Click on our Sponsors to help Support SunWorld
You state in your recent SunWorld article on Java and ActiveX
January letters to the editor
"Microsoft has also adopted this model and is implementing, in the Internet Explorer 3.01 and 4.0 releases, the same type of granularity in security control. Code
signature revocation is available today in Internet Explorer 3.01. A good example of the value of this feature is the infamous Exploder control attack on ActiveX. In
that situation, the developer of the control applied and received a personal digital signature from Verisign. The control was then posted on the Internet and
downloaded by users to demonstrate that ActiveX controls could be signed and still cause damage.
Shortly after the control was posted, Verisign notified the author that he had violated the agreement with
Verisign by using the digital signature to post malicious
code. The code was removed from the Internet and the
author's digital certificate with Verisign was revoked.
Now Internet Explorer can be commanded to check
the status of certificates and remove those that have
been revoked, disabling those ActiveX controls that are
signed by that revoked author."
Such a response is all well and good provided the trust
ensurence mechanism (Verisign, here) knows the identity
of the rogue developer using the mechanism in this torjan
horse manner. If the author of the Exploder ActiveX control
had never made his identity voluntarily known to us all,
what feature of this trust model for software deployment
would have figured out his identity and lead to the same
results, and how long (or how many episodes of destructive
behavior with self-cleaning of the audit trail) would there
be before some audit trail could be established to determine
the identity of the attacker?
The scenario is identical to the torjan horse uploads to
bulletin boards in years past. The most efficient way to
defuse such "digital bombs" is to ask the trust mechanism
(here Verisign) to first test-run every ActiveX control
on a stand-alone or sacrifical machine in an environment
sufficently representative of the broadest base of potential
downloaders for the control before the control is permitted
to be released with a signiture. This is akin to the
BBS operator testing/scanning all posts to the BBS from
Note that efficiency here refers to time/resources spent to
isolate a rogue ActiveX control before it does damage. From
the point of view of business distribution costs, this
may be rather expensive since every ActiveX control from
every developer would be bottle-necked through this trust
Virtual Vision, Inc.
Setting up anonymous FTP for Solaris
To Peter Galvin:
I just set up anonymous FTP for Solaris (previously I had configured it
for SunOS, which is of course totally different), and I have a few extra
tips in addition to what is mentioned in the ftpd man page. You should
add them to your security FAQ.
I'm using Solaris 2.5.1, but this should apply equally to Solaris 2.5.
1) In order to show the correct user and group ID's in "ls -l", you need
to copy the /usr/lib/libmp.so.1 library in addition to all of the other
libraries in ~ftp/usr/lib.
2) If you're using NIS+, and want to search NIS+ for user and group ID's
(again, for "ls -l"), you need to copy NIS_COLD_START and
NIS_SHARED_DIRCACHE from /var/nis to ~ftp/var/nis. You also have to copy
/etc/nsswitch.conf to ~ftp/etc, of course.
3) You need the correct timezone info in ~ftp/usr/share/lib/zoneinfo, but
you don't need ALL the timezones. If you want to save almost 400k, you
can copy just the timezone you need (for example "US/Pacific"). Do an "ls
-l" from within FTP and from a shell prompt, and make sure the two times
Here's another tip if you are having any problems with anonymous FTP (for
example, if you want to set up wu-ftpd). Follow these steps and you can
run a system call trace on ls in the same chroot'ed environment that ftpd
# mount -F lofs /proc ~ftp/proc # loopback mount /proc (truss needs it)
You will need to substitute ~ftp for the absolute pathname of the FTP home
To Peter Galvin:
Thank you very much for all of the tips and advise you have made available
through SunWorld! I was reading the Security FAQ and have a question on one
topic you addressed.
The question was "2.9 Why do Solaris machines act as routers?" You mention that
routing can be turned off on Solaris 2.5 by touching the file /etc/notrouter.
Currently I disable routing by setting a default router in /etc/defaultrouter.
/etc/defaultrouter has the IP address of the network router.
If I set /etc/notrouter and don't have /etc/defaultrouter, will Solaris be
able to access other networks? Will Solaris auto-discover the router?
# cp /usr/bin/truss ~ftp/usr/bin # copy the truss command itself
# chroot ~ftp /bin/truss /bin/ls -l /pub # or any other command
# rm ~ftp/usr/bin/truss
# umount ~ftp/proc
Square D Company
Peter Galvin responds:
These two mechanisms (Why do Solaris machines act as routers and turning off routing by touching the file /etc/notrouter) are separate. If your Solaris machine has only
one network interface, it won't route. If it has two or more, it will
route packets between the networks. This is sometimes indesirable,
especially when the Sun connects two networks in different security
domains. In that case you'll want to prevent the Sun from acting as
a router and effectively connecting the two networks together.
Concerning the question about disabling routing by setting a default router in /etc/defaultrouter, this doesn't disable the Sun from acting as a router. Rather, it tells
the Sun where to send it's packets for routing, that is, it tells the
Sun to how the Sun can get to networks it doesn't know about. It has
nothing to do with whether the Sun will take packets from one of its
networks and send them to another of its networks.
Finally, concerning your question about setting /etc/notrouter and not having /etc/defaultrouter, if you look at the startup files (/etc/init.d/inetinit I believe),
you'll see that route discovery is turned on by default, so yes, you
can disable the Sun acting as a router while still having it
automatically discover routes and use them to send packets from
itself to other networks. It's not *passing* packets for other
Help needed with mail
Please, help me. We have a
workstation running Solaris 2.5 and it is refusing to receive mail. I mean, I can
send mail from the machine, but I can't receive mail on it. Here you have the
return message. Looks like "local configuration error" and "mail loops back
to myself." Is this a problem of sendmail.cf or could be something wrong
with the smart relay host? I tested the "telnet indis 25" and the mail
daemon is alive. I tested "mailx -v adrianc@indis" too, and the dialogue
between the machines seems to be appropriate. What else could I do? Thank you very much.
Do you have any suggestions for Adrian, who lives in Romania? If so, please address them to me and I will print them next month.
Reader needs help with PPP on Solaris
Reading through your introduction at the end of my SunWorld e-mail
to believe you were both knowledgeable and approachable. So I come to you on
bended key, asking for your assistance with Sun PPP.
Just to give you a brief background, I have strong experience with Windows
flavors), (and in fact, am currently dialed in using a PC with PPP), but
only been working with Unix, specifically Solaris, for about six months. I
taken SA-135 and SA-285, but have been unable to setup PPP, and no one
know how. Not even at the Sun training facility I attended in Chicago.
The corporate office Unix gurus(in the company I work for), claim they
a clue how to setup PPP, and that they use ISDN from home. We have a
Sun Support account, which should enable us to get answers to questions like
this, but I have tried three times to get assistance. The first two people I
spoke with didn't have the knowledge to help me. The last gentleman I spoke
with, sounded a little older, and more knowledgeable, and e-mailed a forty
document that was not terribly helpful.
My complaint is that the document (like many in the Unix world), was
a style that leads you to believe it was composed by a half awake engineer
under a severe deadline, who never actually tried to use it to see if it
Ok, that was a slam, but I am frustrated. I have a Sun at home, and want
able to dial in with it to administer the network, but have been unable to
sense of the information. I am also frustrated in general with the Unix
educational community. Surely a class of people this intelligent, who can
the wonderfully complex Unix OS, could spawn someone who can communicate.
Do you or someone you know, have the ability to setup PPP on a Sun, and
you please help me?
Thank you for your time,
Riverside Publishing, Div. of Houghton Mifflin
Do you have experience with PPP on Sun? If you can shed some light on Scot's delimma, please send your answers to me (firstname.lastname@example.org), and I will publish it next month.
Kudos for SunWorld
Great job on the bar charts regarding what's hot.
Thanks. Glad you like them. We're always interested in knowing what works and what doesn't. It's interactive feedback like this that makes it possible to keep improving each month's issues. Readers, we'd like to hear back from you too. If there's something we do very well or poorly, let me know. Send your comments to: email@example.com. Make sure you include your name, and preferrably, your company affiliation too. We'll run them -- the good and the bad.
If you have problems with this magazine, contact
Last updated: 1 January 1997
Click on our Sponsors to help Support SunWorld
If you have technical problems with this magazine, contact