Hi Jose,
On 8/1/07, Jose Flavio Aguilar Paulino <[EMAIL PROTECTED]> wrote:
> Hi David,
>
> I have already changed the copyright headers
Great!
> I was already debugging the powerpc64 libunwind, and the work around I
> used to solve the function address does no work.
> I really need a function in elfxx to translate the function address.
> This happens because there are sections with this toc structure:
> typedef struct
> {
> unsigned long entry;
> unsigned long toc;
> unsigned long env;
> } func_descr_t;
>
> In power everytime I have a branch to a function the branch refers to
> this a toc table with entries like the above.
AFAICS, this is no different from ia64 (except that the function
descriptors have just two entries, normally: entry-point &
global-pointer). Why can't you read those entries from gdb like on
ia64? I'm afraid I'm not following.
> Maybe we should put a common file for the x86 archs that do not
> translate the function pointer, just return the same.
>From an unwind perspective, it doesn't (shouldn't) matter what a
symbol points to. If function pointers point to descriptors, that's
what libunwind should return and I still don't see why it should be
libunwind's job to interpret function descriptors.
--david
_______________________________________________
Libunwind-devel mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/libunwind-devel
