Mozilla has joined Apple in being among the first to fix the DLL load hijacking attack vector that continues to haunt hundreds of Windows applications.

The open-source group released Firefox 3.6.9 with patches for a total of 15 vulnerabilities (11 rated critical), including the publicly known DLL load hijacking flaw that exposes Windows users to remote code execution attacks.

The majority of the 15 vulnerabilities in this Firefox patch batch could be exploited to launch drive-by download attacks from booby-trapped Web sites.

According to Firefox, the DLL load hijacking issue only affects Windows XP users:

Firefox could be used to load a malicious code library that had been planted on a victim’s computer.

Firefox attempts to load dwmapi.dll upon startup as part of its platform detection, so on systems that don’t have this library, such as Windows XP, Firefox will subsequently attempt to load the library from the current working directory. An attacker could use this vulnerability to trick a user into downloading a HTML file and a malicious copy of dwmapi.dll into the same directory on their computer and opening the HTML file with Firefox, thus causing the malicious code to be executed. If the attacker was on the same network as the victim, the malicious DLL could also be loaded via a UNC path. The attack also requires that Firefox not currently be running when it is asked to open the HTML file and accompanying DLL.

Firefox 3.6.9 also fixes the following “critical” security problems issues:

  • MFSA 2010-59: Mozilla developer Blake Kaplan reported that the wrapper class XPCSafeJSObjectWrapper (SJOW), a security wrapper that allows content-defined objects to be safely accessed by privileged code, creates scope chains ending in outer objects. Users of SJOWs which expect the scope chain to end on an inner object may be handed a chrome privileged object which could be leveraged to run arbitrary JavaScript with chrome privileges. Mozilla said Google researcher Michal Zalewski’s recent contributions helped to identify this architectural weakness.
  • MFSA 2010-58: Security researcher Marc Schoenefeld reported that a specially crafted font could be applied to a document and cause a crash on Mac systems. The crash showed signs of memory corruption and presumably could be used by an attacker to execute arbitrary code on a victim’s computer.
  • MFSA 2010-57: Security researcher regenrecht reported via TippingPoint’s Zero Day Initiative that code used to normalize a document contained a logical flaw that could be leveraged to run arbitrary code. When the normalization code ran, a static count of the document’s child nodes was used in the traversal, so a page could be constructed that would remove DOM nodes during this normalization which could lead to the accessing of a deleted object and potentially the execution of attacker-controlled memory.
  • MFSA 2010-56: Security researcher regenrecht reported via TippingPoint’s Zero Day Initiative that the implementation of XUL ’s content view contains a dangling pointer vulnerability. One of the content view’s methods for accessing the internal structure of the tree could be manipulated into removing a node prior to accessing it, resulting in the accessing of deleted memory. If an attacker can control the contents of the deleted memory prior to its access they could use this vulnerability to run arbitrary code on a victim’s machine.
  • MFSA 2010-54: Security researcher regenrecht reported via TippingPoint’s Zero Day Initiative that there was a remaining dangling pointer issue leftover from the fix to CVE-2010-2753. Under certain circumstances one of the pointers held by a XUL tree selection could be freed and then later reused, potentially resulting in the execution of attacker-controlled memory.
  • MFSA 2010-53: Security researcher wushi of team509 reported a heap buffer overflow in code routines responsible for transforming text runs. A page could be constructed with a bidirectional text run which upon reflow could result in an incorrect length being calculated for the run of text. When this value is subsequently used to allocate memory for the text too small a buffer may be created potentially resulting in a buffer overflow and the execution of attacker controlled memory.
  • MFSA 2010-51: Security researcher Sergey Glazunov reported a dangling pointer vulnerability in the implementation of navigator.plugins in which the navigator object could retain a pointer to the plugins array even after it had been destroyed. An attacker could potentially use this issue to crash the browser and run arbitrary code on a victim’s computer.
  • MFSA 2010-50: Security researcher Chris Rohlf of Matasano Security reported that the implementation of the HTML frameset element contained an integer overflow vulnerability. The code responsible for parsing the frameset columns used an 8-byte counter for the column numbers, so when a very large number of columns was passed in the counter would overflow. When this counter was subsequently used to allocate memory for the frameset, the memory buffer would be too small, potentially resulting in a heap buffer overflow and execution of attacker-controlled memory.
  • MFSA 2010-49: Mozilla developers identified and fixed several memory safety bugs in the browser engine used in Firefox and other Mozilla-based products. Some of these bugs showed evidence of memory corruption under certain circumstances, and we presume that with enough effort at least some of these could be exploited to run arbitrary code.

Mozilla Firefox 3.6.9 is available for Windows, Mac and Linux users via the browser’s auto-update mechanism.

Categories: Cryptography, Vulnerabilities, Web Security

Comment (1)

  1. Woodrat 2296

    It’s too bad that FF 3.6.9 is the worst release ever. Since I installed it, a lot of times when I try to go to some sites, FF allows redirects to spam and crap pages that have NOTHING whatsoever to do with the site I’m trying to go to. This usually occurs when I click on a link and try to go to a site like that. Usually if  I type in the url, no problems.  FF 3.6.9 sucks. If this is the way FF is going, it’s time to return to IE or Opera or Chrome or whatever.

Comments are closed.