On Wednesday 21 August 2013 12:30:10 Joseph S. Myers wrote:
> On Wed, 21 Aug 2013, Mike Frysinger wrote:
> > Another example of all the 64bit arches getting the definition via a
> > common file, but the 32bit ones all adding it by themselves and hppa
> > was missed.
> 
> How about adding an architecture-independent testcase (Linux-specific, of
> course) for this function?  I've no idea whether it can test any semantics
> of fanotify_init / fanotify_mark, or only that calls to them link OK, but
> in general when fixing bugs it's a good idea to add testcases that would
> have detected them, and both functions are in my list of untested symbols
> <http://sourceware.org/ml/libc-alpha/2013-07/msg00386.html>, which we need
> to add test coverage for bit by bit.

the fatrace project [1] was used as a spot check.  their is concern of running 
the test on a kernel where fanotify support has been disabled, but hopefully 
that should be easy to detect.
-mike

1: https://launchpad.net/fatrace