged you people a lot :-)

I am also able to generate a shared library ( I have tested by building one 
libkk.so). But If I use the crtbegin and crtend files I am not able to 
generate the shared library. So I decided to do off with that.

Then When I try a test program to use that shared library I get a message 
like:

/usr/cygnus/H-mips-dde-sysv4.2MP/mips-dde-sysv4.2MP/bin/ld: ./libkk.so: 
invalid string offset 5920 >= 239 for section `.dynstr'

But then I am able to run the binary and it works fine. Amazing !!.

Cheers ,
../koundinya

^ permalink raw reply	[flat|nested] 11+ messages in thread

* PATCH for elf32-mips.c
@ 1999-08-05 19:46 Mark Mitchell
  1999-08-07  6:55 ` Koundinya.K
  0 siblings, 1 reply; 11+ messages in thread
From: Mark Mitchell @ 1999-08-05 19:46 UTC (permalink / raw)
  To: binutils; +Cc: Koundinya K.

This patch repairs some bain dramage.  Ralf reports much improved
results with this patch.  I'm working through some other issues, with
help from Ian.

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com

1999-08-04  Mark Mitchell  <mark@codesourcery.com>

	* elf32-mips.c (mips_elf_got16_entry): Don't multiply GOT index by
	the size of a GOT entry here.
	(mips_elf_calculate_relocation): Don't create a local GOT entry
	for the symbol in a GOT16 relocation; just for it's high-order bit
	(_bfd_mips_elf_relocate_section): Fix thinko.

Index: elf32-mips.c
===================================================================
RCS file: /cvs/binutils/binutils/bfd/elf32-mips.c,v
retrieving revision 1.44
diff -c -p -r1.44 elf32-mips.c
*** elf32-mips.c	1999/08/04 16:32:34	1.44
--- elf32-mips.c	1999/08/05 05:50:36
*************** mips_elf_got16_entry (abfd, info, value)
*** 5551,5557 ****
        if ((address & 0xffff0000) == value)
  	{
  	  /* This entry has the right high-order 16 bits.  */
! 	  index = MIPS_ELF_GOT_SIZE (abfd) * (entry - sgot->contents);
  	  break;
  	}
      }
--- 5551,5557 ----
        if ((address & 0xffff0000) == value)
  	{
  	  /* This entry has the right high-order 16 bits.  */
! 	  index = entry - sgot->contents;
  	  break;
  	}
      }
*************** mips_elf_calculate_relocation (abfd, 
*** 5971,5983 ****
      case R_MIPS_GOT_LO16:
      case R_MIPS_CALL_LO16:
        /* Find the index into the GOT where this value is located.  */
!       if (h)
  	{
  	  BFD_ASSERT (addend == 0);
  	  g = mips_elf_global_got_index 
  	    (elf_hash_table (info)->dynobj,
  	     (struct elf_link_hash_entry*) h);
  	}
        else
  	{
  	  g = mips_elf_local_got_index (abfd, info, symbol + addend);
--- 5971,5987 ----
      case R_MIPS_GOT_LO16:
      case R_MIPS_CALL_LO16:
        /* Find the index into the GOT where this value is located.  */
!       if (!local_p)
  	{
  	  BFD_ASSERT (addend == 0);
  	  g = mips_elf_global_got_index 
  	    (elf_hash_table (info)->dynobj,
  	     (struct elf_link_hash_entry*) h);
  	}
+       else if (r_type == R_MIPS_GOT16)
+ 	/* There's no need to create a local GOT entry here; the
+ 	   calculation for a local GOT16 entry does not involve G.  */
+ 	break;
        else
  	{
  	  g = mips_elf_local_got_index (abfd, info, symbol + addend);
*************** _bfd_mips_elf_relocate_section (output_b
*** 6599,6622 ****
  	    /* There's nothing to do for non-local relocations.  */
  	    continue;
  
- 	  r_symndx = ELF32_R_SYM (rel->r_info);
- 	  sym = local_syms + r_symndx;
- 	  if (ELF_ST_TYPE (sym->st_info) == STT_SECTION)
- 	    /* Adjust the addend appropriately.  */
- 	    addend += local_sections[r_symndx]->output_offset;
- 	  
  	  if (r_type == R_MIPS16_GPREL 
  	      || r_type == R_MIPS_GPREL16
  	      || r_type == R_MIPS_GPREL32)
  	    addend -= (_bfd_get_gp_value (output_bfd)
  		       - _bfd_get_gp_value (input_bfd));
! 	  else if (r_type == R_MIPS16_26 || r_type == R_MIPS16_26)
  	    /* The addend is stored without its two least
  	       significant bits (which are always zero.)  In a
  	       non-relocateable link, calculate_relocation will do
  	       this shift; here, we must do it ourselves.  */
  	    addend <<= 2;
  
  	  /* If the relocation is for a R_MIPS_HI16 or R_MIPS_GOT16,
  	     then we only want to write out the high-order 16 bits.
  	     The subsequent R_MIPS_LO16 will handle the low-order bits.  */
--- 6603,6626 ----
  	    /* There's nothing to do for non-local relocations.  */
  	    continue;
  
  	  if (r_type == R_MIPS16_GPREL 
  	      || r_type == R_MIPS_GPREL16
  	      || r_type == R_MIPS_GPREL32)
  	    addend -= (_bfd_get_gp_value (output_bfd)
  		       - _bfd_get_gp_value (input_bfd));
! 	  else if (r_type == R_MIPS_26 || r_type == R_MIPS16_26)
  	    /* The addend is stored without its two least
  	       significant bits (which are always zero.)  In a
  	       non-relocateable link, calculate_relocation will do
  	       this shift; here, we must do it ourselves.  */
  	    addend <<= 2;
  
+ 	  r_symndx = ELF32_R_SYM (rel->r_info);
+ 	  sym = local_syms + r_symndx;
+ 	  if (ELF_ST_TYPE (sym->st_info) == STT_SECTION)
+ 	    /* Adjust the addend appropriately.  */
+ 	    addend += local_sections[r_symndx]->output_offset;
+ 	  
  	  /* If the relocation is for a R_MIPS_HI16 or R_MIPS_GOT16,
  	     then we only want to write out the high-order 16 bits.
  	     The subsequent R_MIPS_LO16 will handle the low-order bits.  */

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: PATCH for elf32-mips.c
  1999-08-02 23:07 Mark Mitchell
@ 1999-08-03 14:57 ` Ralf Baechle
  0 siblings, 0 replies; 11+ messages in thread
From: Ralf Baechle @ 1999-08-03 14:57 UTC (permalink / raw)
  To: Mark Mitchell; +Cc: binutils, linux, linux-mips, linux-mips

On Mon, Aug 02, 1999 at 11:10:41PM -0700, Mark Mitchell wrote:

> These patches (checked in) should fix another couple of thinkos
> discovered by Ralf's mips-linux-gnu testing.  Ralf, just out of
> curiousity, do you think that Linux/GNU will move to the N32/64-bit
> ABIs at some point?  Or do you plan on sticking with the IRIX5-like
> ABI for the forseeable future?

My binutils work is part of Linux/MIPS64, at first for the Indy, then
for the Origin 200 and other 64-bit MIPS based systems.  It'll be
quite a while until we'll also have a 64-bit userland, but that's also
one of the aims.

  Ralf

^ permalink raw reply	[flat|nested] 11+ messages in thread

* PATCH for elf32-mips.c
@ 1999-08-02 23:07 Mark Mitchell
  1999-08-03 14:57 ` Ralf Baechle
  0 siblings, 1 reply; 11+ messages in thread
From: Mark Mitchell @ 1999-08-02 23:07 UTC (permalink / raw)
  To: binutils; +Cc: Ralf Baechle

These patches (checked in) should fix another couple of thinkos
discovered by Ralf's mips-linux-gnu testing.  Ralf, just out of
curiousity, do you think that Linux/GNU will move to the N32/64-bit
ABIs at some point?  Or do you plan on sticking with the IRIX5-like
ABI for the forseeable future?

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com

1999-08-02  Mark Mitchell  <mark@codesourcery.com>

	* elf32-mips.c (_bfd_mips_elf_relocate_section): Handle R_MIPS_26
	and R_MIPS16_26 relocations correctly when relocating.
	(_bfd_mips_elf_check_relocs): Don't assume that R_MIPS_CALL_HI16
	and R_MIPS_CALL_LO16 are for global symbols.
	
Index: elf32-mips.c
===================================================================
RCS file: /cvs/binutils/binutils/bfd/elf32-mips.c,v
retrieving revision 1.39
diff -c -p -r1.39 elf32-mips.c
*** elf32-mips.c	1999/08/01 22:45:06	1.39
--- elf32-mips.c	1999/08/03 06:00:31
*************** _bfd_mips_elf_relocate_section (output_b
*** 6552,6557 ****
--- 6552,6562 ----
  			    | ((addend & 0x7e00000) >> 16)
  			    | (addend & 0x1f));
  		}
+ 	      else if (r_type == R_MIPS16_26 
+ 		       || r_type == R_MIPS16_26)
+ 		/* The addend is stored without its two least
+ 		   significant bits (which are always zero.)  */
+ 		addend <<= 2;
  	    }
  	  else
  	    addend = rel->r_addend;
*************** _bfd_mips_elf_relocate_section (output_b
*** 6590,6595 ****
--- 6595,6605 ----
  	     The subsequent R_MIPS_LO16 will handle the low-order bits.  */
  	  if (r_type == R_MIPS_HI16 || r_type == R_MIPS_GOT16)
  	    addend >>= 16;
+ 	  /* If the relocation is for an R_MIPS_26 relocation, then
+ 	     the two low-order bits are not stored in the object file;
+ 	     they are implicitly zero.  */
+ 	  else if (r_type == R_MIPS_26 || r_type == R_MIPS16_26)
+ 	    addend >>= 2;
  
  	  if (rela_relocation_p)
  	    /* If this is a RELA relocation, just update the addend.
*************** _bfd_mips_elf_check_relocs (abfd, info, 
*** 7373,7388 ****
  
  	case R_MIPS_CALL_HI16:
  	case R_MIPS_CALL_LO16:
! 	  /* This symbol requires a global offset table entry.  */
! 	  if (!mips_elf_record_global_got_symbol (h, info, g))
! 	    return false;
! 
! 	  /* We need a stub, not a plt entry for the undefined
! 	     function.  But we record it as if it needs plt.  See
! 	     elf_adjust_dynamic_symbol in elflink.h.  */
! 	  h->elf_link_hash_flags |= ELF_LINK_HASH_NEEDS_PLT;
! 	  h->type = STT_FUNC;
  
  	  break;
  
  	case R_MIPS_GOT16:
--- 7383,7400 ----
  
  	case R_MIPS_CALL_HI16:
  	case R_MIPS_CALL_LO16:
! 	  if (h != NULL)
! 	    {
! 	      /* This symbol requires a global offset table entry.  */
! 	      if (!mips_elf_record_global_got_symbol (h, info, g))
! 		return false;
  
+ 	      /* We need a stub, not a plt entry for the undefined
+ 		 function.  But we record it as if it needs plt.  See
+ 		 elf_adjust_dynamic_symbol in elflink.h.  */
+ 	      h->elf_link_hash_flags |= ELF_LINK_HASH_NEEDS_PLT;
+ 	      h->type = STT_FUNC;
+ 	    }
  	  break;
  
  	case R_MIPS_GOT16:

^ permalink raw reply	[flat|nested] 11+ messages in thread

* PATCH for elf32-mips.c
@ 1999-08-01 15:48 Mark Mitchell
  0 siblings, 0 replies; 11+ messages in thread
From: Mark Mitchell @ 1999-08-01 15:48 UTC (permalink / raw)
  To: binutils

This simple patch fixes the case where no global GOT entries are
required.  We were tripping over a NULL pointer in that case.

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com

1999-08-01  Mark Mitchell  <mark@codesourcery.com>

	* elf32-mips.c (_bfd_mips_elf_final_link): Handle the case where
	there are no global symbols requiring GOT entries.
	(_bfd_mips_elf_size_dynamic_sections): Likewise.
	(_bfd_mips_elf_finish_dynamic_symbol): Likewise.
	(_bfd_mips_elf_finish_dynamic_sections): Likewise.

Index: elf32-mips.c
===================================================================
RCS file: /cvs/binutils/binutils/bfd/elf32-mips.c,v
retrieving revision 1.37
diff -c -p -r1.37 elf32-mips.c
*** elf32-mips.c	1999/08/01 20:58:19	1.37
--- elf32-mips.c	1999/08/01 22:44:26
*************** _bfd_mips_elf_final_link (abfd, info)
*** 4363,4371 ****
        got = bfd_get_section_by_name (dynobj, ".got");
        g = (struct mips_got_info *) elf_section_data (got)->tdata;
  
!       BFD_ASSERT ((elf_hash_table (info)->dynsymcount
! 		   - g->global_gotsym->dynindx)
! 		  <= g->global_gotno);
      }
  
    /* On IRIX5, we omit the .options section.  On IRIX6, however, we
--- 4363,4372 ----
        got = bfd_get_section_by_name (dynobj, ".got");
        g = (struct mips_got_info *) elf_section_data (got)->tdata;
  
!       if (g->global_gotsym != NULL)
! 	BFD_ASSERT ((elf_hash_table (info)->dynsymcount
! 		     - g->global_gotsym->dynindx)
! 		    <= g->global_gotno);
      }
  
    /* On IRIX5, we omit the .options section.  On IRIX6, however, we
*************** _bfd_mips_elf_size_dynamic_sections (out
*** 7773,7779 ****
    bfd *dynobj;
    asection *s;
    boolean reltext;
!   struct mips_got_info *g;
  
    dynobj = elf_hash_table (info)->dynobj;
    BFD_ASSERT (dynobj != NULL);
--- 7774,7780 ----
    bfd *dynobj;
    asection *s;
    boolean reltext;
!   struct mips_got_info *g = NULL;
  
    dynobj = elf_hash_table (info)->dynobj;
    BFD_ASSERT (dynobj != NULL);
*************** _bfd_mips_elf_size_dynamic_sections (out
*** 7896,7902 ****
   	  if (!mips_elf_sort_hash_table (info, 1))
   	    return false;
  
!  	  i = elf_hash_table (info)->dynsymcount - g->global_gotsym->dynindx;
  	  g->global_gotno = i;
  	  s->_raw_size += i * MIPS_ELF_GOT_SIZE (dynobj);
  	}
--- 7897,7908 ----
   	  if (!mips_elf_sort_hash_table (info, 1))
   	    return false;
  
! 	  if (g->global_gotsym != NULL)
! 	    i = elf_hash_table (info)->dynsymcount - g->global_gotsym->dynindx;
! 	  else
! 	    /* If there are no global symbols, or none requiring
! 	       relocations, then GLOBAL_GOTSYM will be NULL.  */
! 	    i = 0;
  	  g->global_gotno = i;
  	  s->_raw_size += i * MIPS_ELF_GOT_SIZE (dynobj);
  	}
*************** _bfd_mips_elf_size_dynamic_sections (out
*** 8038,8044 ****
        if (! MIPS_ELF_ADD_DYNAMIC_ENTRY (info, DT_MIPS_UNREFEXTNO, 0))
  	return false;
  
!       if (! MIPS_ELF_ADD_DYNAMIC_ENTRY (info, DT_MIPS_GOTSYM, 0))
  	return false;
  
        if (IRIX_COMPAT (dynobj) == ict_irix5
--- 8044,8051 ----
        if (! MIPS_ELF_ADD_DYNAMIC_ENTRY (info, DT_MIPS_UNREFEXTNO, 0))
  	return false;
  
!       if (g != NULL && g->global_gotsym != NULL
! 	  && ! MIPS_ELF_ADD_DYNAMIC_ENTRY (info, DT_MIPS_GOTSYM, 0))
  	return false;
  
        if (IRIX_COMPAT (dynobj) == ict_irix5
*************** _bfd_mips_elf_finish_dynamic_symbol (out
*** 8186,8192 ****
  
    /* Run through the global symbol table, creating GOT entries for all
       the symbols that need them.  */
!   if (h->dynindx >= g->global_gotsym->dynindx)
      {
        bfd_vma offset;
        bfd_vma value;
--- 8193,8200 ----
  
    /* Run through the global symbol table, creating GOT entries for all
       the symbols that need them.  */
!   if (g->global_gotsym != NULL
!       && h->dynindx >= g->global_gotsym->dynindx)
      {
        bfd_vma offset;
        bfd_vma value;

^ permalink raw reply	[flat|nested] 11+ messages in thread

* PATCH for elf32-mips.c
@ 1999-07-01  0:00 mark
  0 siblings, 0 replies; 11+ messages in thread
From: mark @ 1999-07-01  0:00 UTC (permalink / raw)
  To: binutils

This patch just allows us to recognize the N32 ABI when using `objdump
-p'.  OK to check in?

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com

1999-06-05  Mark Mitchell  <mark@codesourcery.com>

	* elf32-mips.c (_bfd_mips_elf_print_private_data): Recognize
	the N32 ABI.

Index: elf32-mips.c
===================================================================
RCS file: /cvs/binutils/binutils/bfd/elf32-mips.c,v
retrieving revision 1.3
diff -u -p -r1.3 elf32-mips.c
--- elf32-mips.c	1999/06/02 18:17:57	1.3
+++ elf32-mips.c	1999/06/06 00:43:25
@@ -2309,6 +2309,8 @@ _bfd_mips_elf_print_private_bfd_data (ab
     fprintf (file, _ (" [abi=EABI64]"));
   else if ((elf_elfheader (abfd)->e_flags & EF_MIPS_ABI))
     fprintf (file, _ (" [abi unknown]"));
+  else if ((elf_elfheader (abfd)->e_flags & EF_MIPS_ABI2))
+    fprintf (file, _ (" [abi=N32]"));
   else
     fprintf (file, _ (" [no abi set]"));
 

^ permalink raw reply	[flat|nested] 11+ messages in thread

end of thread, other threads:[~2001-06-06 13:45 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1999-09-25  6:14 patch for elf32-mips.c Felix Lee
  -- strict thread matches above, loose matches on Subject: below --
2001-06-06 13:45 Patch " H . J . Lu
     [not found] <199909251314.GAA16811@crane.prod.itd.earthlink.net>
1999-09-25  9:40 ` patch " Mark Mitchell
1999-09-25  6:14 Felix Lee
1999-08-05 19:46 PATCH " Mark Mitchell
1999-08-07  6:55 ` Koundinya.K
1999-08-07 14:42   ` Ralf Baechle
1999-08-02 23:07 Mark Mitchell
1999-08-03 14:57 ` Ralf Baechle
1999-08-01 15:48 Mark Mitchell
1999-07-01  0:00 mark

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).