Re: [Patch] rexec check for sockaddr length

[Thorsten Kukuk <kukuk at suse dot de>]

On Sun, Apr 08, Ulrich Drepper wrote:

> Thorsten Kukuk <kukuk@suse.de> writes:
> 
> > --- inet/rexec.c	2000/10/28 00:28:37	1.17
> > +++ inet/rexec.c	2001/04/07 14:47:40
> > @@ -117,8 +117,7 @@
> >  		}
> >  		listen(s2, 1);
> >  		sa2len = sizeof (sa2);
> > -		if (getsockname(s2, (struct sockaddr *)&sa2, &sa2len) < 0 ||
> > -		    sa2len != SA_LEN((struct sockaddr *)&sa2)) {
> > +		if (getsockname(s2, (struct sockaddr *)&sa2, &sa2len) < 0) {
> >  			perror("getsockname");
> >  			(void) __close(s2);
> 
> The only problem which exists here is that errno is not set if "only"
> the size differs.  The size test itself is correct, it was there in
> some form or another forever.  I've checked in a patch for that.

That there was a check for ever means nothing. It is not necessary
(or getsockname will fail) and prevents glibc from working in cases
where it works without problems.

And I don't think that the new return value is correct. I don't see
where there should be a invalid argument ?

  Thorsten
-- 
Thorsten Kukuk       http://www.suse.de/~kukuk/       kukuk@suse.de
SuSE GmbH            Schanzaeckerstr. 10            90443 Nuernberg
Linux is like a Vorlon.  It is incredibly powerful, gives terse,
cryptic answers and has a lot of things going on in the background.