[yocto] [EDT][prelink-cross] possible memory leak in opd_size

Mark Hatle mark.hatle at windriver.com
Wed Sep 9 15:21:30 PDT 2015


I'm finally getting caught up on these items.  I've looked at this, the function
in question is only used on arch-ia64.  I don't have any way to test arch-ia64,
nor do I think it's used much anymore (especially in the embedded space).  So
I'm going to be skipping this particular patch -- unless someone can convince me
it is the right fix -- or can verify behavior on an IA64 system.  (Note, ia64,
not x86-64)

--Mark

On 5/6/15 1:33 AM, Maninder Singh wrote:
> EP-F6AA0618C49C4AEDA73BFF1B39950BAB
> 
> Hi,
> 
> we found one issue through code-walk through
> and it seems we are creating unusable OPD table , Not sure about fix, please check below fix. 
> 
> Signed-off-by: Vaneet Narang <v.narang at samsung.com>
> Signed-off-by: Maninder Singh <maninder1.s at samsung.com>
> Reviewed-by: Ajeet Yadav <ajeet.y at samsung.com>
> ---
>  trunk/src/fptr.c |    1 +
>  1 files changed, 1 insertions(+), 0 deletions(-)
> 
> diff --git a/trunk/src/fptr.c b/trunk/src/fptr.c
> index cfe3aed..39271d5 100644
> --- a/trunk/src/fptr.c
> +++ b/trunk/src/fptr.c
> @@ -458,6 +458,7 @@ opd_size (struct prelink_info *info, GElf_Word entsize)
>     e->val = f->val;
>     e->gp = f->gp;
>     e->opd = ret | OPD_ENT_NEW;
> +   f->ent = e;
>     ret += entsize;
>        }
> 
> --
> 1.7.1
> 
> 
> Thanks and Regards,
> Maninder Singh
> 




More information about the yocto mailing list