On Fri, Dec 04, 2009 at 08:57:18AM +0100, Claudio Jeker wrote:
> On Thu, Dec 03, 2009 at 11:50:52PM -0800, patrick keshishian wrote:
> > On Fri, Dec 04, 2009 at 09:34:31AM +0200, Vladimir Kirillov wrote:
> > > Hello t...@!
> > >
> > > Here's a diff for tun to make it keep record and display the
> > > pid of tunnel owner process in ifconfig.
> > > I'm sure it would be a handy feature, hope you'll like it.
> > [dd]
> > > @@ -561,6 +566,15 @@ tun_ioctl(struct ifnet *ifp, u_long cmd,
> > > case SIOCGIFMEDIA:
> > > case SIOCSIFMEDIA:
> > > error = ifmedia_ioctl(ifp, ifr, &tp->tun_media, cmd);
> > > + break;
> > > +
> > > + case SIOCGTUNPID:
> > > + if (tp->tun_proc == NULL) {
> > > + error = ENOTTY;
> > > + break;
> > > + }
> > > +
> > > + tunr->tun_pid = tp->tun_proc->p_pid;
> > > break;
> >
> >
> > Just a personal comment, why not:
> >
> > + case SIOCGTUNPID:
> > + if (tp->tun_proc == NULL)
> > + error = ENOTTY;
> > + else
> > + tunr->tun_pid = tp->tun_proc->p_pid;
> > break;
> >
> > Just so it is more consistent with the rest of the code?
> >
>
> The main problem with this code is that ENOTTY is the wrong errno to
> return. ENOTTY has a very specific meaning.
It's hard to pick the right one. ESRCH seems close, but is for the
reverse situation (you're requesting info for a pid taht does not exist).
How about ECHILD?
>
> --
> :wq Claudio
