Re: COMMITTED: Add support for STT_IFUNC

[Nick Clifton <nickc at redhat dot com>]

Hi Andreas,

> > + @item STT_IFUNC
> > + Mark the symbol as an indirect function when evaluated during reloc
> > + processing.  (This is only supported on Linux targeted assemblers).

> This sentence is confusing.  What reloc processing is meant here?
> Literally taken, this sentence says that the symbol is not marked right
> now, but at some later point in time, and that it involves evaluating
> the symbol.

I see.  How about this rewording then:

  "Mark the symbol as an indirect function.  This has the same
   semantics as STT_FUNC except when the symbol is referenced by
   a relocation.  In this case the symbol's value is not used in
   the relocation but instead the symbol is invoked as a function
   taking no arguments and the return value is used in the
   relocation.  This also means that processing of the relocation
   must be delayed until run-time".

Cheers
  Nick

PS. I am about to go off on vacation for a couple of weeks, so if I do 
not reply to emails on this subject straight away, that is why.