bug-inetutils
[Top][All Lists]
Advanced

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

[bug-inetutils] [PATCH] Inetd and test scripts


From: Mats Erik Andersson
Subject: [bug-inetutils] [PATCH] Inetd and test scripts
Date: Sun, 31 Oct 2010 04:52:15 +0100
User-agent: Mutt/1.5.18 (2008-05-17)

Hello again,

this time I need to improve the test scripts and I return to
the ever recurring theme of exact sockaddr lengths.

The scripts for testing pings and tracing are given some small
updates. The first now tests `ping` and `ping6',  whereas the
latter tests UDP as well as ICMP.

More alarming is the fact that "tests/tftp.sh" is so severly
GNU/Linux-centric that it fails completely in all aspects
for OpenBSD. It lacks every aspect of robustness and self
diagnosis. I had to impose the switch "-d" on Inetd at all
times in the script, in order make it robust, so the printout
is not as beutiful as was. On the other hand, it does no
longer conceil any false positives.

The changes needed to the test scripts are contaied in

    0001-Improve-the-present-test-script.patch


Once the test script for TFTP is functional on GNU/Linux
and OpenBSD, using the tools built by Inetutils itself,
it became evident that "src/inetd.c" is broken in OpenBSD.
It is the old friend of "optimistic calculation of
address structure length" that shows up again. It is
resolved by an additional integer that records the used
sockaddr length for an established connection. The expected
solution is not very intrusive at all.

The few changes are recorded in

     0002-src-inetd.c-Use-exact-length-of-address-structure.patch


Best regards,

Mats

Attachment: 0001-Improve-the-present-test-script.patch
Description: Text Data

Attachment: 0002-src-inetd.c-Use-exact-length-of-address-structure.patch
Description: Text Data

Attachment: signature.asc
Description: Digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]