Hi! Many thanks for paying attention to such issues!
On Fri, 25 Oct 2013 10:30:22 +0200, Justus Winter <4win...@informatik.uni-hamburg.de> wrote: > Previously any err value set before reaching lose: was overwritten and > thus not properly propagated. > > Found using the Clang Static Analyzer. > > * libports/create-internal.c (_ports_create_port_internal): Fix error > handling. > --- a/libports/create-internal.c > +++ b/libports/create-internal.c > @@ -109,10 +109,11 @@ _ports_create_port_internal (struct port_class *class, > err = EINTR; > lose: > pthread_mutex_unlock (&_ports_lock); > + error_t e; > lose_unlocked: > - err = mach_port_mod_refs (mach_task_self (), port, > - MACH_PORT_RIGHT_RECEIVE, -1); > - assert_perror (err); > + e = mach_port_mod_refs (mach_task_self (), port, > + MACH_PORT_RIGHT_RECEIVE, -1); > + assert_perror (e); > free (pi); > > return err; OK, but I suggest to move the definition of e after the lose_unlocked label, or even move all that (mach_port_mod_refs call and assert_perror) into a new scope to make its locality clear. May also want to replace the earlier »if (! pi)« handling with »goto lose_unlocked;« (»free (NULL)« being a no-op). Grüße, Thomas
pgpEa5CZ90Ha1.pgp
Description: PGP signature
