[PATCH][ARM] Add missing initialization of stub hash entry fields.

Christophe LYON christophe.lyon@st.com
Fri Jul 10 12:02:00 GMT 2009


It looks obvious, indeed. (don't consider this as an approval, for I am 
not entitled to do so ;-)

Maybe a valgrind pass would help catching other similar missing 
initializations?

Christophe.



On 08.07.2009 20:43, Doug Kwan (???) wrote:
> Hi,
> 
>     This patch initializes the target_addend and the orig_insn fields
> of the newly created entry in stub_hash_entry.  Previously,  some
> stubs pointed to random locations because target_addend was not
> initialized.  I ran the binutils testsuites for target arm-none-eabi
> on host x86_64-unknown-linux-gnu.  The patch did not introduce any new
> regressions.
> 
> OK to check in?
> 
> -Doug
> 
> 2009-07-08  Doug Kwan  <dougkwan@google.com>
> 
> 	* elf32-arm.c (stub_hash_newfunc): Initialize fields
> 	target_addend and orig_insn.
> 
> ----------------------------------------------------
> Index: bfd/elf32-arm.c
> ===================================================================
> RCS file: /cvs/src/src/bfd/elf32-arm.c,v
> retrieving revision 1.203
> diff -u -p -r1.203 elf32-arm.c
> --- bfd/elf32-arm.c	1 Jul 2009 14:48:26 -0000	1.203
> +++ bfd/elf32-arm.c	8 Jul 2009 18:33:59 -0000
> @@ -2716,6 +2716,8 @@ stub_hash_newfunc (struct bfd_hash_entry
>        eh->stub_offset = 0;
>        eh->target_value = 0;
>        eh->target_section = NULL;
> +      eh->target_addend = 0;
> +      eh->orig_insn = 0;
>        eh->stub_type = arm_stub_none;
>        eh->stub_size = 0;
>        eh->stub_template = NULL;
> 



More information about the Binutils mailing list