[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#2011: very silly messages from rsh

>>>>> "Ian" == Ian Jackson <ian@chiark.chu.cam.ac.uk> writes:

    Ian> Siggy Brentrup writes ("Re: Bug#2011: very silly messages from rsh"):
    Ian> ...
    Peter> As you already said, it's the getopt bug + another libc4 bug.
    >> Would you please enlighten me, I can't see how a getopt bug is involved here:

    Ian> See bugs 911 and 1685.  Basically, it's the following feature of GNU
    Ian> getopt, which was misconceived and should be disabled (from libc.info,
    Ian> node `Argument Syntax'):

Thanks, I was aware of that and consider it documented behaviour but I promise
to rethink when I get at those bugs.

As I indicated in my message, the bug you are experiencing here is rshd
violating the protocol while the rcmd function isn't prepared to handle it.

This is backed by rsh'ing to an AIX box:

beck:~/tmp$ rsh asterix -- uname -a
AIX asterix 2 3 002014257000
beck:~/tmp$ rsh asterix -l nobody -- uname -a
rshd: 0826-813 Permission is denied.
beck:~/tmp$ rsh asterix -l nonexistent -- uname -a
rshd: 0826-811 The remote user login is not correct.

As you can see, rshd's error messsages are correctly reported.

Before forwarding patches to H.J. Lu, I have to find out whether the socket
isn't correctly closed in rshd or it's the kernel incorrectly throwing away the
last packet after rshd exits.


mailto:bsb@uni-muenster.de           //      programmer/admin for hire
       bsb@beck.westfalen.de         \\  Opinions are strictly my own,
voice: +49-251-8619-99                \\      everything else is GPLed

Reply to: