On Wed, Apr 23, 2014 at 10:30:23AM +0200, Pavel Reichl wrote:
> On Thu, 2014-04-17 at 20:01 +0200, Lukas Slebodnik wrote:
> > On (17/04/14 10:21), Lukas Slebodnik wrote:
> > >ehlo,
> > >
> > >The attached patch fixes my use case. (sssd in container without tty)
> > >I am not sure whether other processes use stdin, therefore I disable
monitorin
> > >of stdin only in main process.
> > >
> > >LS
> >
> > Problem can be reproduced with changing service file
> >
> > --- a/src/sysv/systemd/sssd.service.in
> > +++ b/src/sysv/systemd/sssd.service.in
> > @@ -5,11 +5,9 @@ After=syslog.target
> >
> > [Service]
> > EnvironmentFile=-@environment_file@
> > -ExecStart=@sbindir@/sssd -D -f
> > +ExecStart=@sbindir@/sssd -i -d3
> > # These two should be used with traditional UNIX forking daemons
> > # consult systemd.service(5) for more details
> > -Type=forking
> > -PIDFile=@localstatedir(a)/run/sssd.pid
> >
> > [Install]
> > WantedBy=multi-user.target
> > _______________________________________________
> > sssd-devel mailing list
> > sssd-devel(a)lists.fedorahosted.org
> >
https://lists.fedorahosted.org/mailman/listinfo/sssd-devel
>
> Hello Lukas,
>
> I was able to reproduce the issue and I confirm that your patch solves
> it, however I tested only on Fedora. I wonder if we should ask the
> reporter to test the patch on Debian/Ubuntu?
>
> Code looks good to me. I have just a little nitpick about
> using: bool need_stdin
>
> I think it's more readable to use something like:
> enum need_stdin {NEED_STDIN, DO_NOT_NEED_STDIN};
>
> But maybe it's just me.
>
> Thanks.
I was actually thinking we should close the stdin always. I can't think
of a reason to keep it open.
I tought it would be problem to monitor closed file descriptor with
tevent_add_fd, but it isn't.
New version attached.
LS