Re: [gpsd-dev] [gpsd-commit-watch] [SCM] GPSD branch, master, updated. release-3.11-311-g7de7d97

[Eric S. Raymond]

Gary E. Miller <address@hidden>:
> Yo Eric!
> 
> On Fri, 30 Jan 2015 16:41:27 -0500
> "Eric S. Raymond" <address@hidden> wrote:
> 
> > Gary E. Miller <address@hidden>:
> > > You removed the assert()s I just added.  The ones that start to show
> > > Hal's problem!
> > 
> > Oh, those were debug aids!  I apologize.  By all means put them back
> > in under some kind of debugging guard.
> 
> I changed it to this:
> 
> void gpsd_acquire_reporting_lock(void)
> {
>     int err;
>     /*@ -unrecog  (splint has no pthread declarations as yet) @*/
>     err = pthread_mutex_lock(&report_mutex);
>     /*@ +unrecog @*/
>     if ( 0 != err ) {
>         (void) fprintf(stderr,"pthread_mutex_lock() failed: %s\n",
>             strerror(errno));
>         exit(EXIT_FAILURE);
>     }
> }
> void gpsd_release_reporting_lock(void)
> {
>     int err;
>     /*@ -unrecog (splint has no pthread declarations as yet) @*/
>     err = pthread_mutex_unlock(&report_mutex);
>     /*@ +unrecog @*/
>     if ( 0 != err ) {
>         (void) fprintf(stderr,"pthread_mutex_unlock() failed: %s\n",
>             strerror(errno));
>         exit(EXIT_FAILURE);
>     }
> }
> 
> I'm guessing you would like some comments.  So how About this:
> 
> /* POSIX says pthread_mutex_lock() should only fail if the thread
> holding the lock has died.  Best to just die in that case.  */
> 
> /* POSIX says pthread_mutex_unlock() should only fail when trying to
> unlock a lock that does not exist, or is not owned by this thread.
> This should never happen, so best to die if that is the case.  */
> 
> Clear enough?

Yup.
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>