[Bug gold/15200] Runtime undefined reference to __exidx_start/_end

[dougkwan at google dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=15200

--- Comment #14 from Doug Kwan <dougkwan at google dot com> 2013-03-20 04:41:03 
UTC ---
The special symbols in question __exidx_{start,end} were exported from
DSO's by mistake in both ld & gold and the mistake has been fixed in
both linkers.  We really don't want references of such symbols in the
main executable to be bound to the definitions in a DSO.  So Peter
wants to define those symbols even when they are seen in a DSO.  Doing
so makes it possible to use DSO's generated by older versions of gold
& ld on ARM but I really don't know whether this affects correctness
in general.

-Doug

On Tue, Mar 19, 2013 at 8:29 PM, ian at airs dot com
<address@hidden> wrote:
> http://sourceware.org/bugzilla/show_bug.cgi?id=15200
>
> --- Comment #12 from Ian Lance Taylor <ian at airs dot com> 2013-03-20 
> 03:29:49 UTC ---
> oldsym->in_dyn() will return true if the symbol was seen in a dynamic object.
> I don't see why we should create the symbol if it is seen in a dynamic object.
>
> It seems that the code should be something like
>
>     if (oldsym == NULL)
>       return NULL;
>     if (oldsym->is_undefined())
>       ;
>     else if (oldsym->is_from_dynobj())
>       ;
>     else
>       return NULL;
>
> but that's not right either.  We should only create the symbol if it is
> referenced by a regular object.  So it really needs to be
>
>     if (oldsym == NULL)
>       return NULL;
>
> --
> Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
> ------- You are receiving this mail because: -------
> You are on the CC list for the bug.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.