computer hacks

1323 Articles

computer hacks

This Week In Security: KNOB, Old Scams Are New Again, 0-days, Backdoors, And More

August 23, 2019 by 12 Comments

Bluetooth is a great protocol. You can listen to music, transfer files, get on the internet, and more. A side effect of those many uses is that the specification is complicated and intended to cover many use cases. A team of researchers took a look at the Bluetooth specification, and discovered a problem they call the KNOB attack, Key Negotiation Of Bluetooth.

This is actually one of the simpler vulnerabilities to understand. Randomly generated keys are only as good as the entropy that goes into the key generation. The Bluetooth specification allows negotiating how many bytes of entropy is used in generating the shared session key. By necessity, this negotiation happens before the communication is encrypted. The real weakness here is that the specification lists a minimum entropy of 1 byte. This means 256 possible initial states, far within the realm of brute-forcing in real time.

The attack, then, is to essentially man-in-the-middle the beginning of a Bluetooth connection, and force that entropy length to a single byte. That’s essentially it. From there, a bit of brute forcing results in the Bluetooth session key, giving the attacker complete access to the encrypted stream.

One last note, this isn’t an implementation vulnerability, it’s a specification vulnerability. If your device properly implements the Bluetooth protocol, it’s vulnerable.

CenturyLink Unlinked

You may not be familiar with CenturyLink, but it maintains one of the backbone fiber networks serving telephone and internet connectivity. On December 2018, CenturyLink had a large outage affecting its fiber network, most notable disrupting 911 services for many across the United States for 37 hours. The incident report was released on Monday, and it’s… interesting.
Continue reading “This Week In Security: KNOB, Old Scams Are New Again, 0-days, Backdoors, And More” →

Build A Fungus Foraging App With Machine Learning

August 22, 2019 by 42 Comments

As the 2019 mushroom foraging season approaches it’s timely to combine my thirst for knowledge about low level machine learning (ML) with a popular pastime that we enjoy here where I live. Just for the record, I’m not an expert on ML, and I’m simply inviting readers to follow me back down some rabbit holes that I recently explored.

But mushrooms, I do know a little bit about, so firstly, a bit about health and safety:

  • The app created should be used with extreme caution and results always confirmed by a fungus expert.
  • Always test the fungus by initially only eating a very small piece and waiting for several hours to check there is no ill effect.
  • Always wear gloves  – It’s surprisingly easy to absorb toxins through fingers.

Since this is very much an introduction to ML, there won’t be too much terminology and the emphasis will be on having fun rather than going on a deep dive. The system that I stumbled upon is called XGBoost (XGB). One of the XGB demos is for binary classification, and the data was drawn from The Audubon Society Field Guide to North American Mushrooms. Binary means that the app spits out a probability of ‘yes’ or ‘no’ and in this case it tends to give about 95% probability that a common edible mushroom (Agaricus campestris) is actually edible. 

The app asks the user 22 questions about their specimen and collates the data inputted as a series of letters separated by commas. At the end of the questionnaire, this data line is written to a file called ‘fungusFile.data’ for further processing.

XGB can not accept letters as data so they have to be mapped into ‘classic LibSVM format’ which looks like this: ‘3:218’, for each letter. Next, this XGB friendly data is split into two parts for training a model and then subsequently testing that model.

Installing XGB is relatively easy compared to higher level deep learning systems and runs well on both Linux Ubuntu 16.04 and on a Raspberry Pi. I wrote the deployment app in bash so there should not be any additional software to install. Before getting any deeper into the ML side of things, I highly advise installing XGB, running the app, and having a bit of a play with it.

Training and testing is carried out by running bash runexp.sh in the terminal and it takes less than one second to process the 8124 lines of fungal data. At the end, bash spits out a set of statistics to represent the accuracy of the training and also attempts to ‘draw’ the decision tree that XGB has devised. If we have a quick look in directory ~/xgboost/demo/binary_classification, there should now be a 0002.model file in it ready for deployment with the questionnaire.

I was interested to explore the decision tree a bit further and look at the way XGB weighted different characteristics of the fungi. I eventually got some rough visualisations working on a Python based Jupyter Notebook script:

 

 

 

 

 

 

 

Obviously this app is not going to win any Kaggle competitions since the various parameters within the software need to be carefully tuned with the help of all the different software tools available. A good place to start is to tweak the maximum depth of the tree and the number or trees used. Depth = 4 and number = 4 seems to work well for this data. Other parameters include the feature importance type, for example: gain, weight, cover, total_gain or total_cover. These can be tuned using tools such as SHAP.

Finally, this app could easily be adapted to other questionnaire based systems such as diagnosing a particular disease, or deciding whether to buy a particular stock or share in the market place.

An even more basic introduction to ML goes into the baseline theory in a bit more detail – well worth a quick look.

Sushi Roll Helps Inspect Your CPU Internals

August 21, 2019 by 8 Comments

[Gamozolabs’] post about Sushi Roll — a research kernel for monitoring Intel CPU internals — is pretty long. While we were disappointed at the end that the kernel’s source is not exactly available due to “sensitive features”, we were so impressed with the description of the modern x86 architecture and some of the work done with Sushi Roll, that we just had to post it. If the post gets you wanting to actually try some of this, you can check out another [Gamozolabs] creation, Orange Slice.

While you probably know that a modern Intel CPU bears little resemblance to the old 8086 processor it emulates, it is surprising, sometimes, to realize just how far it has gone. The very first thing the CPU does is to break your instruction up into microoperations. The execution engine uses some sophisticated techniques for register renaming and scheduling that allow you to run instructions out of order and to run more than one instruction per clock cycle.

Continue reading “Sushi Roll Helps Inspect Your CPU Internals” →

This Week In Security: Black Hat, DEF CON, And Patch Tuesday

August 16, 2019 by 11 Comments

Blackhat and DEF CON both just wrapped, and Patch Tuesday was this week. We have a bunch of stories to cover today.

First some light-hearted shenanigans. Obviously inspired by Little Bobby Tables, Droogie applied for the vanity plate “NULL”. A year went by without any problems, but soon enough it was time to renew his registration. The online registration form refused to acknowledge “NULL” as a valid license plate. The hilarity didn’t really start until he got a parking ticket, and received a bill for $12,000. It seems that the California parking ticket collection system can’t properly differentiate between “NULL” and a null value, and so every ticket without a license plate is now unintentionally linked to his plate.

In the comments on the Ars Technica article, it was suggested that “NULL” simply be added to the list of disallowed vanity plates. A savvy reader pointed out that the system that tracks disallowed plates would probably similarly choke on a “NULL” value.

Hacking an F-15

In a surprising move, Air Force officials brought samples of the Trusted Aircraft Information Download Station (TADS) from an F-15 to DEF CON. Researchers were apparently able to compromise those devices in a myriad of ways. This is a radical departure from the security-through-obscurity approach that has characterized the U.S. military for years.

Next year’s DEF CON involvement promises to be even better as the Air Force plans to bring researchers out to an actual aircraft, inviting them to compromise it in every way imaginable.

Patch Tuesday

Microsoft’s monthly dump of Windows security fixes landed this week, and it was a doozy. First up are a pair of remotely exploitable Remote Desktop vulnerabilities, CVE-2019-1222 and CVE-2019-1226. It’s been theorized that these bugs were found as part of an RDP code review launched in response to the BlueKeep vulnerability from earlier this year. The important difference here is that these bugs affect multiple versions of Windows, up to and including Windows 10.

What the CTF

Remember Tavis Ormandy and his Notepad attack? We finally have the rest of the story! Go read the whole thing, it’s a great tale of finding something strange, and then pulling it apart looking for vulnerabilities.

Microsoft Windows has a module, MSCTF, that is part of the Text Services Framework. What does the CTF acronym even stand for? That’s not clear. It seems that CTF is responsible for handling keyboard layouts, and translating keystrokes based on what keyboard type is selected. What is also clear is that every time an application builds a window, that application also connects to a CTF process. CTF has been a part of Microsoft’s code base since at least 2001, with relatively few code changes since then.

CTF doesn’t do any validation, so an attacker can connect to the CTF service and claim to be any process. Tavis discovered he could effectively attempt to call arbitrary function pointers of any program talking to the same CTF service. Due to some additional security measures built into modern Windows, the path to an actual compromise is rather convoluted, but by the end of the day, any CFT client can be compromised, including notepad.

The most interesting CFT client Tavis found was the login screen. The exploit he demos as part of the write-up is to lock the computer, and then compromise the login in order to spawn a process with system privileges.

The presence of this unknown service running on every Windows machine is just another reminder that operating systems should be open source.

Biostar 2

Biostar 2 is a centralized biometric access control system in use by thousands of organizations and many countries around the globe. A pair of Israeli security researchers discovered that the central database that controls the entire system was unencrypted and unsecured. 23 Gigabytes of security data was available, including over a million fingerprints. This data was stored in the clear, rather than properly hashed, so passwords and fingerprints were directly leaked as a result. This data seems to have been made available through an Elasticsearch instance that was directly exposed to the internet, and was found through port scanning.

If you have any exposure to Biostar 2 systems, you need to assume your data has been compromised. While passwords can be changed, fingerprints are forever. As biometric authentication becomes more widespread, this is an unexplored side effect.

Overclocking In An SNES Emulator

August 14, 2019 by 12 Comments

The bsnes emulator has a new overclocking mode to eliminate slowdowns in SNES games while keeping the gameplay speed accurate. We’re emulating old SNES hardware on modern machines that are vastly more powerful. Eliminating slowdowns should be trivial, right? For an emulator such as bsnes, which is written to achieve essentially pixel-perfect accuracy when emulating, the problem is decidedly non-trivial. Stick around to learn why.

Continue reading “Overclocking In An SNES Emulator” →

Laser Trip Wire Hides What You’re (Not) Working On

August 9, 2019 by 15 Comments

We assume your office policy allows for reading Hackaday during work hours. But what about cruising reddit, or playing Universal Paperclips? There’s a special kind of stress experienced when attempting to keep one eye on your display and the other on the doorway; all the while convinced the boss is about to waltz into the room and be utterly disappointed in you.

But fear not, for [dekuNukem] has found the solution with Daytripper. This wireless laser tripwire communicates back to your computer using NRF24 (2.4 Ghz on the ISM band) and can be used to invisibly cordon off a door or hallway and fire a scripted action on your computer if its beam has been broken. Nominally this is used to send the keyboard command that hides all open windows, but we’re sure the imaginative readers of Hackaday could come up with all sorts of alternate uses for this capability.

The Daytripper transmitter uses a laser time-of-flight sensor, in this case the very small VL53L0X by STMicroelectronics. It’s best situated so the laser will be bounced straight back at it. It has a range of about four feet, which is perfect for covering a door, though a wide hallway could give it some trouble. [dekuNukem] admits that the 5 Hz scan rate means a sufficiently fast moving adversary might slip past the sensor, but if they’re trying that hard to see what’s on your monitor, they probably deserve a peek.

On the receiver side, there’s a small board that plugs into your computer and mimics a USB keyboard. It has a selector switch on the side that allows the user to set what key sequence will be “typed” once the system has been tripped. It has built-in support for minimizing all windows or locking the computer, or you can set it to send ALT + Pause, which you can listen for and act on however you see fit.

If you want to build your own Daytripper, the firmware and hardware are both available on GitHub under an MIT license. For those who prefer instant gratification, [dekuNukem] is doing a small production run and offering them up on Tindie.

This Week In Security: SWAPGS, Malicious Shaders, More IOS Woes, And WPA3

August 9, 2019 by 19 Comments

I’m sure you’ve heard of Spectre, which was the first of many speculative execution vulnerabilities found in modern processors. A new one just popped up this week. At Blackhat on Tuesday, CVE-2019-1125 was announced by Bitdefender as SWAPGS.

SWAPGS is an x86_64 instruction that is intended for use in context switching, that is when execution is transferred from a user-space program back into the kernel. Specifically, SWAPGS swaps the value of the GS register so that it refers to either a memory location in the running application, or a location in the kernel’s space. An unprivileged program can attempt to call this instruction and leak kernel memory contents as a result of the processor speculatively executing the instruction (this is similar to Spectre). Even though the instruction will ultimately not be executed, because a userspace program doesn’t have sufficient privilege to do so, the contents of the system cache have already been sufficiently altered, and an attack could feasibly leverage this to read arbitrary kernel memory.

While the initial reports have mentioned both AMD and Intel products, AMD has released a statement:

AMD is aware of new research claiming new speculative execution attacks that may allow access to privileged kernel data. Based on external and internal analysis, AMD believes it is not vulnerable to the SWAPGS variant attacks because AMD products are designed not to speculate on the new GS value following a speculative SWAPGS. For the attack that is not a SWAPGS variant, the mitigation is to implement our existing recommendations for Spectre variant 1.

Patches for Windows and Linux have been released, and Red Hat has an informative write-up on the vulnerability. I would have reviewed Bitdefender’s whitepaper on the vulnerability, but rather than make it freely available, they have opted to require a name and email address. While I would like to see their work, I refuse to sell my contact information in exchange for access.

A Malicious Shader?

This is the first time I can remember hearing of a malicious pixel shader. Cisco Talos announced a set of vulnerabilities targeting VMware and NVIDIA graphics drivers.

Shaders are specialized programs that run on a video card, and are generally used to apply effects like blur, lighting, bump mapping, and more. Most of the graphical improvements in the last few years of gaming is a result of shaders.

Talos researchers were specifically looking at how to compromise a VM Hyper-visor from inside a guest OS, and they discovered that when a host provides 3d acceleration to the guest, shaders are passed directly through to the system drivers without verification. Because the NVIDIA drivers are also vulnerable, this could allow a malicious program on the host to run arbitrary code on the hypervisor.

While this is troubling enough, the topper is that a malicious shader could potentially be run via WebGL. Taken together, this represents a real danger where simply loading a malicious WebGL enabled page could compromise not only a conventional machine, but could also compromise the bare-metal OS even when run on a guest instance.

Both NVIDIA and VMware have already released driver updates that fixes the flaw, so go update!

iOS Problems

Natalie Silvanovich of Google’s Project Zero released a set of 5 iOS vulnerabilities on Wednesday the 7th. These are not garden variety bugs, but so-called “zero click” problems where no user interaction is required for exploit.

The first exploit, for example, is a spoofed visual voicemail message. Visual voicemail notifications are sent as specially formatted text messages and contain information about the message and the address of an IMAP server to connect to and download the message. That information can be spoofed, leading a device to try to download a message from an IMAP server in the control of an attacker. From that point, finding a bug in the iOS IMAP handling code was relatively easy.

5 vulnerabilities have been fixed in iOS updates. There is a 6th vulnerability, CVE-2019-8641, that has yet to be fixed. While a few hints about this problem are given, the details have been withheld until an update has been released to fully fix the problem. One could be a bit cynical and point out that it’s the Google research team announcing these flaws. While there is certainly a self-serving angle to consider, it’s much better for iOS and consumers if flaws are fixed and publicized, rather than kept secret and sold to an offensive security vendor.

One more iOS story is Apple Bleee. Bluetooth Low Energy is an extremely useful communication protocol, allowing Apple devices to perform many of their seemingly magic functionality. The downside is that to make the magic happen, iOS devices are constantly sending BLE signals, probing for other devices. The researchers at Hexway realized that these signals leak lots of data about your device, potentially including your phone number.

iOS uses a SHA256 hash of the device’s phone number as an identifier when using AirDrop. A SHA256 is still a reasonably secure one-way hash, so there’s no problem, right? The clever realization is that while the hash is secure, and the output space is too large to attack, the input space is small enough to be manageable. An attacker could target the most common area codes in their area, limiting the target space further. From there, the SHA256 hashes for all valid numbers can be pre-calculated and stored in a lookup table.

More WPA3 Problems

We’ve discussed Dragonblood, a WPA3 analysis project. A new problem has been identified, a timing analysis attack that leaks information about the internal state of the encryption algorithm.