In early October 2014, a researcher discovered a serious privilege escalation vulnerability in Apple’s Mac OS X operating system. The flaw, which has been around since at least 2011, has been finally fixed by the company.
Software updates released by Apple on Wednesday address tens of security holes, including a hidden backdoor API that can be exploited to gain root privileges. The vulnerability, dubbed “rootpipe,” was discovered last year by TrueSec security software engineer Emil Kvarnhammar. Now that Apple has patched the issue with the release of OS X Yosemite 10.10.3, the researcher has provided additional technical details on the bug.
The vulnerability (CVE-2015-1130), which plagues the Admin framework in OS X, existed since at least the release of OS X 10.7. According to Kvarnhammar, the flaw can be exploited by an attacker with physical access to the targeted device, but a remote attacker can also leverage the bug in combination with remote code execution exploits.
“[Apple’s] intention was probably to serve the ‘System Preferences’ app and systemsetup (command-line tool), but there is no access restriction. This means the API is accessible (through XPC) from any user process in the system,” the researcher explained in a blog post.
Initially, Kvarnhammar’s exploit only worked for users with administrator permissions, but he later got it to work for all users. It’s worth noting that user accounts created when OS X is installed have admin privileges by default.
Apple was informed of the issue on October 3. The company initially told the researcher that it was planning to patch the vulnerability in mid-January. However, it only managed to fix it in early March with the release of OS X 10.10.3 public beta. The patch was delivered on Wednesday to all users.
Apple has pointed out that the fix for this flaw will not be backported to OS X 10.9.x and older versions because it required significant changes to be made.
Another interesting vulnerability patched by Apple on Wednesday was discovered by Kaspersky Lab researchers. In a blog post published on Friday, the security firm disclosed the details of a Darwin kernel flaw (CVE-2015-1102) that can be exploited to cause devices running OS X or iOS to enter a denial-of-service (DoS) state. Because TCP headers are not processed correctly, an attacker can cause a system to crash simply by sending it a specially crafted network packet.
Kaspersky researchers have pointed out that many firewalls and routers are designed to drop invalid packets, but they have found a way to bypass these mechanisms. “A true professional can easily use [the vulnerability] to break down a user’s device or even interrupt the work of a corporate network,” experts said.
One of the vulnerabilities fixed by Apple in the Safari browser was reported in January by Jouko Pynnonen of the Finland-based company Klikki Oy. The vulnerability (CVE-2015-1126) allows an attacker to bypass the browser’s cross-domain policies and access resources belonging to another website. A malicious actor simply needs to get the victim to visit a website containing the exploit in order for the attack to work.
“The attacker could access or modify cookies belonging to another website in order to hijack a logged-in session of that website,” Pynnonen told SecurityWeek.
The researcher’s example exploit targeted cookies, but Apple noted in its advisory that “visiting a maliciously crafted website may lead to resources of another origin being accessed.” The fact that Apple used the term “resources” could indicate that not just cookies can be accessed, but the company hasn’t confirmed this when asked by the researcher.
“Technically, the [attacker’s] page would contain an IFRAME tag pointing to a specially formatted URL. Due to the bug, Safari would ‘miscalculate’ the domain of the attacker-supplied IFRAMEd document. That document would then have access to cookies of another domain, which could be anything the attacker chooses (e.g. apple.com),” the researcher explained.
“Another attack scenario is ‘session fixation’, i.e. the attacker would write the victim’s session cookie and wait for the victim to authenticate that session. After a successful attack the attacker would gain login access to the website with the victim’s credentials,” Pynnonen added.