[Bug middle-end/65256] [5 regression] Undefined symbols linking f951

[dave.anglin at bell dot net]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65256

--- Comment #2 from dave.anglin at bell dot net ---
On 2015-03-02 2:36 PM, law at redhat dot com wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65256
>
> Jeffrey A. Law <law at redhat dot com> changed:
>
>             What    |Removed                     |Added
> ----------------------------------------------------------------------------
>                   CC|                            |law at redhat dot com
>
> --- Comment #1 from Jeffrey A. Law <law at redhat dot com> ---
> John, is this with SOM or ELF?  If the former, it looks like something might be
> missing its .import/.export statement.  What do the underlying symbols look
> like for the reference and the definition of those symbols?

It is with SOM.

I would say we have missing .import directives at least for first:

_ZTVN12_GLOBAL__N_18pass_dceE:
         .word   0
         .word   0
         .word   P%_ZN12_GLOBAL__N_18pass_dceD1Ev
         .word   P%_ZN12_GLOBAL__N_18pass_dceD0Ev
         .word   P%_ZN12_GLOBAL__N_18pass_dce5cloneEv
         .word   P%_ZN12_GLOBAL__N_18pass_dce4gateEP8function
         .word   P%_ZN12_GLOBAL__N_18pass_dce7executeEP8function

There is no local definition of 
_ZN12_GLOBAL__N_18pass_dce4gateEP8function and
no .import directive for it.

r220868 was ok and r220883 bad.

Dave