Re: lib/52386: pthread(3) doesn't respect PTHREAD_DESTRUCTOR_ITERATIONS

[Joerg Sonnenberger <joerg%bec.de@localhost>]

On Sun, Jul 09, 2017 at 11:30:13PM +0200, Kamil Rytarowski wrote:
> On 09.07.2017 22:20, Christos Zoulas wrote:
> > On Jul 9,  5:20pm, joerg%bec.de@localhost (Joerg Sonnenberger) wrote:
> > -- Subject: Re: lib/52386: pthread(3) doesn't respect PTHREAD_DESTRUCTOR_ITER
> > 
> > |  On Sun, Jul 09, 2017 at 03:30:00PM +0000, n54%gmx.com@localhost wrote:
> > |  > PTHREAD_DESTRUCTOR_ITERATIONS is not respected as the maximum number of destructor iterations.
> > |  
> > |  The test case is incorrect. PTHREAD_DESTRUCTOR_ITERATIONS specifies a
> > |  lower limit on the number of iterations performed. It is explicitly not
> > |  a hard limit. To quote POSIX:
> > |  
> > |  Calling pthread_setspecific() from a thread-specific data destructor
> > |  routine may result either in lost storage (after at least
> > |  PTHREAD_DESTRUCTOR_ITERATIONS attempts at destruction) or in an infinite
> > |  loop.
> > 
> > I agree, but making the code using the iteration constant we advertise
> > is simple enough.
> > 
> > christos
> > 
> 
> Before digging into POSIX, the attached test passes on FreeBSD, OpenBSD,
> Linux, Android, Darwin. Only NetBSD is distinct.

That doesn't make the test case correct though.

Joerg