------- Comment From <email address hidden> 2017-01-29 22:25 EDT-------
I think both e4aa8a9 ("PowerPC64 ELFv1 undefined weak functions") and 7f27ccfc ("PR 20472, PowerPC64 ifunc confusion") should be left off 2.26 based binutils.
While the undefined weak fix is desirable, it broke ifunc and the followup fix for ifunc dynamic relocation accounting confusion requires other infrastructure changes that are not on 2.26. The backports necessary are large enough that I think it would be better to simply require 2.27 if we really want undefined weak to be fixed.
------- Comment From <email address hidden> 2017-01-29 22:25 EDT-------
I think both e4aa8a9 ("PowerPC64 ELFv1 undefined weak functions") and 7f27ccfc ("PR 20472, PowerPC64 ifunc confusion") should be left off 2.26 based binutils.
While the undefined weak fix is desirable, it broke ifunc and the followup fix for ifunc dynamic relocation accounting confusion requires other infrastructure changes that are not on 2.26. The backports necessary are large enough that I think it would be better to simply require 2.27 if we really want undefined weak to be fixed.