Re: [PATCH v2] tst-ttyname: skip the test when /dev/ptmx is not available

["Dmitry V. Levin" <ldv at altlinux dot org>]

On Tue, Jan 02, 2018 at 12:46:29AM -0500, Luke Shumaker wrote:
> On Mon, 25 Dec 2017 16:41:54 -0500, Dmitry V. Levin wrote:
> > 
> > * sysdeps/unix/sysv/linux/tst-ttyname.c (do_in_chroot_1): Skip the
> > test instead of failing in case of ENOENT returned by posix_openpt.
> > ---
> >  ChangeLog                             | 5 +++++
> >  sysdeps/unix/sysv/linux/tst-ttyname.c | 9 ++++++++-
> >  2 files changed, 13 insertions(+), 1 deletion(-)
> > 
> > diff --git a/sysdeps/unix/sysv/linux/tst-ttyname.c b/sysdeps/unix/sysv/linux/tst-ttyname.c
> > index 0fdf1a8..6848a6d 100644
> > --- a/sysdeps/unix/sysv/linux/tst-ttyname.c
> > +++ b/sysdeps/unix/sysv/linux/tst-ttyname.c
> > @@ -253,7 +253,14 @@ do_in_chroot_1 (int (*cb)(const char *, int))
> >    /* Open the PTS that we'll be testing on.  */
> >    int master;
> >    char *slavename;
> > -  VERIFY ((master = posix_openpt (O_RDWR|O_NOCTTY|O_NONBLOCK)) >= 0);
> > +  master = posix_openpt (O_RDWR|O_NOCTTY|O_NONBLOCK);
> > +  if (master < 0)
> > +    {
> > +      if (errno == ENOENT)
> > +	FAIL_UNSUPPORTED ("posix_openpt: %m");
> > +      else
> > +	FAIL_EXIT1 ("posix_openpt: %m");
> > +    }
> >    VERIFY ((slavename = ptsname (master)));
> >    VERIFY (unlockpt (master) == 0);
> >    if (strncmp (slavename, "/dev/pts/", 9) != 0)
> > -- 
> > ldv
> 
> As I explained in a different thread: I generally support applying
> this and backporting it to the 2.26 release branch.
> 
> However, shouldn't it also make this same change in do_in_chroot_2?

If the first posix_openpt invocation succeeded, we may expect all
subsequent posix_openpt invocations will succeed, too.
If they don't, there must be something odd going on.


-- 
ldv

Attachment: signature.asc
Description: PGP signature