IFUNC resolver scheduling (bugs 21041, 20019)

[Maciej W. Rozycki]

On Thu, 26 Jan 2017, Carlos O'Donell wrote:

> My opinion is that IFUNC is a useful feature, but just needs better
> documentation of the restrictions and tooling to support developers.
> 
> By it's nature it is a very low-level feature, and we may indeed need to
> talk about individual machine restrictions.

 There are hidden traps we can't seem to escape ourselves too, e.g.:
<https://www.sourceware.org/ml/libc-alpha/2013-09/msg00781.html>.

  Maciej