linking to libc and libc_r (was Re: Certain 4.10 apps failing to run as root)

From: Anthony Chavez (acc_at_anthonychavez.org)
Date: 07/14/04

  • Next message: Jason Dusek: "FAILURE - ATAPI_RESET no interrupt" 
    To: freebsd-questions@freebsd.org
Date: Tue, 13 Jul 2004 21:00:01 -0600

    -----BEGIN PGP SIGNED MESSAGE-----
    Hash: SHA1

    Here's an update...

    On Thu, 10 Jun 2004 09:12:33 -0600 Anthony Chavez <acc@anthonychavez.org> wrote:
    > I think that Kris has been swallowed up by other efforts at the moment,
    > so I'm moving this to the freebsd-stable list, where it's probably more
    > relevant anyhow.

    Actually, this particular machine is actually running RELENG_4_10. My
    bad. So I'm posting this back to -questions. :-)

    > On Thu, 03 Jun 2004 01:35:33 -0600 Anthony Chavez <acc@anthonychavez.org> wrote:
    >> I just upgraded one of my machines to 4.10-STABLE from 4.9-STABLE and

    s/0-STABLE/0-RELEASE/

    >> certain binaries (such as vim or vipw) when run as root (either with
    >> sudo or su -) freeze up with "rcmdsh: unknown user:" followed by a bunch
    >> of garbage.
    >>
    >> Any thoughts on what could be causing this?
    >
    > On Thu, 03 Jun 2004 07:22:09 -0600 Anthony Chavez <acc@anthonychavez.org> wrote:
    >> On Thu, 3 Jun 2004 01:27:57 -0700 Kris Kennaway <kris@obsecurity.org> wrote:
    >>> I've only seen this when something incorrectly links to both libc and
    >>> libc_r. Neither of those four binaries should be linked to libc_r at
    >>> all though. Can you confirm with ldd(1)?

    I have successfully worked around the problem by passing -DLITE to make
    when building editors/vim. When I remove -DLITE, the error resurfaces.

    The following binaries remain affected (there are others, but they are
    irrelevant to vim):

    /usr/X11R6/bin/glxinfo:
            libc_r.so.4 => /usr/lib/libc_r.so.4 (0x28221000)
            libc.so.4 => /usr/lib/libc.so.4 (0x2831e000)
    /usr/X11R6/bin/glxgears:
            libc_r.so.4 => /usr/lib/libc_r.so.4 (0x281ac000)
            libc.so.4 => /usr/lib/libc.so.4 (0x28264000)

    What is causing this? My first suspicion is ccache. Could it be that I
    have set the following variables in /etc/make.conf?

    CC=/usr/local/bin/cc
    CPP=/usr/local/bin/cpp
    CXX=/usr/local/bin/c++

    ( /usr/local/bin/c{c,pp,++} are symlinked to /usr/local/bin/ccache. )

    Apart from that, the only difference between the system in question and
    a stock FreeBSD installation is the following MARK_ARGS pair in
    /usr/local/etc/pkgtools.conf:

    'editors/vim' => '-DLITE -DWITH_CSCOPE -DWITH_PERL -DWITH_PYTHON -DWITH_RUBY -DWITH_TCL'

    - --
    Anthony Chavez http://www.anthonychavez.org/
    mailto:acc@anthonychavez.org jabber:acc@anthonychavez.org
    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.2.4 (Darwin)

    iD8DBQFA9KG1bZTbIaRBRXERAvYRAJ9Q+3AcEYJLB1iYSoSqWZM/ZMfxkACfXKeT
    WTypK5pincbgv/4DyQ0abDU=
    =FEw9
    -----END PGP SIGNATURE-----

    _______________________________________________
    freebsd-questions@freebsd.org mailing list
    http://lists.freebsd.org/mailman/listinfo/freebsd-questions
    To unsubscribe, send any mail to "freebsd-questions-unsubscribe@freebsd.org"

  • Next message: Jason Dusek: "FAILURE - ATAPI_RESET no interrupt"