Slashdot Log In
Firefox Susceptible To QuickTime Security Flaw
Posted by
kdawson
on Tue Nov 27, 2007 03:56 PM
from the exploit-available-in-the-wild dept.
from the exploit-available-in-the-wild dept.
Hugh Pickens writes "Apple's QuickTime media player software contains a previously undocumented security weakness in the way QuickTime handles the RTSP media-streaming protocol. The vulnerability is present in QuickTime versions 4.0 through 7.3 (the latest version) on both Windows and Mac systems. Symantec has tested the publicly available exploit code and found that it failed to work properly against Internet Explorer 6/7 or Safari 3 Beta but the exploit works against Firefox if users have chosen QuickTime as the default player for multimedia formats. Firefox users are more susceptible to this attack because Firefox farms off the request directly to the QuickTime Player as a separate process outside of its control, while IE loads the QuickTime Player as an internal plugin and when the overflow occurs, standard buffer-overflow protection is triggered, shutting down the affected processes before any damage can occur."
Related Stories
This discussion has been archived.
No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
Full
Abbreviated
Hidden
Loading... please wait.
And this is a firefox problem... (Score:5, Insightful)
Re:And this is a firefox problem... (Score:5, Insightful)
Guess they want the more hits by throwing Fox into the mess though, but really, why have Mozilla fix Apple's flaws?
Parent
Re:And this is a firefox problem... (Score:5, Informative)
Why? Just why?
Every website that has a quicktime video, I just go straight to youtube and search for the equivalent.
This is mainly due to the fact that the quicktime plugin traditionally hasn't been able to automatically install. You have to actually go to their website and install some adware filled crap that will never leave your system tray alone.
*bends over ready for -5 apple bashing*
Parent
Re:And this is a firefox problem... (Score:5, Insightful)
Besides, this is Slashdot. Since when did the headlines make sense?
Parent
Re:And this is a firefox problem... (Score:5, Interesting)
Glass half empty, half full type thing. Of course, Quicktime is causing the problem, but would you rather have a browser that arbitrarily trusts the plugin, or does some bounds checking?
Parent
Re:And this is a firefox problem... (Score:5, Funny)
The optimist says the glass is half full. The pessimist says the glass is half empty. The scientist says there is
Parent
Re:And this is a firefox problem... (Score:5, Funny)
Jack Bauer found out where the glass was, who drank the water, and which government they worked for.
Parent
Re:And this is a firefox problem... (Score:4, Insightful)
OTOH, babysitting probably takes up more resources so a paranoid OS will slow down. But IMHO the solution is still to taint dangerous stuff (what you got just downloaded) and have the OS babysit it.
Parent
Re:And this is a firefox problem... (Score:4, Insightful)
So taking your logic further, the OS should be responsible for all of this, so it's not even Firefox's problem. ^_^ Apps should be purpose built and responsible for that purpose. If you do the blame game up the line, you'll find tremendous bloat (more so than it already is) creeping into all first-line programs and even more so to the OS. If you don't blame Microsoft and OSX (the only two platforms Quicktime runs on, IIRC) as much as Firefox, you have violated your own thinking line.
Parent
Re:And this is a firefox problem... (Score:5, Insightful)
IE uses a plugin interface to deal with QuickTime. As such, it has a standard framework which does some bounds checking and can find buffer overflows like this one and kill a plugin (or iexplore.exe if necessary) preventing damage.
Firefox just passes parameters on to an external program.
Pick your poison, you can probably make justifications for either, but to me the IE method makes more sense. It's embedded content, it should be handled as a plugin to the parent application. You are a programmer, I'm sure you are familiar with the concepts of parents and children
Parent
Because of the end appearance (Score:5, Insightful)
So while it isn't FF's responsibility to fix the specific bug, it could be an indication of how things should be done better.
Parent
Re:And this is a firefox problem... (Score:4, Informative)
It does not really matter that the 'actual' vulnerability is in Quicktime. Firefox is the application that controls whether this vulnerability will affect the user, since it is obvious that is it possible to have code in Firefox that stops this exploit from working.
It is also a Firefox problem because any other plugin of this type is equally vulnerable using Firefox. From a secure coding point of view, is it your problem if you create an avenue whereby an exploit can occur? Damn straight! In this case, perhaps running the plugins in a controlled and monitored sandbox would be a good design change, instead of forking another process...
Parent
Re:And this is a firefox problem... (Score:5, Interesting)
I 90% agree with you; however, I do think operating systems should handle transactions with internet applications differently than normal processes. Both Vista and Leopard and any Linux distro with SELinux enhancements has the ability to sandbox certain processes for added security. The reason this exploit does not work with IE is because runs it as a plug-in and sandboxes all of those plug-ins within IE. I'd argue that any process to which data is "handed off" by a Web browser, e-mail client, or chat client should run in a sandbox as an extra layer of protection against this common type of attack.
Yeah, Quicktime is the culprit here and Firefox is not to blame, but I'd argue that the OS (all of them currently) is partly to blame for not sandboxing data coming into the machine via the Web.
Parent
Re:And this is a firefox problem... (Score:5, Funny)
Parent
That does it for me... (Score:5, Funny)
Re:That does it for me... (Score:4, Insightful)
Funny that security is not touted as much as a feature anymore compared to the early Firefox releases.
Parent
How is this a firefox problem? (Score:3, Insightful)
RT2FA: It's NOT a Firefox plugin issue (Score:4, Informative)
Non-Firefox browser: exploit fails to execute, instead protected by bounds checking
Firefox: exploit executes unchecked
How is that NOT a Firefox problem? If you don't use Firefox, you're immune. If you do, you're vulnerable. Even if the final cause is currently QuickTime, it's only a matter of time until some other plugin is found vulnerable and exploitable under Firefox but nowhere else.
Besides, Firefox and IE use different plugin models. Apparently the flaw is with Firefox's plugin model - clearly a Firefox problem.
The headline should read "Vulnerability in QuickTime. IE mitigates attacks via its QT plugin. Firefox doesn't fix problem in QT."
Per the Symantec article, the issue as related to Firefox is not with a plugin. The article states that QuickTime is run as a plugin inside IE and Safari. The vulnerable software is run inside the browser, and thus falls under the browser's control. http://www.symantec.com/enterprise/security_response/weblog/upload/2007/11/Image_IE.html [symantec.com] shows this. However, in the case of Firefox, QuickTime is run as a standalone app outside the browser. See http://www.symantec.com/enterprise/security_response/weblog/upload/2007/11/Image_FF.html [symantec.com]. In this case, Firefox gets Item A and sees that the system is configured to handle that type of item with Program B. Therefore, Firefox hands Item A to Program B. It works exactly the same as launching the malicious file from the Run box.
Once again, it is not a problem with Firefox's plugin system because this is not running as a Firefox plugin. Let me correct your quote. See how that makes it a little less cut and dried?
If there were a vulnerability in your email or FTP program, would you blame Firefox because it hands off mailto: and ftp: links to those external programs? Should Firefox be held responsible for malicious files (of any type - Word, MP3, .exe, etc.) that you download and then run externally? The Symantec article also mentions emailing attachments as an attack vector. Uh oh, Outlook and Thunderbird are also flawed, because they hand the file off to QuickTime to open too!
Also, judging by the IE pic, it appears that their "buffer overrun protection" is "crashing the browser". In this case, the QT vuln is also a DoS against IE, while Firefox does not have that vulnerability.
I agree that every program should do what it can to limit damage. However, Firefox can't do much about completely external programs. In this case, Firefox has no understanding of the data being downloaded, just that the system is configured to handle the data with a certain program. The only way to fix this is with filename/URL blacklisting so it doesn't open the bad URL (gee, that's practical) or by coding Firefox to understand every type of data it encounters. Essentially, code every other program into Firefox itself so that it can determine if the data is good or bad before handing it off (gee, that's practical). If this were a problem with a Firefox plugin, I would agree with you fully. However, it's a completely external program which Firefox has no control over, so I can't disagree more.
Parent
Apple software not secure. (Score:4, Insightful)
Troll -1 (Score:4, Funny)
Design for maliciousness (Score:4, Insightful)
Software should be pessimistic. Design the code to handle incoming requests as potentially malicious, and you'll never be disappointed.
A bigger problem (Score:5, Insightful)
Why can't about:plugins just have a 'disable' box on each plugin? Or, better yet, a standard preferences menu list which just lets me disable them there and then?
Re:A bigger problem (Score:5, Informative)
Parent
Quicktime is the FF plugin from hell (Score:5, Informative)
2. You HAVE to install Quicktime if you want to use iTunes
3. You (sort of) HAVE to install iTunes if you want to use an iPod (although I strongly recommend people consider Winamp, which has native support now, or the excellent ml_ipod plugin for Winamp)
4. Quicktime's browser plugin commandeers associations with a whole range of media types whether you want it to or not
5. QT doesn't give you the option of launching QT in a totally separate window - it automatically opens things embedded in the browser and starts playing them
6. QT seems to totally screw the ability to get Firefox to go back to launching media files with the good old "Open with..." dialog box, which lets you decide whether to open it, what to open it with, or whether to save it to disk
7. QT has absolutely no regard for what other media players and file association you might already have configured for your browser
and I guess we can add 8, although it was already implied
8. QT is a buggy p.o.s. with worse functionality and security than any half-decent media player including VLC, Winamp, and (in my humble opinion) even the dreaded WMP.
All of this reflects Apple's horrible attitude to developing software for the PC, which is essentially that they will utterly ignore the now well-established conventions of the platform in terms of installation behaviour, GUI and menu structure, and plugin behaviour and just run roughshod over the whole thing. Which would probably be more acceptable if their software JUST WORKED and was as fully featured as other options on the PC - but unfortunately that is not the case.
Parent
Symantec is wrong... (Score:4, Informative)
http://erratasec.blogspot.com/2007/11/new-rtsp-quicktime-flaw-affects-both.html [blogspot.com]
Standard buffer overflow protection doesn't work, Symantec was wrong. It seems that parts of Quicktime are not enabled for ASLR making these attacks possible.