public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
* [PATCH] ld: Fix ADDR32/64 incorrect outputs in pe-aarch64
@ 2023-01-18  5:11 Mark Harmstone
  2023-01-24 13:43 ` Nick Clifton
  0 siblings, 1 reply; 4+ messages in thread
From: Mark Harmstone @ 2023-01-18  5:11 UTC (permalink / raw)
  To: binutils, amodra; +Cc: Mark Harmstone

This changes coff_pe_aarch64_relocate_section to also perform
IMAGE_REL_ARM64_ADDR32 and IMAGE_REL_ARM64_ADDR64 relocations, which
fixes the test failures shown by 52936661.

---
 bfd/coff-aarch64.c | 45 +++++++++++++++++++++++++++++++++++++++++----
 1 file changed, 41 insertions(+), 4 deletions(-)

diff --git a/bfd/coff-aarch64.c b/bfd/coff-aarch64.c
index 7f6511c82cb..a0cc4f01ae6 100644
--- a/bfd/coff-aarch64.c
+++ b/bfd/coff-aarch64.c
@@ -526,10 +526,7 @@ coff_pe_aarch64_relocate_section (bfd *output_bfd,
       asection *sec = NULL;
       uint64_t dest_vma;
 
-      /* skip trivial relocations */
-      if (rel->r_type == IMAGE_REL_ARM64_ADDR32
-	  || rel->r_type == IMAGE_REL_ARM64_ADDR64
-	  || rel->r_type == IMAGE_REL_ARM64_ABSOLUTE)
+      if (rel->r_type == IMAGE_REL_ARM64_ABSOLUTE)
 	continue;
 
       symndx = rel->r_symndx;
@@ -564,6 +561,46 @@ coff_pe_aarch64_relocate_section (bfd *output_bfd,
 
       switch (rel->r_type)
 	{
+	case IMAGE_REL_ARM64_ADDR32:
+	  {
+	    uint64_t val;
+	    int32_t addend;
+
+	    addend = bfd_getl32 (contents + rel->r_vaddr);
+
+	    dest_vma += addend;
+
+	    val = dest_vma;
+
+	    if (val > 0xffffffff)
+	      (*info->callbacks->reloc_overflow)
+		(info, h ? &h->root : NULL, syms[symndx]._n._n_name,
+		"IMAGE_REL_ARM64_ADDR32", addend, input_bfd,
+		input_section, rel->r_vaddr - input_section->vma);
+
+	    bfd_putl32 (val, contents + rel->r_vaddr);
+	    rel->r_type = IMAGE_REL_ARM64_ABSOLUTE;
+
+	    break;
+	  }
+
+	case IMAGE_REL_ARM64_ADDR64:
+	  {
+	    uint64_t val;
+	    int64_t addend;
+
+	    addend = bfd_getl64 (contents + rel->r_vaddr);
+
+	    dest_vma += addend;
+
+	    val = dest_vma;
+
+	    bfd_putl64 (val, contents + rel->r_vaddr);
+	    rel->r_type = IMAGE_REL_ARM64_ABSOLUTE;
+
+	    break;
+	  }
+
 	case IMAGE_REL_ARM64_ADDR32NB:
 	  {
 	    uint64_t val;
-- 
2.38.2


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] ld: Fix ADDR32/64 incorrect outputs in pe-aarch64
  2023-01-18  5:11 [PATCH] ld: Fix ADDR32/64 incorrect outputs in pe-aarch64 Mark Harmstone
@ 2023-01-24 13:43 ` Nick Clifton
  2023-01-25  0:12   ` Mark Harmstone
  0 siblings, 1 reply; 4+ messages in thread
From: Nick Clifton @ 2023-01-24 13:43 UTC (permalink / raw)
  To: Mark Harmstone, binutils, amodra

Hi Mark,

> This changes coff_pe_aarch64_relocate_section to also perform
> IMAGE_REL_ARM64_ADDR32 and IMAGE_REL_ARM64_ADDR64 relocations, which
> fixes the test failures shown by 52936661.

Approved - please apply.

Cheers
   Nick


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] ld: Fix ADDR32/64 incorrect outputs in pe-aarch64
  2023-01-24 13:43 ` Nick Clifton
@ 2023-01-25  0:12   ` Mark Harmstone
  2023-01-25  2:29     ` Alan Modra
  0 siblings, 1 reply; 4+ messages in thread
From: Mark Harmstone @ 2023-01-25  0:12 UTC (permalink / raw)
  To: Nick Clifton, binutils, amodra

Thanks Nick, but I think the changes that Alan made have superseded this - I no longer get the test failures. Alan, please correct me if I'm wrong.

Mark

On 24/1/23 13:43, Nick Clifton wrote:
> Hi Mark,
>
>> This changes coff_pe_aarch64_relocate_section to also perform
>> IMAGE_REL_ARM64_ADDR32 and IMAGE_REL_ARM64_ADDR64 relocations, which
>> fixes the test failures shown by 52936661.
>
> Approved - please apply.
>
> Cheers
>   Nick
>


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] ld: Fix ADDR32/64 incorrect outputs in pe-aarch64
  2023-01-25  0:12   ` Mark Harmstone
@ 2023-01-25  2:29     ` Alan Modra
  0 siblings, 0 replies; 4+ messages in thread
From: Alan Modra @ 2023-01-25  2:29 UTC (permalink / raw)
  To: Mark Harmstone; +Cc: Nick Clifton, binutils

On Wed, Jan 25, 2023 at 12:12:07AM +0000, Mark Harmstone wrote:
> Thanks Nick, but I think the changes that Alan made have superseded this

That's right.  It would be a backward step to install this patch.

-- 
Alan Modra
Australia Development Lab, IBM

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2023-01-25  2:29 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-01-18  5:11 [PATCH] ld: Fix ADDR32/64 incorrect outputs in pe-aarch64 Mark Harmstone
2023-01-24 13:43 ` Nick Clifton
2023-01-25  0:12   ` Mark Harmstone
2023-01-25  2:29     ` Alan Modra

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).