This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: Fix ld/13803: ARM test-cases for "erratum 760522 fix" failing for arm-linux-gnueabi
- From: Matthew Gretton-Dann <matthew dot gretton-dann at arm dot com>
- To: Hans-Peter Nilsson <hans-peter dot nilsson at axis dot com>
- Cc: "binutils at sourceware dot org" <binutils at sourceware dot org>, Richard Earnshaw <rearnsha at arm dot com>
- Date: Mon, 5 Mar 2012 09:44:44 +0000
- Subject: Re: Fix ld/13803: ARM test-cases for "erratum 760522 fix" failing for arm-linux-gnueabi
- References: <201203050605.q2565oGp022944@ignucius.se.axis.com>
On Mon, Mar 05, 2012 at 06:05:50AM +0000, Hans-Peter Nilsson wrote:
> Seems like part of the objdump result template wasn't
> sufficiently generalized: for arm-eabi the function
> func_to_branch_to is resolved to 0x00008000 while for
> arm-linux-gnueabi it's resolved to 0x00008074. Suggested fix as
> follows. Not committed as obvious; borderline, maybe the
> difference shouldn't really be there. But then why regexpifying
> everything else about the address 0x8000?
>
> Tested arm-eabi and arm-linux-gnueabi.
>
> Ok?
>
> ld/testsuite:
> * ld-arm/fix-arm1176-off.d, ld-arm/fix-arm1176-on.d: Regexpify
> address of func_to_branch_to.
>
> Index: ld-arm/fix-arm1176-off.d
> ===================================================================
> RCS file: /cvs/src/src/ld/testsuite/ld-arm/fix-arm1176-off.d,v
> retrieving revision 1.1
> diff -p -u -r1.1 fix-arm1176-off.d
> --- ld-arm/fix-arm1176-off.d 9 Aug 2011 13:10:44 -0000 1.1
> +++ ld-arm/fix-arm1176-off.d 5 Mar 2012 06:03:09 -0000
> @@ -9,7 +9,7 @@ Disassembly of section .foo:
>
> [0-9a-f]+ <__func_to_branch_to_veneer>:
> [0-9a-f]+: e51ff004 ldr pc, \[pc, #-4\] ; 200101c <__func_to_branch_to_veneer\+0x4>
> - [0-9a-f]+: 00008000 .word 0x00008000
> + [0-9a-f]+: 00008... .word 0x00008...
>
> Disassembly of section .text:
>
> Index: ld-arm/fix-arm1176-on.d
> ===================================================================
> RCS file: /cvs/src/src/ld/testsuite/ld-arm/fix-arm1176-on.d,v
> retrieving revision 1.1
> diff -p -u -r1.1 fix-arm1176-on.d
> --- ld-arm/fix-arm1176-on.d 9 Aug 2011 13:10:44 -0000 1.1
> +++ ld-arm/fix-arm1176-on.d 5 Mar 2012 06:03:09 -0000
> @@ -11,7 +11,7 @@ Disassembly of section .foo:
> [0-9a-f]+: 4778 bx pc
> [0-9a-f]+: 46c0 nop ; \(mov r8, r8\)
> [0-9a-f]+: e51ff004 ldr pc, \[pc, #-4\] ; 2001020 <__func_to_branch_to_veneer\+0x8>
> - [0-9a-f]+: 00008000 .word 0x00008000
> + [0-9a-f]+: 00008... .word 0x00008...
> [0-9a-f]+: 00000000 .word 0x00000000
>
> Disassembly of section .text:
>
> brgds, H-P
>
I'm not a maintainer so can't approve the patch.
But I did write the original patch - and this does look fine to me.
Thanks,
Matt
--
Matthew Gretton-Dann
Principal Engineer, PD Software, ARM Ltd.