Four New Unpatched Windows Vulnerabilities

peeon writes "Right before Christmas, four new Windows NT/2k/XP vulnerabilities were posted to the Bugtraq list. This story discusses two of the vulnerabilities in the LoadImage function (buffer overflow) and Windows Help program (heap overflow), but the Chinese company discovered two more exploits in the parsing of a specially crafted ANI file (causes DoS). A Bugtraq posting has more details."

Comments

.. posted from newly esspee2d xp abomination

[maharg]

so it's christmas eve 2004, i'm at the in-laws, just spent 3 hours adawaring, spybotting, esspee2ing from a cd burnt on the latest stage 1. go figure.

30 megs of critical/av signatures to be done over diallup another time

damn you micro$hite

Re:Forced Upgrade.

[DrEvil]

It has to be a conspiracy. Anyone who claims that this might be a consequence of the year-long security push for SP2 and that a high-level fix made during this push might prevent certain classes of bugs from being exploitable is clearly evil and has been exposed to too much software engineering. I'd suspect such a person of spreading facts instead of FUD.

Re:Forced Upgrade.

[bryanp]

a) Nobody's forcing you to upgrade. I still haven't had Steve Ballmer show up on my doorstep with an Uzi yet.

b) The list you give is mostly patches. There are four base OS' on that list and 6 patches, all of which are free.

c) If it bothers you, feel free to run an unpatched OS of your choice, whether it be Windows, MacOS or one of the many *nix variants.

Bah!

[rubberband]

Hi, you've missed the point. I hope you're not trolling, because I'm going to bite.

Every box at my workplace is patched with SP2. In this case, it doesn't matter - one of the exploits is still useable.

The problem is not (this time, thankfully) the corporate enterprise deployment of windows. It's friends and family. Every time a new windows exploit like this comes out, jerk spyware/worm/virus writers are on it within 24 hours, populating their zombie networks with your mom's, friends' and families' computers. Manditory regular patching at work is easy. The same for people you see occaisionally who are not computer literate is not. These are the people who it really screws with - for example, all one of my buddies wants to do with his dell is play games, send email and surf. He knows nothing beyond that, and is certainly not going to run down to the basement on christmas eve to make sure his operating system is secure RIGHT NOW.

This business of "patch or you deserve it" is utter BS. I maintain that virus writers should be dragged into the street and beaten with keyboards, followed shortly by geeks who empower them by putting any of the blame on the end user. If I paid thousands for an OS site license, I should not be spending my holidays fixing it. If I spend hundreds for an oem copy at home, the same applies. The only ones who deserve ANYTHING bad here are the exploiters and the providers of the crappy OS in question.

Re:Bah!

[rubberband]

I still think the point is valid. Consider that a) That means that the vendor has had 7 years to secure their product. I any other industry they would have litigated into oblivion by now. It is *NOT* the end user's fault that the current world standard for personal computer operating systems is frequently bugged.

Sure, carrying $1000 in cash is dumb, but there are easily accessible alternatives. Credit cards, debit cards, traveller's cheques, travel wallets, etc are all viable alternatives. Carrying cash is like opening attachments from unknown senders. Getting your windows box 0wned without your action because a new exploit came out 8 hours ago is like the jacket manufacturer attaching a big red "steal from me!" sign to the back and cutting a pickpockt access hole out, too. (Except then they take over the world jacket manufacturing business and force you to wear one unless you want to freeze or learn to sew).

To use the token comparison to a vehicle - yes, when you buy a car you should be responsible enough to get it serviced from time to time, and act on any critical recall issues that might arise. You shouldn't however have to open the hood and check the internals 3 times per day to ensuire it doesn't explode and require expensive maintenance the next time you turn the key in the ignition.

Don't get me wrong - I'm not saying sysadmins should have no responsability whatsoever. They are after all paid to deal with systems. But when was the last time you head of a dell salesperson telling an unexperienced buyer that if they wish to have their computer on regularly they'll need to spend 5 minutes every single day, and an hour of two each week making sure they're machine doesn't get destroyed?

Re:Is it really this hard...

[twiddlingbits]

Nice try, but you should check the return code from malloc(). If it is -1 then there is a problem and you don't need to do the If statement. A lot of times the trouble comes not when allocating memory but when using a pointer to WRITE to memory. It's a C programmer trick to set up a pointer to a block of size X and write to it via the pointer, of course if you lose track of the pointer address you can easily go too far. Common errors are off by one in the count, assuming you are writing 8/16/32 bits without checking the underlying data type first,
or just writing to whatever address the pointer says w/o checking that *p > MAX_MEMORY_ADDRESS. These are errors a beginner programmer would make, and from the looks of how common these errors are in Windows that is the type of folks MS uses. It also says to me that they don't use any sort of Automated Code Analysis tools which can catch these sorts of errors. Or maybe they don't do any indpendant QA at all? It's pretty pathetic when the worlds most popular software is made by a company that probably doesn't meet SEI Level 2 criteria. I only wish that the laws allowed someone to sue for lost time/income from the "basic" errors that shouldn't have been present.

i wonder...

[hitmark]

why in this day and age, 99%-100% of automated exploits still happens to be some kind of overflow. why do we keep thinking that we dont have to check the sizes when moveing data about as its defined by a standard anyways? its like not checking to see if you have room for something in your house or car before buying it at the very least.

Definition of "Patched/Unpatched"

[jamesl]

Slashdot has made subtle changes to the definitions of Patched and Unpatched.

Patched Open Source: A vulnerability has been identified and someone is thinking about fixing it. Because the time between discovery and fix is vanishingly small, there are no unpatched open source vulnerabilities.

Patched Windows/Proprietary: A patch has been available for not less than 12 months and is installed on not less than 99% of affected systems. It will be several months, if not years, before vulnerabilities fixed by Windows XP SP2 will be considered patched.