On TechRepublic: 12 tech terms that make you sound old
BNET Business Network:
BNET
TechRepublic
ZDNet
TalkBack 37 of 99:
Next »
« Previous
Not quite true.
While I'm no big fan of either MS or Intel, this wasn't planned by either of them, nor is it in any way their own fault.

The fault lies in the driver for the Hollywood DVD PCI card. It violates a basic rule of programming on memory-managed OSes and CPUs (including all 32-bit and 64-bit CPUs, regardless of manufacturer): do not put executable code into memory that is not specifically allocated to your application and specifically allocated for the purpose of holding executable code, and executing it.

The driver has executable code mixed in with data and placed in an area allocated for data, not code. This was bad all along, but the badness only became apparent when you install SP2 and use it with a CPU that enforces No-Execute status on memory blocks allocated for data. To date, only the AMD 64-bit CPUs actually enforce this, and then only with an OS that helps, and the first version of Windows to do so is XP SP2. But Intel is already readying Xeons and Pentiums with their own version of No-Execute (I believe that they call theirs "Execute Disable," but it's the same thing), and the same "problem" would happen with them.

The purpose of No Execute / Execute Disable is to block what is BY FAR the most common cause of the security vulnerabilities that have long plagued Windows in general: unchecked buffer and stack overflows in systems-level software or software applications or drivers running with enhanced privileges. If code is allowed to execute in a data-only area, then a tricky hacker can force a buffer or stack overflow in any of a vast number of ways (look carefully at the descriptions of most of those smaller Critical Updates that Windows Update asks you to install from time to time -- most of them will be about "buffer overflows"), and stick executable code into the overflow, tricking the CPU into executing it with the same privileges as the program which allocated the buffer (often part of Windows itself, or a Windows system-level program, or device driver, etc.). This means that such code can blow right past most of the protections inherent in Windows, and do absolutely ANYTHING that the hacker desires. That could be anything from deleting important system files and thus crashing your system, to more subtle but more dangerous things like watching you type your passwords and credit card numbers and the like into your Web browser, then sending those back to the virus writer on the sly (after all, you can always re-install your OS or applications, but if your identity is stolen, your life is basically screwed for quite some time while you try to repair the damage to your very identity).

By enforcing data-only (no execute) status on such buffers, ALL buffer-overflow attacks and exploits are blocked, since the CPU will flatly refuse to execute ANY executable code that is not residing in a memory area specifically allocated for executable code!

In other words, the problem happened precisely because both SP2 and the AMD 64-bit chips are doing EXACTLY WHAT THEY WERE ADVERTISED TO DO. Again, this SAME problem would show up on the forthcoming Intel CPUs with Execute Disable.

I'm just surprised that only one ancient driver for one ancient piece of hardware has so far shown this problem. It's amazing that (so far) only one program still out there was written by some hack programmer who didn't think that following the rules of memory management applied to him/her.
Posted by: Joel R   Posted on: 08/28/04 You are currently: a Guest | Members login | Terms of Use

Alert moderator to an offensive message

Subscribe to this discussion via Email or RSS

Nice headline  bidemytime | 08/27/04
AMD Could be the problem here  gary.douglas@... | 08/27/04
But the story said...  bidemytime | 08/27/04
The problem also affects Intel CPUs  threedaysdwn | 08/27/04
This is Bull Schitt!!!  rdurso@... | 08/27/04
what?  doh123 | 08/27/04
It's neither...  techboy_z | 08/27/04
pitted  gary.douglas@... | 08/27/04
Whew!  bhanes@... | 08/27/04
DOH123 was right. You are wrong, as was the article writer.  Joel R | 09/08/04
Drivers are the achilles heel of the PC (NT)  P. Douglas | 08/27/04
Tell me about it  php_developer | 08/27/04
Non-sequitur  Yagotta B. Kidding | 08/27/04
Is mpegport.sys supplied by MS? i was trying to find out  The King's Servant | 08/27/04
How to find out  Yagotta B. Kidding | 08/27/04
Yeah but I don't have it.  The King's Servant | 08/27/04
Re: Non-sequitur  P. Douglas | 08/27/04
Re: Non-sequitur (Second Try)  P. Douglas | 08/27/04
Re: Non-sequitur (Third Try)  P. Douglas | 08/27/04
Re: Non-sequitur (Fourth [and Last]Try)  P. Douglas | 08/27/04
One other thing ...  P. Douglas | 08/27/04
One other thing ... (Second Try)  P. Douglas | 08/27/04
One other thing ... (Third Try)  P. Douglas | 08/27/04
M$ needs to pull back and admit they screwed up!  drichards1953 | 08/27/04
Dude, they'd have to go back to 1994, then!  Xunil_Sierutuf | 08/27/04
Fire Up The Delorean Doc, Marty's Got A New Mission  itanalyst | 08/27/04
It's a sad day  threedaysdwn | 08/27/04
A joke, right?  AmusedAtItAll | 08/28/04
No, experience.  threedaysdwn | 08/28/04
Migrating?  NemesisNL | 08/29/04
Back up Bubba Boy!  drichards1953 | 08/28/04
To date there are NO conflicts with SP2  Joel R | 09/08/04
What would you expect from MICROSUCKS ???  realitycheck101 | 08/27/04
Not MS  rpmyers1 | 08/27/04
This is not a bug  threedaysdwn | 08/27/04
hmm...inhell & m$ had to do something  mvaar | 08/27/04
Not quite true.  Joel R | 08/28/04
Then please explain this  Pinkerton | 08/29/04
Not portrayed to be  JBL312@... | 08/27/04
Incompatible  Yagotta B. Kidding | 08/27/04
What is MS telling people?  The King's Servant | 08/27/04
Might be able to fix the problem by?  JBL312@... | 08/27/04
Solving With Sufficient Progress In TS  ParadigmOdyssey | 08/29/04
who's not thinking?  ryusen | 08/27/04
What?  agottschald | 08/27/04
Why not tell the truth?  R.T.F.M. | 08/27/04
You are correct...  mvaar | 08/27/04
Is That The  ParadigmOdyssey | 08/29/04
Umm  threedaysdwn | 08/29/04
Make a BookMark Your In Luck  ParadigmOdyssey | 08/29/04
Not a SP2 or AMD64 problem!!  Guy Fawkes | 08/27/04
RE: Not a SP2 or AMD64 problem!!  JimOTR | 08/27/04
It is a piece of hardware in a PCI slot  The King's Servant | 08/27/04
Your arrogant and clueless remarks  madlikehell | 09/27/04
What is with ZDNet's talkback forums ???  MrEMann | 08/27/04
Wow. What did Intel have to do?  Yagotta B. Kidding | 08/27/04
This problem affects Intel chips as well  threedaysdwn | 08/29/04
Does this driver have WHQL certification?  The King's Servant | 08/27/04
When Pigopolists Fly  Yagotta B. Kidding | 08/27/04
Correct me if I am wrong but  The King's Servant | 08/27/04
Linux and DVDs  Yagotta B. Kidding | 08/27/04
because of Vendor IP, not evil empires.  balsover | 08/27/04
Not enough data  dberlind | 08/27/04
Not enough data? Think again.  dpbakeril | 08/27/04
I can tell ya this  mac@... | 08/27/04
That Was  ParadigmOdyssey | 08/29/04
At least there will be software available  American Pride | 08/27/04
Excuse me?  Judas I. | 08/27/04
Sure you do  American Pride | 08/27/04
What is clear ...  Judas I. | 08/27/04
Ahh, only atheists like Linux, huh?  Joel R | 08/28/04
Pride goeth before destruction, and an haughty spirit before a fall.  agottschald | 08/27/04
ANOTHER nickname change?  Linux_Developer | 08/27/04
and yet another reason...  ryusen | 08/27/04
What I mean to say is  loser_z | 08/27/04
poor ORIGINAL ZZ, now you have an idiot following you  Monkey_MCSE | 08/27/04
THE one and only  loser_z | 08/27/04
nonono...  ryusen | 08/30/04
No Problem with SP2  dony_z | 08/27/04
MS will make sure it works with AMD  American Pride | 08/27/04
Windows SP2 and AMD  mondoville | 08/27/04
Windows is working in the right direction  loser_z | 08/27/04
Leave it to Microsoft to break the finest processor available  BitTwiddler | 08/27/04
INTEL the finest processor  loser_z | 08/27/04
uhhh...  doh123 | 08/27/04
Rejoice!  Chiatzu | 08/28/04
It's neither Microsoft's nor AMD's fault.  Joel R | 08/28/04
But the story also said  bob@... | 08/27/04
Ignorant and Stupid Reporting!!!  rdurso@... | 08/27/04
Proof that AMD is just a chip company, not platform  Prognosticator | 08/27/04
You're right, and...  KOS-MOS | 08/27/04
You're both wrong, and...  Joel R | 08/28/04
good call  zeusfuse | 10/04/04
Experienced Windows users know . . .  CheeseToast | 08/27/04
SP2 on AMD 64  art@... | 08/27/04
Why would anyone use a realmagic decoder on Athlon 64?  SteveS_z | 08/28/04
if you want 64 bit support buy longhorn - lol  V Sanders | 08/28/04
I think you are wrong  NemesisNL | 08/29/04
why does a web page "windows update"  V Sanders | 08/30/04

What do you think?

SponsoredWhite Papers, Webcasts, and Downloads

advertisement

SmartPlanet

Click Here