The latest Android and Google Play news for app and game developers.
Posted by Vikrant Nanda and René Mayrhofer, Android Security & Privacy Team
There is no better time to talk about Android dessert releases than the holidays because who doesn't love dessert? And what is one of our favorite desserts during the holiday season? Well, pie of course.
In all seriousness, pie is a great analogy because of how the various ingredients turn into multiple layers of goodness: right from the software crust on top to the hardware layer at the bottom. Read on for a summary of security and privacy features introduced in Android Pie this year.
Making Android more secure requires a combination of hardening the platform and advancing anti-exploitation techniques.
With Android Pie, we updated File-Based Encryption to support external storage media (such as, expandable storage cards). We also introduced support for metadata encryption where hardware support is present. With filesystem metadata encryption, a single key present at boot time encrypts whatever content is not encrypted by file-based encryption (such as, directory layouts, file sizes, permissions, and creation/modification times).
Android Pie also introduced a BiometricPrompt API that apps can use to provide biometric authentication dialogs (such as, fingerprint prompt) on a device in a modality-agnostic fashion. This functionality creates a standardized look, feel, and placement for the dialog. This kind of standardization gives users more confidence that they're authenticating against a trusted biometric credential checker.
New protections and test cases for the Application Sandbox help ensure all non-privileged apps targeting Android Pie (and all future releases of Android) run in stronger SELinux sandboxes. By providing per-app cryptographic authentication to the sandbox, this protection improves app separation, prevents overriding safe defaults, and (most significantly) prevents apps from making their data widely accessible.
With Android Pie, we expanded our compiler-based security mitigations, which instrument runtime operations to fail safely when undefined behavior occurs.
Control Flow Integrity (CFI) is a security mechanism that disallows changes to the original control flow graph of compiled code. In Android Pie, it has been enabled by default within the media frameworks and other security-critical components, such as for Near Field Communication (NFC) and Bluetooth protocols. We also implemented support for CFI in the Android common kernel, continuing our efforts to harden the kernel in previous Android releases.
Integer Overflow Sanitization is a security technique used to mitigate memory corruption and information disclosure vulnerabilities caused by integer operations. We've expanded our use of Integer Overflow sanitizers by enabling their use in libraries where complex untrusted input is processed or where security vulnerabilities have been reported.
One of the highlights of Android Pie is Android Protected Confirmation, the first major mobile OS API that leverages a hardware-protected user interface (Trusted UI) to perform critical transactions completely outside the main mobile operating system. Developers can use this API to display a trusted UI prompt to the user, requesting approval via a physical protected input (such as, a button on the device). The resulting cryptographically signed statement allows the relying party to reaffirm that the user would like to complete a sensitive transaction through their app.
We also introduced support for a new Keystore type that provides stronger protection for private keys by leveraging tamper-resistant hardware with dedicated CPU, RAM, and flash memory. StrongBox Keymaster is an implementation of the Keymaster hardware abstraction layer (HAL) that resides in a hardware security module. This module is designed and required to have its own processor, secure storage, True Random Number Generator (TRNG), side-channel resistance, and tamper-resistant packaging.
Other Keystore features (as part of Keymaster 4) include Keyguard-bound keys, Secure Key Import, 3DES support, and version binding. Keyguard-bound keys enable use restriction so as to protect sensitive information. Secure Key Import facilitates secure key use while protecting key material from the application or operating system. You can read more about these features in our recent blog post as well as the accompanying release notes.
User privacy has been boosted with several behavior changes, such as limiting the access background apps have to the camera, microphone, and device sensors. New permission rules and permission groups have been created for phone calls, phone state, and Wi-Fi scans, as well as restrictions around information retrieved from Wi-Fi scans. We have also added associated MAC address randomization, so that a device can use a different network address when connecting to a Wi-Fi network.
On top of that, Android Pie added support for encrypting Android backups with the user's screen lock secret (that is, PIN, pattern, or password). By design, this means that an attacker would not be able to access a user's backed-up application data without specifically knowing their passcode. Auto backup for apps has been enhanced by providing developers a way to specify conditions under which their app's data is excluded from auto backup. For example, Android Pie introduces a new flag to determine whether a user's backup is client-side encrypted.
As part of a larger effort to move all web traffic away from cleartext (unencrypted HTTP) and towards being secured with TLS (HTTPS), we changed the defaults for Network Security Configuration to block all cleartext traffic. We're protecting users with TLS by default, unless you explicitly opt-in to cleartext for specific domains. Android Pie also adds built-in support for DNS over TLS, automatically upgrading DNS queries to TLS if a network's DNS server supports it. This protects information about IP addresses visited from being sniffed or intercepted on the network level.
We believe that the features described in this post advance the security and privacy posture of Android, but you don't have to take our word for it. Year after year our continued efforts are demonstrably resulting in better protection as evidenced by increasing exploit difficulty and independent mobile security ratings. Now go and enjoy some actual pie while we get back to preparing the next Android dessert release!
Acknowledgements: This post leveraged contributions from Chad Brubaker, Janis Danisevskis, Giles Hogben, Troy Kensinger, Ivan Lozano, Vishwath Mohan, Frank Salim, Sami Tolvanen, Lilian Young, and Shawn Willden.
Posted by Lilian Young and Shawn Willden, Android Security; and Frank Salim, Google Pay
The Android Keystore provides application developers with a set of cryptographic tools that are designed to secure their users' data. Keystore moves the cryptographic primitives available in software libraries out of the Android OS and into secure hardware. Keys are protected and used only within the secure hardware to protect application secrets from various forms of attacks. Keystore gives applications the ability to specify restrictions on how and when the keys can be used.
Android Pie introduces new capabilities to Keystore. We will be discussing two of these new capabilities in this post. The first enables restrictions on key use so as to protect sensitive information. The second facilitates secure key use while protecting key material from the application or operating system.
There are times when a mobile application receives data but doesn't need to immediately access it if the user is not currently using the device. Sensitive information sent to an application while the device screen is locked must remain secure until the user wants access to it. Android Pie addresses this by introducing keyguard-bound cryptographic keys. When the screen is locked, these keys can be used in encryption or verification operations, but are unavailable for decryption or signing. If the device is currently locked with a PIN, pattern, or password, any attempt to use these keys will result in an invalid operation. Keyguard-bound keys protect the user's data while the device is locked, and only available when the user needs it.
Keyguard binding and authentication binding both function in similar ways, except with one important difference. Keyguard binding ties the availability of keys directly to the screen lock state while authentication binding uses a constant timeout. With keyguard binding, the keys become unavailable as soon as the device is locked and are only made available again when the user unlocks the device.
It is worth noting that keyguard binding is enforced by the operating system, not the secure hardware. This is because the secure hardware has no way to know when the screen is locked. Hardware-enforced Android Keystore protection features like authentication binding, can be combined with keyguard binding for a higher level of security. Furthermore, since keyguard binding is an operating system feature, it's available to any device running Android Pie.
Keys for any algorithm supported by the device can be keyguard-bound. To generate or import a key as keyguard-bound, call setUnlockedDeviceRequired(true) on the KeyGenParameterSpec or KeyProtection builder object at key generation or import.
Secure Key Import is a new feature in Android Pie that allows applications to provision existing keys into Keystore in a more secure manner. The origin of the key, a remote server that could be sitting in an on-premise data center or in the cloud, encrypts the secure key using a public wrapping key from the user's device. The encrypted key in the SecureKeyWrapper format, which also contains a description of the ways the imported key is allowed to be used, can only be decrypted in the Keystore hardware belonging to the specific device that generated the wrapping key. Keys are encrypted in transit and remain opaque to the application and operating system, meaning they're only available inside the secure hardware into which they are imported.
Secure Key Import is useful in scenarios where an application intends to share a secret key with an Android device, but wants to prevent the key from being intercepted or from leaving the device. Google Pay uses Secure Key Import to provision some keys on Pixel 3 phones, to prevent the keys from being intercepted or extracted from memory. There are also a variety of enterprise use cases such as S/MIME encryption keys being recovered from a Certificate Authorities escrow so that the same key can be used to decrypt emails on multiple devices.
To take advantage of this feature, please review this training article. Please note that Secure Key Import is a secure hardware feature, and is therefore only available on select Android Pie devices. To find out if the device supports it, applications can generate a KeyPair with PURPOSE_WRAP_KEY.
Posted by Mo Yu, Damien Octeau, and Chuangang Ren, Android Security & Privacy Team
In a previous blog post, we talked about using machine learning to combat Potentially Harmful Applications (PHAs). This blog post covers how Google uses machine learning techniques to detect and classify PHAs. We'll discuss the challenges in the PHA detection space, including the scale of data, the correct identification of PHA behaviors, and the evolution of PHA families. Next, we will introduce two of the datasets that make the training and implementation of machine learning models possible, such as app analysis data and Google Play data. Finally, we will present some of the approaches we use, including logistic regression and deep neural networks.
Detecting PHAs is challenging and requires a lot of resources. Our security experts need to understand how apps interact with the system and the user, analyze complex signals to find PHA behavior, and evolve their tactics to stay ahead of PHA authors. Every day, Google Play Protect (GPP) analyzes over half a million apps, which makes a lot of new data for our security experts to process.
Leveraging machine learning helps us detect PHAs faster and at a larger scale. We can detect more PHAs just by adding additional computing resources. In many cases, machine learning can find PHA signals in the training data without human intervention. Sometimes, those signals are different than signals found by security experts. Machine learning can take better advantage of this data, and discover hidden relationships between signals more effectively.
There are two major parts of Google Play Protect's machine learning protections: the data and the machine learning models.
The quality and quantity of the data used to create a model are crucial to the success of the system. For the purpose of PHA detection and classification, our system mainly uses two anonymous data sources: data from analyzing apps and data from how users experience apps.
Google Play Protect analyzes every app that it can find on the internet. We created a dataset by decomposing each app's APK and extracting PHA signals with deep analysis. We execute various processes on each app to find particular features and behaviors that are relevant to the PHA categories in scope (for example, SMS fraud, phishing, privilege escalation). Static analysis examines the different resources inside an APK file while dynamic analysis checks the behavior of the app when it's actually running. These two approaches complement each other. For example, dynamic analysis requires the execution of the app regardless of how obfuscated its code is (obfuscation hinders static analysis), and static analysis can help detect cloaking attempts in the code that may in practice bypass dynamic analysis-based detection. In the end, this analysis produces information about the app's characteristics, which serve as a fundamental data source for machine learning algorithms.
In addition to analyzing each app, we also try to understand how users perceive that app. User feedback (such as the number of installs, uninstalls, user ratings, and comments) collected from Google Play can help us identify problematic apps. Similarly, information about the developer (such as the certificates they use and their history of published apps) contribute valuable knowledge that can be used to identify PHAs. All these metrics are generated when developers submit a new app (or new version of an app) and by millions of Google Play users every day. This information helps us to understand the quality, behavior, and purpose of an app so that we can identify new PHA behaviors or identify similar apps.
In general, our data sources yield raw signals, which then need to be transformed into machine learning features for use by our algorithms. Some signals, such as the permissions that an app requests, have a clear semantic meaning and can be directly used. In other cases, we need to engineer our data to make new, more powerful features. For example, we can aggregate the ratings of all apps that a particular developer owns, so we can calculate a rating per developer and use it to validate future apps. We also employ several techniques to focus in on interesting data.To create compact representations for sparse data, we use embedding. To help streamline the data to make it more useful to models, we use feature selection. Depending on the target, feature selection helps us keep the most relevant signals and remove irrelevant ones.
By combining our different datasets and investing in feature engineering and feature selection, we improve the quality of the data that can be fed to various types of machine learning models.
Building a good machine learning model is like building a skyscraper: quality materials are important, but a great design is also essential. Like the materials in a skyscraper, good datasets and features are important to machine learning, but a great algorithm is essential to identify PHA behaviors effectively and efficiently.
We train models to identify PHAs that belong to a specific category, such as SMS-fraud or phishing. Such categories are quite broad and contain a large number of samples given the number of PHA families that fit the definition. Alternatively, we also have models focusing on a much smaller scale, such as a family, which is composed of a group of apps that are part of the same PHA campaign and that share similar source code and behaviors. On the one hand, having a single model to tackle an entire PHA category may be attractive in terms of simplicity but precision may be an issue as the model will have to generalize the behaviors of a large number of PHAs believed to have something in common. On the other hand, developing multiple PHA models may require additional engineering efforts, but may result in better precision at the cost of reduced scope.
We use a variety of modeling techniques to modify our machine learning approach, including supervised and unsupervised ones.
One supervised technique we use is logistic regression, which has been widely adopted in the industry. These models have a simple structure and can be trained quickly. Logistic regression models can be analyzed to understand the importance of the different PHA and app features they are built with, allowing us to improve our feature engineering process. After a few cycles of training, evaluation, and improvement, we can launch the best models in production and monitor their performance.
For more complex cases, we employ deep learning. Compared to logistic regression, deep learning is good at capturing complicated interactions between different features and extracting hidden patterns. The millions of apps in Google Play provide a rich dataset, which is advantageous to deep learning.
In addition to our targeted feature engineering efforts, we experiment with many aspects of deep neural networks. For example, a deep neural network can have multiple layers and each layer has several neurons to process signals. We can experiment with the number of layers and neurons per layer to change model behaviors.
We also adopt unsupervised machine learning methods. Many PHAs use similar abuse techniques and tricks, so they look almost identical to each other. An unsupervised approach helps define clusters of apps that look or behave similarly, which allows us to mitigate and identify PHAs more effectively. We can automate the process of categorizing that type of app if we are confident in the model or can request help from a human expert to validate what the model found.
PHAs are constantly evolving, so our models need constant updating and monitoring. In production, models are fed with data from recent apps, which help them stay relevant. However, new abuse techniques and behaviors need to be continuously detected and fed into our machine learning models to be able to catch new PHAs and stay on top of recent trends. This is a continuous cycle of model creation and updating that also requires tuning to ensure that the precision and coverage of the system as a whole matches our detection goals.
As part of Google's AI-first strategy, our work leverages many machine learning resources across the company, such as tools and infrastructures developed by Google Brain and Google Research. In 2017, our machine learning models successfully detected 60.3% of PHAs identified by Google Play Protect, covering over 2 billion Android devices. We continue to research and invest in machine learning to scale and simplify the detection of PHAs in the Android ecosystem.
This work was developed in joint collaboration with Google Play Protect, Safe Browsing and Play Abuse teams with contributions from Andrew Ahn, Hrishikesh Aradhye, Daniel Bali, Hongji Bao, Yajie Hu, Arthur Kaiser, Elena Kovakina, Salvador Mandujano, Melinda Miller, Rahul Mishra, Sebastian Porst, Monirul Sharif, Sri Somanchi, Sai Deep Tetali, and Zhikun Wang.
Posted by Janis Danisevskis, Information Security Engineer, Android Security
In Android Pie, we introduced Android Protected Confirmation, the first major mobile OS API that leverages a hardware protected user interface (Trusted UI) to perform critical transactions completely outside the main mobile operating system. This Trusted UI protects the choices you make from fraudulent apps or a compromised operating system. When an app invokes Protected Confirmation, control is passed to the Trusted UI, where transaction data is displayed and user confirmation of that data's correctness is obtained.
Once confirmed, your intention is cryptographically authenticated and unforgeable when conveyed to the relying party, for example, your bank. Protected Confirmation increases the bank's confidence that it acts on your behalf, providing a higher level of protection for the transaction.
Protected Confirmation also adds additional security relative to other forms of secondary authentication, such as a One Time Password or Transaction Authentication Number. These mechanisms can be frustrating for mobile users and also fail to protect against a compromised device that can corrupt transaction data or intercept one-time confirmation text messages.
Once the user approves a transaction, Protected Confirmation digitally signs the confirmation message. Because the signing key never leaves the Trusted UI's hardware sandbox, neither app malware nor a compromised operating system can fool the user into authorizing anything. Protected Confirmation signing keys are created using Android's standard AndroidKeyStore API. Before it can start using Android Protected Confirmation for end-to-end secure transactions, the app must enroll the public KeyStore key and its Keystore Attestation certificate with the remote relying party. The attestation certificate certifies that the key can only be used to sign Protected Confirmations.
There are many possible use cases for Android Protected Confirmation. At Google I/O 2018, the What's new in Android security session showcased partners planning to leverage Android Protected Confirmation in a variety of ways, including Royal Bank of Canada person to person money transfers; Duo Security, Nok Nok Labs, and ProxToMe for user authentication; and Insulet Corporation and Bigfoot Biomedical, for medical device control.
Insulet, a global leading manufacturer of tubeless patch insulin pumps, has demonstrated how they can modify their FDA cleared Omnipod DASH TM Insulin management system in a test environment to leverage Protected Confirmation to confirm the amount of insulin to be injected. This technology holds the promise for improved quality of life and reduced cost by enabling a person with diabetes to leverage their convenient, familiar, and secure smartphone for control rather than having to rely on a secondary, obtrusive, and expensive remote control device. (Note: The Omnipod DASH™ System is not cleared for use with Pixel 3 mobile device or Protected Confirmation).
This work is fulfilling an important need in the industry. Since smartphones do not fit the mold of an FDA approved medical device, we've been working with FDA as part of DTMoSt, an industry-wide consortium, to define a standard for phones to safely control medical devices, such as insulin pumps. A technology like Protected Confirmation plays an important role in gaining higher assurance of user intent and medical safety.
To integrate Protected Confirmation into your app, check out the Android Protected Confirmation training article. Android Protected Confirmation is an optional feature in Android Pie. Because it has low-level hardware dependencies, Protected Confirmation may not be supported by all devices running Android Pie. Google Pixel 3 and 3XL devices are the first to support Protected Confirmation, and we are working closely with other manufacturers to adopt this market-leading security innovation on more devices.
Posted by Nagendra Modadugu and Bill Richardson, Google Device Security Group
At the Made by Google event last week, we talked about the combination of AI + Software + Hardware to help organize your information. To better protect that information at a hardware level, our new Pixel 3 and Pixel 3 XL devices include a Titan M chip.We briefly introduced Titan M and some of its benefits on our Keyword Blog, and with this post we dive into some of its technical details.
Titan M is a second-generation, low-power security module designed and manufactured by Google, and is a part of the Titan family. As described in the Keyword Blog post, Titan M performs several security sensitive functions, including:
Including Titan M in Pixel 3 devices substantially reduces the attack surface. Because Titan M is a separate chip, the physical isolation mitigates against entire classes of hardware-level exploits such as Rowhammer, Spectre, and Meltdown. Titan M's processor, caches, memory, and persistent storage are not shared with the rest of the phone's system, so side channel attacks like these—which rely on subtle, unplanned interactions between internal circuits of a single component—are nearly impossible. In addition to its physical isolation, the Titan M chip contains many defenses to protect against external attacks.
But Titan M is not just a hardened security microcontroller, but rather a full-lifecycle approach to security with Pixel devices in mind. Titan M's security takes into consideration all the features visible to Android down to the lowest level physical and electrical circuit design and extends beyond each physical device to our supply chain and manufacturing processes. At the physical level, we incorporated essential features optimized for the mobile experience: low power usage, low-latency, hardware crypto acceleration, tamper detection, and secure, timely firmware updates. We improved and invested in the supply chain for Titan M by creating a custom provisioning process, which provides us with transparency and control starting from the earliest silicon stages.
Finally, in the interest of transparency, the Titan M firmware source code will be publicly available soon. While Google holds the root keys necessary to sign Titan M firmware, it will be possible to reproduce binary builds based on the public source for the purpose of binary transparency.
Titan (left) and Titan M (right)
Titan M's CPU is an ARM Cortex-M3 microprocessor specially hardened against side-channel attacks and augmented with defensive features to detect and respond to abnormal conditions. The Titan M CPU core also exposes several control registers, which can be used to taper access to chip configuration settings and peripherals. Once powered on, Titan M verifies the signature of its flash-based firmware using a public key built into the chip's silicon. If the signature is valid, the flash is locked so it can't be modified, and then the firmware begins executing.
Titan M also features several hardware accelerators: AES, SHA, and a programmable big number coprocessor for public key algorithms. These accelerators are flexible and can either be initialized with keys provided by firmware or with chip-specific and hardware-bound keys generated by the Key Manager module. Chip-specific keys are generated internally based on entropy derived from the True Random Number Generator (TRNG), and thus such keys are never externally available outside the chip over its entire lifetime.
While implementing Titan M firmware, we had to take many system constraints into consideration. For example, packing as many security features into Titan M's 64 Kbytes of RAM required all firmware to execute exclusively off the stack. And to reduce flash-wear, RAM contents can be preserved even during low-power mode when most hardware modules are turned off.
The diagram below provides a high-level view of the chip components described here.
At the heart of our implementation of Titan M are two broader trends: transparency and building a platform for future innovation.
Transparency around every step of the design process — from logic gates to boot code to the applications — gives us confidence in the defenses we're providing for our users. We know what's inside, how it got there, how it works, and who can make changes.
Custom hardware allows us to provide new features, capabilities, and performance not readily available in off-the-shelf components. These changes allow higher assurance use cases like two-factor authentication, medical device control, P2P payments, and others that we will help develop down the road.
As more of our lives are bound up in our phones, keeping those phones secure and trustworthy is increasingly important. Google takes that responsibility seriously. Titan M is just the latest step in our continuing efforts to improve the privacy and security of all our users.
Posted by Sami Tolvanen, Staff Software Engineer, Android Security
Android's security model is enforced by the Linux kernel, which makes it a tempting target for attackers. We have put a lot of effort into hardening the kernel in previous Android releases and in Android 9, we continued this work by focusing on compiler-based security mitigations against code reuse attacks.
Google's Pixel 3 will be the first Android device to ship with LLVM's forward-edge Control Flow Integrity (CFI) enforcement in the kernel, and we have made CFI support available in Android kernel versions 4.9 and 4.14. This post describes how kernel CFI works and provides solutions to the most common issues developers might run into when enabling the feature.
A common method of exploiting the kernel is using a bug to overwrite a function pointer stored in memory, such as a stored callback pointer or a return address that had been pushed to the stack. This allows an attacker to execute arbitrary parts of the kernel code to complete their exploit, even if they cannot inject executable code of their own. This method of gaining code execution is particularly popular with the kernel because of the huge number of function pointers it uses, and the existing memory protections that make code injection more challenging.
CFI attempts to mitigate these attacks by adding additional checks to confirm that the kernel's control flow stays within a precomputed graph. This doesn't prevent an attacker from changing a function pointer if a bug provides write access to one, but it significantly restricts the valid call targets, which makes exploiting such a bug more difficult in practice.
Figure 1. In an Android device kernel, LLVM's CFI limits 55% of indirect calls to at most 5 possible targets and 80% to at most 20 targets.
In order to determine all valid call targets for each indirect branch, the compiler needs to see all of the kernel code at once. Traditionally, compilers work on a single compilation unit (source file) at a time and leave merging the object files to the linker. LLVM's solution to CFI is to require the use of LTO, where the compiler produces LLVM-specific bitcode for all C compilation units, and an LTO-aware linker uses the LLVM back-end to combine the bitcode and compile it into native code.
Figure 2. A simplified overview of how LTO works in the kernel. All LLVM bitcode is combined, optimized, and generated into native code at link time.
Linux has used the GNU toolchain for assembling, compiling, and linking the kernel for decades. While we continue to use the GNU assembler for stand-alone assembly code, LTO requires us to switch to LLVM's integrated assembler for inline assembly, and either GNU gold or LLVM's own lld as the linker. Switching to a relatively untested toolchain on a huge software project will lead to compatibility issues, which we have addressed in our arm64 LTO patch sets for kernel versions 4.9 and 4.14.
In addition to making CFI possible, LTO also produces faster code due to global optimizations. However, additional optimizations often result in a larger binary size, which may be undesirable on devices with very limited resources. Disabling LTO-specific optimizations, such as global inlining and loop unrolling, can reduce binary size by sacrificing some of the performance gains. When using GNU gold, the aforementioned optimizations can be disabled with the following additions to LDFLAGS:
LDFLAGS += -plugin-opt=-inline-threshold=0 \ -plugin-opt=-unroll-threshold=0
Note that flags to disable individual optimizations are not part of the stable LLVM interface and may change in future compiler versions.
LLVM's CFI implementation adds a check before each indirect branch to confirm that the target address points to a valid function with a correct signature. This prevents an indirect branch from jumping to an arbitrary code location and even limits the functions that can be called. As C compilers do not enforce similar restrictions on indirect branches, there were several CFI violations due to function type declaration mismatches even in the core kernel that we have addressed in our CFI patch sets for kernels 4.9 and 4.14.
Kernel modules add another complication to CFI, as they are loaded at runtime and can be compiled independently from the rest of the kernel. In order to support loadable modules, we have implemented LLVM's cross-DSO CFI support in the kernel, including a CFI shadow that speeds up cross-module look-ups. When compiled with cross-DSO support, each kernel module contains information about valid local branch targets, and the kernel looks up information from the correct module based on the target address and the modules' memory layout.
Figure 3. An example of a cross-DSO CFI check injected into an arm64 kernel. Type information is passed in X0 and the target address to validate in X1.
CFI checks naturally add some overhead to indirect branches, but due to more aggressive optimizations, our tests show that the impact is minimal, and overall system performance even improved 1-2% in many cases.
CFI for arm64 requires clang version >= 5.0 and binutils >= 2.27. The kernel build system also assumes that the LLVMgold.so plug-in is available in LD_LIBRARY_PATH. Pre-built toolchain binaries for clang and binutils are available in AOSP, but upstream binaries can also be used.
The following kernel configuration options are needed to enable kernel CFI:
CONFIG_LTO_CLANG=y CONFIG_CFI_CLANG=y
Using CONFIG_CFI_PERMISSIVE=y may also prove helpful when debugging a CFI violation or during device bring-up. This option turns a violation into a warning instead of a kernel panic.
As mentioned in the previous section, the most common issue we ran into when enabling CFI on Pixel 3 were benign violations caused by function pointer type mismatches. When the kernel runs into such a violation, it prints out a runtime warning that contains the call stack at the time of the failure, and the call target that failed the CFI check. Changing the code to use a correct function pointer type fixes the issue. While we have fixed all known indirect branch type mismatches in the Android kernel, similar problems may be still found in device specific drivers, for example.
CFI failure (target: [<fffffff3e83d4d80>] my_target_function+0x0/0xd80): ------------[ cut here ]------------ kernel BUG at kernel/cfi.c:32! Internal error: Oops - BUG: 0 [#1] PREEMPT SMP … Call trace: … [<ffffff8752d00084>] handle_cfi_failure+0x20/0x28 [<ffffff8752d00268>] my_buggy_function+0x0/0x10 …
Figure 4. An example of a kernel panic caused by a CFI failure.
Another potential pitfall are address space conflicts, but this should be less common in driver code. LLVM's CFI checks only understand kernel virtual addresses and any code that runs at another exception level or makes an indirect call to a physical address will result in a CFI violation. These types of failures can be addressed by disabling CFI for a single function using the __nocfi attribute, or even disabling CFI for entire code files using the $(DISABLE_CFI) compiler flag in the Makefile.
static int __nocfi address_space_conflict() { void (*fn)(void); … /* branching to a physical address trips CFI w/o __nocfi */ fn = (void *)__pa_symbol(function_name); cpu_install_idmap(); fn(); cpu_uninstall_idmap(); … }
Figure 5. An example of fixing a CFI failure caused by an address space conflict.
Finally, like many hardening features, CFI can also be tripped by memory corruption errors that might otherwise result in random kernel crashes at a later time. These may be more difficult to debug, but memory debugging tools such as KASAN can help here.
We have implemented support for LLVM's CFI in Android kernels 4.9 and 4.14. Google's Pixel 3 will be the first Android device to ship with these protections, and we have made the feature available to all device vendors through the Android common kernel. If you are shipping a new arm64 device running Android 9, we strongly recommend enabling kernel CFI to help protect against kernel vulnerabilities.
LLVM's CFI protects indirect branches against attackers who manage to gain access to a function pointer stored in kernel memory. This makes a common method of exploiting the kernel more difficult. Our future work involves also protecting function return addresses from similar attacks using LLVM's Shadow Call Stack, which will be available in an upcoming compiler release.
Posted by Jason Woloz and Mayank Jain, Android Security & Privacy Team
Our Android and Play security reward programs help us work with top researchers from around the world to improve Android ecosystem security every day. Thank you to all the amazing researchers who submitted vulnerability reports.
In the ASR program's third year, we received over 470 qualifying vulnerability reports from researchers and the average pay per researcher jumped by 23%. To date, the ASR program has rewarded researchers with over $3M, paying out roughly $1M per year.
Here are some of the highlights from the Android Security Rewards program's third year:
As part of our ongoing commitment to security we regularly update our programs and policies based on ecosystem feedback. We also updated our severity guidelines for evaluating the impact of reported security vulnerabilities against the Android platform.
In October 2017, we rolled out the Google Play Security Reward Program to encourage security research into popular Android apps available on Google Play. So far, researchers have reported over 30 vulnerabilities through the program, earning a combined bounty amount of over $100K.
If undetected, these vulnerabilities could have potentially led to elevation of privilege, access to sensitive data and remote code execution on devices.
In addition to rewarding for vulnerabilities, we continue to work with the broad and diverse Android ecosystem to protect users from issues reported through our program. We collaborate with manufacturers to ensure that these issues are fixed on their devices through monthly security updates. Over 250 device models have a majority of their deployed devices running a security update from the last 90 days. This table shows the models with a majority of deployed devices running a security update from the last three months:
Thank you to everyone internally and externally who helped make Android safer and stronger in the past year. Together, we made a huge investment in security research that helps Android users everywhere. If you want to get involved to make next year even better, check out our detailed program rules. For tips on how to submit complete reports, see Bug Hunter University.
Posted by Dave Kleidermacher, VP, Head of Security - Android, Chrome OS, Play
At Google I/O 2018, in our What's New in Android Security session, we shared a brief update on the Android security updates program. With the official release of Android 9 Pie, we wanted to share a more comprehensive update on the state of security updates, including best practice guidance for manufacturers, how we're making Android easier to update, and how we're ensuring compliance to Android security update releases.
As we noted in our 2017 Android Security Year-in-Review, Android's anti-exploitation strength now leads the mobile industry and has made it exceedingly difficult and expensive to leverage operating system bugs into compromises. Nevertheless, an important defense-in-depth strategy is to ensure critical security updates are delivered in a timely manner. Monthly security updates are the recommended best practice for Android smartphones. We deliver monthly Android source code patches to smartphone manufacturers so they may incorporate those patches into firmware updates. We also deliver firmware updates over-the-air to Pixel devices on a reliable monthly cadence and offer the free use of Google's firmware over-the-air (FOTA) servers to manufacturers. Monthly security updates are also required for devices covered under the Android One program.
While monthly security updates are best, at minimum, Android manufacturers should deliver regular security updates in advance of coordinated disclosure of high severity vulnerabilities, published in our Android bulletins. Since the common vulnerability disclosure window is 90 days, updates on a 90-day frequency represents a minimum security hygiene requirement.
Product security factors into purchase decisions of enterprises, who often consider device security update cadence, flexibility of policy controls, and authentication features. Earlier this year, we introduced the Android Enterprise Recommended program to help businesses make these decisions. To be listed, Android devices must satisfy numerous requirements, including regular security updates: at least every 90 days, with monthly updates strongly recommended. In addition to businesses, consumers interested in understanding security update practices and commitment may also refer to the Enterprise Recommended list.
We've also been working to make Android easier to update, overall. A key pillar of that strategy is to improve modularity and clarity of interfaces, enabling operating system subsystems to be updated without adversely impacting others. Project Treble is one example of this strategy in action and has enabled devices to update to Android P more easily and efficiently than was possible in previous releases. The modularity strategy applies equally well for security updates, as a framework security update can be performed independently of device specific components.
Another part of the strategy involves the extraction of operating system services into user-mode applications that can be updated independently, and sometimes more rapidly, than the base operating system. For example, Google Play services, including secure networking components, and the Chrome browser can be updated individually, just like other Google Play apps.
Partner programs are a third key pillar of the updateability strategy. One example is the GMS Express program, in which Google is working closely with system-on-chip (SoC) suppliers to provide monthly pre-integrated and pre-tested Android security updates for SoC reference designs, reducing cost and time to market for delivering them to users.
Recently, researchers reported a handful of missing security bug fixes across some Android devices. Initial reports had several inaccuracies, which have since been corrected. We have been developing security update testing systems that are now making compliance failures less likely to occur. In particular, we recently delivered a new testing infrastructure that enables manufacturers to develop and deploy automated tests across lower levels of the firmware stack that were previously relegated to manual testing. In addition, the Android build approval process now includes scanning of device images for specific patterns, reducing the risk of omission.
In 2017, about a billion Android devices received security updates, representing approximately 30% growth over the preceding year. We continue to work hard devising thoughtful strategies to make Android easier to update by introducing improved processes and programs for the ecosystem. In addition, we are also working to drive increased and more expedient partner adoption of our security update and compliance requirements. As a result, over coming quarters, we expect the largest ever growth in the number of Android devices receiving regular security updates.
Bugs are inevitable in all complex software systems, but exploitability of those bugs is not. We're working hard to ensure that the incidence of potentially harmful exploitation of bugs continues to decline, such that the frequency for security updates will reduce, not increase, over time. While monthly security updates represents today's best practice, we see a future in which security updates becomes easier and rarer, while maintaining the same goal to protect all users across all devices.
Posted by Purnima Kochikar, Director, Google Play, Apps & Games
On Monday we released Android 9 Pie. As we continue to push the Android platform forward, we're always looking to provide new ways to distribute your apps efficiently, help people discover and engage with your work, and improve the overall security of our ecosystem. Google Play has had a busy year so far with some big milestones around helping you reach more users, including:
Google Play is dedicated to helping you build and grow quality app businesses, reach the more than 2 billion Android devices globally and provide your users with better experiences. Here are some of the important areas we're prioritizing this year:
We've added more testing tools to the popular Play Console to help developers de-risk app launches with internal and external test tracks and staged rollouts to get valuable early feedback. This year we've expanded the Start on Android program globally that provides developers new to Android additional guidance to optimize their apps before launch. Google Play Instant remains a huge bet to transform app discovery and improve conversions by letting users engage without the friction of installing. We're seeing great results from early adopters and are working on new places to surface instant experience, including ads, and making them easier to build throughout the year.
Google Play plays an important role helping developers understand and fix quality and performance issues. At I/O, we showcased how we expanded the battery, stability and rendering of Android vitals reporting to include app start time & permission denials, enabling developers to cut application not responding errors by up to 95%. We also expanded the functionality of automated device testing with the pre-launch report to enable games testing. Recently, we increased the importance of app quality in our search and discovery recommendations that has resulted in higher engagement and satisfaction with downloaded games.
Over the last year we've rolled out more editorial content and improved our machine learning to deliver personalized recommendations for apps and games that engage users. Since most game downloads come from browsing (as opposed to searching or deep linking into) the store, we've put particular focus on games discovery, with a new games home page, special sections for premium and new games, immersive video trailers and screenshots, and the ability to try games instantly. We've also introduced new programs to help drive app downloads through richer discovery. For example, since launching our app pre-registration program in 2016, we've seen nearly 250 million app pre-registrations. Going forward, we'll be expanding on these programs and others like LiveOps cards to help developers engage more deeply with their audience.
Google Play now collects payments in 150 markets via credit card, direct carrier billing (DCB), Paypal, and gift cards. Direct carrier billing is now enabled across 167 carriers in 64 markets. In 2018, we have focused on expanding our footprint in Africa and Latam with launches in Ghana, Kenya, Tanzania, Nigeria, Peru & Colombia. And users can now buy Google Play credit via gift cards or other means in more 800,000 retail locations around the world. This year, we also launched seller support in 18 new markets bringing the total markets with seller support to 98. Our subscription offering continues to improve with ML-powered fraud detection and even more control for subscribers and developers. Google Play's risk modeling automatically helps detect fraudulent transactions and purchase APIs help you better analyze your refund data to identify suspicious activity.
Google Play Protect and our other systems scan and analyze more than 50 billion apps a day to keep our ecosystem safe for users and developers. In fact, people who only download apps from Google Play are nine times less likely to download a potentially harmful app than those who download from other sources. We've made significant improvements in our ability to detect abuse—such as impersonation, inappropriate content, fraud, or malware—through new machine learning models and techniques. The result is that 99% of apps with abusive content are identified and rejected before anyone can install them. We're also continuing to run the Google Play Security Rewards Program through a collaboration with Hacker One to discover other vulnerabilities.
We are continually inspired by what developers build—check out #IMakeApps for incredible examples—and want every developer to have the tools needed to succeed. We can't wait to see what you do next!
Android's switch to LLVM/Clang as the default platform compiler in Android 7.0 opened up more possibilities for improving our defense-in-depth security posture. In the past couple of releases, we've rolled out additional compiler-based mitigations to make bugs harder to exploit and prevent certain types of bugs from becoming vulnerabilities. In Android P, we're expanding our existing compiler mitigations, which instrument runtime operations to fail safely when undefined behavior occurs. This post describes the new build system support for Control Flow Integrity and Integer Overflow Sanitization.
A key step in modern exploit chains is for an attacker to gain control of a program's control flow by corrupting function pointers or return addresses. This opens the door to code-reuse attacks where an attacker executes arbitrary portions of existing program code to achieve their goals, such as counterfeit-object-oriented and return-oriented programming. Control Flow Integrity (CFI) describes a set of mitigation technologies that confine a program's control flow to a call graph of valid targets determined at compile-time.
While we first supported LLVM's CFI implementation in select components in Android O, we're greatly expanding that support in P. This implementation focuses on preventing control flow manipulation via indirect branches, such as function pointers and virtual functions—the 'forward-edges' of a call graph. Valid branch targets are defined as function entry points for functions with the expected function signature, which drastically reduces the set of allowable destinations an attacker can call. Indirect branches are instrumented to detect runtime violations of the statically determined set of allowable targets. If a violation is detected because a branch points to an unexpected target, then the process safely aborts.
Figure 1. Assembly-level comparison of a virtual function call with and without CFI enabled.
For example, Figure 1 illustrates how a function that takes an object and calls a virtual function gets translated into assembly with and without CFI. For simplicity, this was compiled with -O0 to prevent compiler optimization. Without CFI enabled, it loads the object's vtable pointer and calls the function at the expected offset. With CFI enabled, it performs a fast-path first check to determine if the pointer falls within an expected range of addresses of compatible vtables. Failing that, execution falls through to a slow path that does a more extensive check for valid classes that are defined in other shared libraries. The slow path will abort execution if the vtable pointer points to an invalid target.
With control flow tightly restricted to a small set of legitimate targets, code-reuse attacks become harder to utilize and some memory corruption vulnerabilities become more difficult or even impossible to exploit.
In terms of performance impact, LLVM's CFI requires compiling with Link-Time Optimization (LTO). LTO preserves the LLVM bitcode representation of object files until link-time, which allows the compiler to better reason about what optimizations can be performed. Enabling LTO reduces the size of the final binary and improves performance, but increases compile time. In testing on Android, the combination of LTO and CFI results in negligible overhead to code size and performance; in a few cases both improved.
For more technical details about CFI and how other forward-control checks are handled, see the LLVM design documentation.
For Android P, CFI is enabled by default widely within the media frameworks and other security-critical components, such as NFC and Bluetooth. CFI kernel support has also been introduced into the Android common kernel when building with LLVM, providing the option to further harden the trusted computing base. This can be tested today on the HiKey reference boards.
The UndefinedBehaviorSanitizer's (UBSan) signed and unsigned integer overflow sanitization was first utilized when hardening the media stack in Android Nougat. This sanitization is designed to safely abort process execution if a signed or unsigned integer overflows by instrumenting arithmetic instructions which may overflow. The end result is the mitigation of an entire class of memory corruption and information disclosure vulnerabilities where the root cause is an integer overflow, such as the original Stagefright vulnerability.
Because of their success, we've expanded usage of these sanitizers in the media framework with each release. Improvements have been made to LLVM's integer overflow sanitizers to reduce the performance impact by using fewer instructions in ARM 32-bit and removing unnecessary checks. In testing, these improvements reduced the sanitizers' performance overhead by over 75% in Android's 32-bit libstagefright library for some codecs. Improved Android build system support, such as better diagnostics support, more sensible crashes, and globally sanitized integer overflow targets for testing have also expedited the rollout of these sanitizers.
We've prioritized enabling integer overflow sanitization in libraries where complex untrusted input is processed or where there have been security bulletin-level integer overflow vulnerabilities reported. As a result, in Android P the following libraries now benefit from this mitigation:
Moving forward, we're expanding our use of these mitigation technologies and we strongly encourage vendors to do the same with their customizations. More information about how to enable and test these options will be available soon on the Android Open Source Project.
Posted by Vishwath Mohan, Security Engineer
To keep users safe, most apps and devices have an authentication mechanism, or a way to prove that you're you. These mechanisms fall into three categories: knowledge factors, possession factors, and biometric factors. Knowledge factors ask for something you know (like a PIN or a password), possession factors ask for something you have (like a token generator or security key), and biometric factors ask for something you are (like your fingerprint, iris, or face).
Biometric authentication mechanisms are becoming increasingly popular, and it's easy to see why. They're faster than typing a password, easier than carrying around a separate security key, and they prevent one of the most common pitfalls of knowledge-factor based authentication—the risk of shoulder surfing.
As more devices incorporate biometric authentication to safeguard people's private information, we're improving biometrics-based authentication in Android P by:
Currently, biometric unlocks quantify their performance today with two metrics borrowed from machine learning (ML): False Accept Rate (FAR), and False Reject Rate (FRR).
In the case of biometrics, FAR measures how often a biometric model accidentally classifies an incorrect input as belonging to the target user—that is, how often another user is falsely recognized as the legitimate device owner. Similarly, FRR measures how often a biometric model accidentally classifies the user's biometric as incorrect—that is, how often a legitimate device owner has to retry their authentication. The first is a security concern, while the second is problematic for usability.
Both metrics do a great job of measuring the accuracy and precision of a given ML (or biometric) model when applied to random input samples. However, because neither metric accounts for an active attacker as part of the threat model, they do not provide very useful information about its resilience against attacks.
In Android 8.1, we introduced two new metrics that more explicitly account for an attacker in the threat model: Spoof Accept Rate (SAR) and Imposter Accept Rate (IAR). As their names suggest, these metrics measure how easily an attacker can bypass a biometric authentication scheme. Spoofing refers to the use of a known-good recording (e.g. replaying a voice recording or using a face or fingerprint picture), while impostor acceptance means a successful mimicking of another user's biometric (e.g. trying to sound or look like a target user).
We use the SAR/IAR metrics to categorize biometric authentication mechanisms as either strong or weak. Biometric authentication mechanisms with an SAR/IAR of 7% or lower are strong, and anything above 7% is weak. Why 7% specifically? Most fingerprint implementations have a SAR/IAR metric of about 7%, making this an appropriate standard to start with for other modalities as well. As biometric sensors and classification methods improve, this threshold can potentially be decreased in the future.
This binary classification is a slight oversimplification of the range of security that different implementations provide. However, it gives us a scalable mechanism (via the tiered authentication model) to appropriately scope the capabilities and the constraints of different biometric implementations across the ecosystem, based on the overall risk they pose.
While both strong and weak biometrics will be allowed to unlock a device, weak biometrics:
These measures are intended to allow weaker biometrics, while reducing the risk of unauthorized access.
Starting in Android P, developers can use the BiometricPrompt API to integrate biometric authentication into their apps in a device and biometric agnostic way. BiometricPrompt only exposes strong modalities, so developers can be assured of a consistent level of security across all devices their application runs on. A support library is also provided for devices running Android O and earlier, allowing applications to utilize the advantages of this API across more devices.
While applications can hook into BiometricPrompt directly for Android 9 and higher, developers should use the BiometricPrompt library to support the widest range of devices.
The API is intended to be easy to use, allowing the platform to select an appropriate biometric to authenticate with instead of forcing app developers to implement this logic themselves. Here's an example of how a developer might use it in their app:
Biometrics have the potential to both simplify and strengthen how we authenticate our digital identity, but only if they are designed securely, measured accurately, and implemented in a privacy-preserving manner.
We want Android to get it right across all three. So we're combining secure design principles, a more attacker-aware measurement methodology, and a common, easy to use biometrics API that allows developers to integrate authentication in a simple, consistent, and safe manner.
Acknowledgements: This post was developed in joint collaboration with Jim Miller.
Posted by Giles Hogben, Privacy Engineer and Milinda Perera, Software Engineer
Developers already use HTTPS to communicate with Firebase Cloud Messaging (FCM). The channel between FCM server endpoint and the device is encrypted with SSL over TCP. However, messages are not encrypted end-to-end (E2E) between the developer server and the user device unless developers take special measures.
To this end, we advise developers to use keys generated on the user device to encrypt push messages end-to-end. But implementing such E2E encryption has historically required significant technical knowledge and effort. That is why we are excited to announce the Capillary open source library which greatly simplifies the implementation of E2E-encryption for push messages between developer servers and users' Android devices.
We also added functionality for sending messages that can only be decrypted on devices that have recently been unlocked. This is designed to support decrypting messages on devices using File-Based Encryption (FBE): encrypted messages are cached in Device Encrypted (DE) storage and message decryption keys are stored in Android Keystore, requiring user authentication. This allows developers to specify messages with sensitive content, that remain encrypted in cached form until the user has unlocked and decrypted their device.
The library handles:
The library supports both RSA encryption with ECDSA authentication and Web Push encryption, allowing developers to re-use existing server-side code developed for sending E2E-encrypted Web Push messages to browser-based clients.
Along with the library, we are also publishing a demo app (at last, the Google privacy team has its own messaging app!) that uses the library to send E2E-encrypted FCM payloads from a gRPC-based server implementation.
You can find more technical details describing how we've architected and implemented the library and demo here.
Posted by Shawn Willden, Staff Software Engineer
Our smart devices, such as mobile phones and tablets, contain a wealth of personal information that needs to be kept safe. Google is constantly trying to find new and better ways to protect that valuable information on Android devices. From partnering with external researchers to find and fix vulnerabilities, to adding new features to the Android platform, we work to make each release and new device safer than the last. This post talks about Google's strategy for making the encryption on Google Pixel 2 devices resistant to various levels of attack—from platform, to hardware, all the way to the people who create the signing keys for Pixel devices.
We encrypt all user data on Google Pixel devices and protect the encryption keys in secure hardware. The secure hardware runs highly secure firmware that is responsible for checking the user's password. If the password is entered incorrectly, the firmware refuses to decrypt the device. This firmware also limits the rate at which passwords can be checked, making it harder for attackers to use a brute force attack.
To prevent attackers from replacing our firmware with a malicious version, we apply digital signatures. There are two ways for an attacker to defeat the signature checks and install a malicious replacement for firmware: find and exploit vulnerabilities in the signature-checking process or gain access to the signing key and get their malicious version signed so the device will accept it as a legitimate update. The signature-checking software is tiny, isolated, and vetted with extreme thoroughness. Defeating it is hard. The signing keys, however, must exist somewhere, and there must be people who have access to them.
In the past, device makers have focused on safeguarding these keys by storing the keys in secure locations and severely restricting the number of people who have access to them. That's good, but it leaves those people open to attack by coercion or social engineering. That's risky for the employees personally, and we believe it creates too much risk for user data.
To mitigate these risks, Google Pixel 2 devices implement insider attack resistance in the tamper-resistant hardware security module that guards the encryption keys for user data. This helps prevent an attacker who manages to produce properly signed malicious firmware from installing it on the security module in a lost or stolen device without the user's cooperation. Specifically, it is not possible to upgrade the firmware that checks the user's password unless you present the correct user password. There is a way to "force" an upgrade, for example when a returned device is refurbished for resale, but forcing it wipes the secrets used to decrypt the user's data, effectively destroying it.
The Android security team believes that insider attack resistance is an important element of a complete strategy for protecting user data. The Google Pixel 2 demonstrated that it's possible to protect users even against the most highly-privileged insiders. We recommend that all mobile device makers do the same. For help, device makers working to implement insider attack resistance can reach out to the Android security team through their Google contact.
Acknowledgements: This post was developed in joint collaboration with Paul Crowley, Senior Software Engineer
Play Console
