pthread cancellation points missing for fdatasync and lockf in some cases

[Ulrich Drepper]

Where to start summarizing what is wrong here:

- 'C' is not part of the interface description.  along with 'E' and 'V'
   it is a different beast

- the nptl has a separate ChangeLog

- the date in the ChangeLog entry has the wrong format

- lockf() is not meant to be a cancellation point except for F_LOCK
   (I filed a bug against the standard)

- years in copyright comments not changed in files

- you cannot have tested the change since otherwise you'd have seen
   that the __THROW for fdatasync in unistd.h had to be removed


After fixing things up I checked in the changes.

-- 
➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖