candidate sched.h and sys/sched.h for review

[Corinna Vinschen]

On Mar 30 15:46, Ralf Corsepius wrote:
> On 03/30/2010 01:58 PM, Joel Sherrill wrote:
> >On 03/30/2010 03:30 AM, Corinna Vinschen wrote:
> >>#if defined(_POSIX_PRIORITY_SCHEDULING)
> >>>>/*
> >>>>* 13.3.1 Set Scheduling Parameters, P1003.1b-1993, p. 252
> >>This is just a formal thingy, but shouldn't new references to the
> >>standard better refer to the latest standard, which would be IEEE
> >>Std 1003.1-2008 at the present time?
> >>
> >I will fix the other issues and resubmit but I don't
> >have a paper copy of 2008. I agree it would be better.
> >Do you have a paper copy to get page numbers from
> >if I put XXX? Or should I delete them?
> 
> Why do you need a paper copy?
> 
> http://www.opengroup.org/onlinepubs/9699919799/
> has everything you need.
> 
> http://www.opengroup.org/ had an order form for paper copies, but I
> don't now if they sell 9699919799 in "dead wood".

Members of the Open Group also have free access to the document as PDF.
I don't think the pages of the hardcopy are overly important, just a
reference to POSIX-1.2008 and that the functions conform to that standard
is enough, IMHO.

Otherwise, the sched.h header is defined in XBD, chapter 13, the
functions in XSH, chapter 3.  There's no further numbering scheme.

  sched_setparam            p.1803
  sched_getparam            p.1800
  sched_setscheduler        p.1805
  sched_getscheduler        p.1801
  sched_get_priority_max    p.1799
  sched_get_priority_min    p.1799
  sched_rr_get_interval     p.1802
  sched_yield               p.1807

Btw., is there any problem to renumber SCHED_SPORADIC to 4?  If we
ever come around to implement that on Cygwin, we wouldn't have to
add another conditional.  If it is a problem, then, never mind.


Corinna

-- 
Corinna Vinschen
Cygwin Project Co-Leader
Red Hat