Apple SSL Bug!?

It's not clear what your question is. What else would you like to know?

The flaw is in the use of a browser on an unsecure wireless network which will permit a man-in-the-middle to read the wireless traffic. It DOES NOT permit a person to enter or compromise your computer or device. It does not affect wired network connections. If you are not using an unsecure public WiFi you should not be vulnerable.

It's not a hoax, but there is a lot of typical hyperventilating media wind being blown about it. Apple discovered a flaw in the way encrypted SSL communications are implemented if they were to take place over an unencrypted wireless network, such as can be found in coffee shops etc. Conceivably, a determined hacker could exploit that flaw to eavesdrop on email conversations that take place over such an unencrypted wireless network. It's worth noting that there is no evidence or knowledge of it having actually been accomplished by anyone for any specific nefarious purpose. It's also worth noting that there are many more, far simpler means of eavesdropping on private conversations including placing listening devices in your home or office, or simply looking over your shoulder while you're texting someone.

As a rule it's not a good idea to send or receive sensitive information while connected to a public, unencrypted Wi-Fi network. This will always remain true whether or not Apple decides to release an OS X update to address the flaw.

Nothing about this flaw is capable of altering your iPhone or Mac and is not something that can be used to take control of your Mac or access the information on it.

If it bothers you, use Firefox while on a public hotspot until a system update is released, probably in the next few days.

Ralph Landry1 wrote:

It does not affect wired network connections.

Yes it does. Obviousy it's more difficult to clandestinly plug into a router as opposed to joining by Wi-Fi, but once you are on the local network the Man-In-The-Middle attacks are identical.

Safari 7.0.1 and Firefox 26 are affected by this flaw.

APPLE-SA-2014-02-25-1 OS X Mavericks 10.9.2 and Security Update

2014-001

OS X Mavericks 10.9.2 and Security Update 2014-001 is now available

and addresses the following:

Apache

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: Multiple vulnerabilities in Apache

Description: Multiple vulnerabilities existed in Apache, the most

serious of which may lead to cross-site scripting. These issues were

addressed by updating Apache to version 2.2.26.

CVE-ID

CVE-2013-1862

CVE-2013-1896

App Sandbox

Available for: OS X Mountain Lion v10.8.5

Impact: The App Sandbox may be bypassed

Description: The LaunchServices interface for launching an

application allowed sandboxed apps to specify the list of arguments

passed to the new process. A compromised sandboxed application could

abuse this to bypass the sandbox. This issue was addressed by

preventing sandboxed applications from specifying arguments. This

issue does not affect systems running OS X Mavericks 10.9 or later.

CVE-ID

CVE-2013-5179 : Friedrich Graeter of The Soulmen GbR

ATS

Available for: OS X Mountain Lion v10.8.5,

OS X Mavericks 10.9 and 10.9.1

Impact: Viewing or downloading a document containing a maliciously

crafted embedded font may lead to arbitrary code execution

Description: A memory corruption issue existed in the handling of

handling of Type 1 fonts. This issue was addressed through improved

bounds checking.

CVE-ID

CVE-2014-1254 : Felix Groebert of the Google Security Team

ATS

Available for: OS X Mavericks 10.9 and 10.9.1

Impact: The App Sandbox may be bypassed

Description: A memory corruption issue existed in the handling of

Mach messages passed to ATS. This issue was addressed through

improved bounds checking.

CVE-ID

CVE-2014-1262 : Meder Kydyraliev of the Google Security Team

ATS

Available for: OS X Mavericks 10.9 and 10.9.1

Impact: The App Sandbox may be bypassed

Description: An arbitrary free issue existed in the handling of Mach

messages passed to ATS. This issue was addressed through additional

validation of Mach messages.

CVE-ID

CVE-2014-1255 : Meder Kydyraliev of the Google Security Team

ATS

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: The App Sandbox may be bypassed

Description: A buffer overflow issue existed in the handling of Mach

messages passed to ATS. This issue was addressed by additional bounds

checking.

CVE-ID

CVE-2014-1256 : Meder Kydyraliev of the Google Security Team

Certificate Trust Policy

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: Root certificates have been updated

Description: The set of system root certificates has been updated.

The complete list of recognized system roots may be viewed via the

Keychain Access application.

CFNetwork Cookies

Available for: OS X Mountain Lion v10.8.5

Impact: Session cookies may persist even after resetting Safari

Description: Resetting Safari did not always delete session cookies

until Safari was closed. This issue was addressed through improved

handling of session cookies. This issue does not affect systems

running OS X Mavericks 10.9 or later.

CVE-ID

CVE-2014-1257 : Rob Ansaldo of Amherst College, Graham Bennett

CoreAnimation

Available for: OS X Mountain Lion v10.8.5,

OS X Mavericks 10.9 and 10.9.1

Impact: Visiting a maliciously crafted site may lead to an

unexpected application termination or arbitrary code execution

Description: A heap buffer overflow existed in CoreAnimation's

handling of images. This issue was addressed through improved bounds

checking.

CVE-ID

CVE-2014-1258 : Karl Smith of NCC Group

CoreText

Available for: OS X Mavericks 10.9 and 10.9.1

Impact: Applications that use CoreText may be vulnerable to an

unexpected application termination or arbitrary code execution

Description: A signedness issue existed in CoreText in the handling

of Unicode fonts. This issue is addressed through improved bounds

checking.

CVE-ID

CVE-2014-1261 : Lucas Apa and Carlos Mario Penagos of IOActive Labs

curl

Available for: OS X Mavericks 10.9 and 10.9.1

Impact: An attacker with a privileged network position may intercept

user credentials or other sensitive information

Description: When using curl to connect to an HTTPS URL containing

an IP address, the IP address was not validated against the

certificate. This issue does not affect systems prior to OS X

Mavericks v10.9.

CVE-ID

CVE-2014-1263 : Roland Moriz of Moriz GmbH

Data Security

Available for: OS X Mavericks 10.9 and 10.9.1

Impact: An attacker with a privileged network position may capture

or modify data in sessions protected by SSL/TLS

Description: Secure Transport failed to validate the authenticity of

the connection. This issue was addressed by restoring missing

validation steps.

CVE-ID

CVE-2014-1266

Date and Time

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: An unprivileged user may change the system clock

Description: This update changes the behavior of the systemsetup

command to require administrator privileges to change the system

clock.

CVE-ID

CVE-2014-1265

File Bookmark

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: Viewing a file with a maliciously crafted name may lead to

an unexpected application termination or arbitrary code execution

Description: A buffer overflow existed in the handling of file

names. This issue was addressed through improved bounds checking.

CVE-ID

CVE-2014-1259

Finder

Available for: OS X Mavericks 10.9 and 10.9.1

Impact: Accessing a file's ACL via Finder may lead to other users

gaining unauthorized access to files

Description: Accessing a file's ACL via Finder may corrupt the ACLs

on the file. This issue was addressed through improved handling of

ACLs.

CVE-ID

CVE-2014-1264

ImageIO

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: Viewing a maliciously crafted JPEG file may lead to the

disclosure of memory contents

Description: An uninitialized memory access issue existed in

libjpeg's handling of JPEG markers, resulting in the disclosure of

memory contents. This issue was addressed by better JPEG handling.

CVE-ID

CVE-2013-6629 : Michal Zalewski

IOSerialFamily

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5

Impact: Executing a malicious application may result in arbitrary

code execution within the kernel

Description: An out of bounds array access existed in the

IOSerialFamily driver. This issue was addressed through additional

bounds checking. This issue does not affect systems running OS X

Mavericks v10.9 or later.

CVE-ID

CVE-2013-5139 : @dent1zt

LaunchServices

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5

Impact: A file could show the wrong extension

Description: An issue existed in the handling of certain unicode

characters that could allow filenames to show incorrect extensions.

The issue was addressed by filtering unsafe unicode characters from

display in filenames. This issue does not affect systems running OS X

Mavericks v10.9 or later.

CVE-ID

CVE-2013-5178 : Jesse Ruderman of Mozilla Corporation, Stephane Sudre

of Intego

NVIDIA Drivers

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: Executing a malicious application could result in arbitrary

code execution within the graphics card

Description: An issue existed that allowed writes to some trusted

memory on the graphics card. This issue was addressed by removing the

ability of the host to write to that memory.

CVE-ID

CVE-2013-5986 : Marcin Kościelnicki from the X.Org Foundation

Nouveau project

CVE-2013-5987 : Marcin Kościelnicki from the X.Org Foundation

Nouveau project

PHP

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: Multiple vulnerabilities in PHP

Description: Multiple vulnerabilities existed in PHP, the most

serious of which may have led to arbitrary code execution. These

issues were addressed by updating PHP to version 5.4.22 on OS X

Mavericks v10.9, and 5.3.28 on OS X Lion and Mountain Lion.

CVE-ID

CVE-2013-4073

CVE-2013-4113

CVE-2013-4248

CVE-2013-6420

QuickLook

Available for: OS X Mountain Lion v10.8.5

Impact: Downloading a maliciously crafted Microsoft Office file may

lead to an unexpected application termination or arbitrary code

execution

Description: A memory corruption issue existed in QuickLook's

handling of Microsoft Office files. Downloading a maliciously crafted

Microsoft Office file may have led to an unexpected application

termination or arbitrary code execution. This issue does not affect

systems running OS X Mavericks 10.9 or later.

CVE-ID

CVE-2014-1260 : Felix Groebert of the Google Security Team

QuickLook

Available for: OS X Mountain Lion v10.8.5,

OS X Mavericks 10.9 and 10.9.1

Impact: Downloading a maliciously crafted Microsoft Word document

may lead to an unexpected application termination or arbitrary code

execution

Description: A double free issue existed in QuickLook's handling of

Microsoft Word documents. This issue was addressed through improved

memory management.

CVE-ID

CVE-2014-1252 : Felix Groebert of the Google Security Team

QuickTime

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: Playing a maliciously crafted movie file may lead to an

unexpected application termination or arbitrary code execution

Description: A buffer overflow existed in the handling of 'ftab'

atoms. This issue was addressed through improved bounds checking.

CVE-ID

CVE-2014-1246 : An anonymous researcher working with HP's Zero Day

Initiative

QuickTime

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: Playing a maliciously crafted movie file may lead to an

unexpected application termination or arbitrary code execution

Description: A memory corruption issue existed in the handling of

'dref' atoms. This issue was addressed through improved bounds

checking.

CVE-ID

CVE-2014-1247 : Tom Gallagher & Paul Bates working with HP's Zero Day

Initiative

QuickTime

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: Playing a maliciously crafted movie file may lead to an

unexpected application termination or arbitrary code execution

Description: A buffer overflow existed in the handling of 'ldat'

atoms. This issue was addressed through improved bounds checking.

CVE-ID

CVE-2014-1248 : Jason Kratzer working with iDefense VCP

QuickTime

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: Viewing a maliciously crafted PSD image may lead to an

unexpected application termination or arbitrary code execution

Description: A buffer overflow existed in the handling of PSD

images. This issue was addressed through improved bounds checking.

CVE-ID

CVE-2014-1249 : dragonltx of Tencent Security Team

QuickTime

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: Playing a maliciously crafted movie file may lead to an

unexpected application termination or arbitrary code execution

Description: An out of bounds byte swapping issue existed in the

handling of 'ttfo' elements. This issue was addressed through

improved bounds checking.

CVE-ID

CVE-2014-1250 : Jason Kratzer working with iDefense VCP

QuickTime

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks 10.9 and 10.9.1

Impact: Playing a maliciously crafted movie file may lead to an

unexpected application termination or arbitrary code execution

Description: A signedness issue existed in the handling of 'stsz'

atoms. This issue was addressed through improved bounds checking.

CVE-ID

CVE-2014-1245 : Tom Gallagher & Paul Bates working with HP's Zero Day

Initiative

Secure Transport

Available for: OS X Mountain Lion v10.8.5

Impact: An attacker may be able to decrypt data protected by SSL

Description: There were known attacks on the confidentiality of SSL

3.0 and TLS 1.0 when a cipher suite used a block cipher in CBC mode.

To address these issues for applications using Secure Transport, the

1-byte fragment mitigation was enabled by default for this

configuration.

CVE-ID

CVE-2011-3389 : Juliano Rizzo and Thai Duong

OS X Mavericks v10.9.2 includes the content of Safari 7.0.2.

OS X Mavericks v10.9.2 and Security Update 2014-001 may be obtained from

the Mac App Store or Apple's Software Downloads web site:

http://www.apple.com/support/downloads/

Information will also be posted to the Apple Security Updates

web site: http://support.apple.com/kb/HT1222

The Mavericks update is at: http://support.apple.com/kb/DL1726

@MadMac0

Re:

"Feb 24, 2014 11:30 PM (in response to Ralph Landry1)

Ralph Landry1 wrote:

It does not affect wired network connections.

Yes it does. Obviousy it's more difficult to clandestinly plug into a router as opposed to joining by Wi-Fi, but once you are on the local network the Man-In-The-Middle attacks are identical."

That is false. You are basing that on an incorrect understanding of the actual bug. The bug only affected wireless connections using Safari, and only then, if you were on a completely unencrypted wifi that had no security or password... This isn't a simple man in the middle open bug, it was in the way SSL was negotiated only when using completely unencrypted wifi... It's likely nobody ever noticed this until Apple announced the flaw and fixed it because most people with any security conscienceness have at least some encryption on their wifi or are a little leary about trusting completely unencrypted, open wifi connections to begin with.

I don't know why you are choosing to reply to this now since it's all been taken care of, so I figured I'd let it go, but your statements are so far off base that I felt I needed to respond.

I spent the better part of three days gaining an understanding of exactly what this flaw was by researching what virtually all the security and coding experts had to say about it and then checking it all out for myself to the extent I was able, before posting one word here. Everything I've written is thoroughly documented and was vetted with some of the sources and a group of computer security colleagues that I collaborate with every day. We were even able to obtain a bit of confirmation from Apple product-security, but as you might expect they were unable to discuss most of the details.

Safari was not the only browser that was impacted by the flaw. My colleagues and I checked several browsers that we had handy and every one of them that use the WebKit framework was found to be vulnerable, using both the test sites that had been setup. This includes OmniWeb, Maxthon and some lesser known browsers. There are also several other applications that also use WebKit including Apple Mail, Entourage 2008, Instant Messenger, Fire, iChat, MSN Messenger, Yahoo Messenger, NetNewWire, iWeb, Sandvox, BBEdit, and the list goes on. Although no tests were setup to prove vulnerability, I think it's reasonable to suspect that any of those apps that use SSL were subject to spoofing attack.

Not only that, but I also tested these same browsers both over my local WPA-2 encrypted W-Fi network and when attached via Ethernet to my router. All tests showed those browsers to be vulnerable.

But the flaw wasn't in Safari or WebKit. The only security update involving Safari was for a very different WebKit issue:

Safari 6.1.2 and Safari 7.0.2 is now available and addresses the

following:

WebKit

Available for: OS X Lion v10.7.5, OS X Lion Server v10.7.5,

OS X Mountain Lion v10.8.5, OS X Mavericks v10.9.1

Impact: Visiting a maliciously crafted website may lead to an

unexpected application termination or arbitrary code execution

Description: Multiple memory corruption issues existed in WebKit.

These issues were addressed through improved memory handling.

CVE-ID

CVE-2013-6635 : cloudfuzzer

CVE-2014-1268 : Apple

CVE-2014-1269 : Apple

CVE-2014-1270 : Apple

The flaw was in the coding of the SecureTransport section of /System/Library/Frameworks/Security.framework/Version/A/Security. It's the section of the code that handles SSL certification. That section does not handle Wi-Fi encryption in any way. There was speculation just before 10.9.2 was released that SSL might be involved between Airport Utility and Airport Base Stations, but there wasn't time to check that out.

Finally, let's take a look at what Apple had to say about the fix:

Data Security

Available for: OS X Mavericks 10.9 and 10.9.1

Impact: An attacker with a privileged network position may capture

or modify data in sessions protected by SSL/TLS

Description: Secure Transport failed to validate the authenticity of

the connection. This issue was addressed by restoring missing

validation steps.

CVE-ID

CVE-2014-1266

"Privileged network position" does not say Wi-Fi only. It's SSL/TLS not WEP, WPA or any other Wi-Fi encryption protocol. If an attacker is able to hack your local network then he becomes the Man In the Middle whether open or encrypted.

What would have happened next is anybodies guess. Harvesting of credentials would undoubtedly be the first order of business. UserID's and passwords for web sites, e-mail accounts, AppleID's, chat accounts. Any application that used the Secure Transport function to do SSL/TLS certificate validation would be vulnerable to intercept by the MITM with fake certificates. FireFox and Google Chrome are said to use their own SSL/TLS certification process, which is why they were not involved with this flaw.

I will check this out. So far what you have gleaned contradicts all other basic public disclosures of this bug. But you bring up browsers nobody uses to say everyting was effected. They said it was webkit, but in reality webkit is Safari on OSX. Firefox and Opera and others don't use it, and Webkit is understood by anybody that knows the far reaching browsers you are trying to pull out of it. Of course they said webkit so if you twist you can say other browsers were affected on OSX, but you are misleading pulling those out and not saying FireFox or Opera weren't affected, which Apple also said. You checked those two and know different?

But, before actually going off on a tangent, are you just talking about IOS webkit and just webkit browsers? Basically, are you saying Firefox on OSX was affected, Opera was affected, or or any other OSX browser actually used by most when we talk "alternatives", other than Safari was affected? I will check out the mail stuff and other protocols, because I really only heard basically Safari browsing was affected. Priviledged is a term in the fix, but they also said it pertained to non encrypted wifi in the public warning, and I actually think that was accurate.

Oh my gosh, and reading deeper you are just saying "it is reasonable to suspect" other protocols. Never mind. You are way off. I wasted time learning your twisted thinking.

You are really are only talking about Safari webkit in practice, and saying other stuff *could* use those protocols and pulling obscure "non safari" webkit browsers, you are misdirecting and pulling randon non-meaning cases that nobody uses to prove a false fact... Anybody using webkit in some off browser knows that the announcement mentioned webkit. For the public, saying Firefox and Opera and basically any non-webkit browser was good was accurate. You see what you want.

I really don't see how you could tell browsers were vulnerable on eithernet unless you actually spoofed a certificate and didn't get any warnings, but I guess it doesn't matter. You could be right and I apologize for going off, and I can't test it now, it just felt like a lot of stretching.

I got security warnings all the time on wifi for self signed certs until I accepted them, and I switch certificates often enough and test new sites enough to run into them here and there. I just didn't assume you actually performed MIM attacks on yourself to find browsers vulnerable on ethernet or whatever this bug in theory was supposed to allow, so I didn't know what you meant by "found them vulnerable" and assumed you were just assuming.

I guess I just never saw this vulnerability using wifi or ethernet, but I never really use open wifi so I figured that was why. Thanks for the info to consider.

I do apologize for rehashing.

Like you said, it's all taken care of.

I have been dealing with so much misinfo for days I had to go back and understand.

I've been dealing with people quoting bug name news articles saying anyone on your network could spoof you which isn't how MIM attacks work, but there's no excuse. Your correction helped me get a grip and see how bad it could have been and thanks.