https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90399
--- Comment #2 from Martin Karsten <mkarsten at uwaterloo dot ca> --- I single-stepped through the call 0x00007ffff7fc71f5 <+107>: callq 0x7ffff7fc7040 <__splitstack_makecontext@plt> and via indirect rip-relative addressing, it ends up at the (executable) __splitstack_makecontext() instance at 0x0000555555556379. That at least explains my initial confusion related to gdb output. It still leaves open the broader question about the two copies of split-stack code/variables not playing nicely with each other.
