[Bug middle-end/58585] [4.9 Regression] ICE in ipa with virtual inheritance

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58585

--- Comment #11 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Created attachment 31762
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31762&action=edit
Proposed fix

Hi,
this patch makes us to look harder for correct virtual table.  I walk the stack
of binfos with vtables and find first one with matching offset.  If I
understand thigns right, that should be always the binfo that holds the valid
vtable, since these are shared.
I will probably need to dive into C++ ABI to double check it.