This page contains information about the features of Keystore in Android 6.0.
Keystore provides the following categories of operations:
- Key generation
- Import and export of asymmetric keys (no key wrapping)
- Import of raw symmetric keys (again, no wrapping)
- Asymmetric encryption and decryption with appropriate padding modes
- Asymmetric signing and verification with digesting and appropriate padding modes
- Symmetric encryption and decryption in appropriate modes, including an AEAD mode
- Generation and verification of symmetric message authentication codes
Protocol elements, such as purpose, mode and padding, as well as access control constraints, must be specified when keys are generated or imported and are permanently bound to the key, ensuring the key cannot be used in any other way.
In addition to the list above, there is one more service that Keymaster implementations must provide but which is not exposed as an API: Random number generation. This is used internally for generation of keys, Initialization Vectors (IVs), random padding and other elements of secure protocols that require randomness.
All implementations must provide:
- 2048, 3072 and 4096-bit key support are required
- Support for public exponent F4 (2^16+1)
- Required padding modes for RSA signing are:
- No padding (deprecated, will be removed in the future)
- RSASSA-PSS (
- RSASSA-PKCS1-v1_5 (
- Required digest modes for RSA signing are:
- No digest (deprecated, will be removed in the future)
- Required padding modes for RSA encryption/decryption are:
- RSAES-OAEP (
- RSAES-PKCS1-v1_5 (
- 224, 256, 384 and 521-bit key support are required, using the NIST P-224, P-256, P-384 and P-521 curves, respectively
- Required digest modes for ECDSA are:
- No digest (deprecated, will be removed in the future)
- 128 and 256-bit keys are required
- CBC, CTR, ECB and and GCM. The GCM implementation must not allow the use of tags smaller than 96 bits or nonce lengths other than 96 bits.
- Padding modes
KM_PAD_PKCS7must be supported for CBC and ECB modes. With no padding, CBC or ECB mode encryption must fail if the input isn't a multiple of the block size.
- HMAC SHA-256, with any key size up to at least 32 bytes.
SHA1 and the other members of the SHA2 family (SHA-224, SHA384 and SHA512) are strongly recommended, but not required. Keystore will provide them in software if the hardware Keymaster implementation doesn't provide them.
Some primitives are also recommended for interoperability with other systems:
- Smaller key sizes for RSA
- Arbitrary public exponents for RSA
Key access control
Hardware-based keys that can never be extracted from the device don't provide much security if an attacker can use them at will (though they're more secure than keys which can be exfiltrated). Thus, it's crucial that Keystore enforce access controls.
Access controls are defined as an "authorization list" of tag/value pairs. Authorization tags are 32-bit integers and the values are a variety of types. Some tags may be repeated to specify multiple values. Whether a tag may be repeated is specified in the documentation for the tag. When a key is created, the caller specifies an authorization list. The Keymaster implementation underlying Keystore will modify the list to specify some additional information, such as whether the key has rollback protection, and return a "final" authorization list, encoded into the returned key blob. Any attempt to use the key for any cryptographic operation must fail if the final authorization list is modified.
The set of possible tags is defined in the enumeration
the set must be permanently fixed (though it can be extended).
Names are prefixed with
KM_TAG_. The top
four bits of tag IDs are used to indicate the type.
Possible types include:
KM_ENUM: Many tags' values are defined in enumerations. For example, the possible
KM_TAG_PURPOSE are defined in enum
KM_ENUM_REP: Same as
KM_ENUM, except that the tag may
be repeated in an authorization list. Repetition
indicates multiple authorized values. For example, an encryption key will
KM_UINT: 32-bit unsigned integers. Example:
KM_UINT_REP: Same as
KM_UINT, except that the tag may be
repeated in an authorization list. Repetition indicates multiple authorized values.
KM_ULONG: 64-bit unsigned integers. Example:
KM_ULONG_REP: Same as
KM_ULONG, except that the tag may be
repeated in an authorization list. Repetition
indicates multiple authorized values.
KM_DATE: Date/time values, expressed as milliseconds since January 1, 1970.
KM_BOOL: True or false. A tag of type
KM_BOOL is assumed
to be "false" if the tag is not present and "true" if present. Example:
KM_BIGNUM: Arbitrary-length integers, expressed as a byte array
in big-endian order. Example:
KM_BYTES: A sequence of bytes. Example:
Hardware vs. software enforcement
Not all secure hardware will implement the same features. To support a variety of approaches, Keymaster 1.0 distinguishes between secure and non-secure world access control enforcement, which we call hardware and software enforcement, respectively.
Implementations are required to:
- Enforce exact matching (not enforcement) of all authorizations. Authorization lists in key blobs must exactly match the authorizations returned during key generation, including ordering. Any mismatch must cause an error diagnostic.
- Declare the authorizations whose semantic values are enforced.
The API mechanism for declaring hardware-enforced authorizations is in the
keymaster_key_characteristics_t structure. It divides the
authorization list into two sub-lists,
sw_enforced. The secure hardware is responsible for placing the
appropriate values in each, based on what it can enforce.
In addition, Keystore implements software-based enforcement of all authorizations, whether they're enforced by the secure hardware or not.
For example, consider a TrustZone-based implementation that does not support
key expiration. A key with an expiration date may still be created. That key's
authorization list will include the tag
KM_TAG_ORIGINATION_EXPIRE_DATETIME with the expiration date. A
request to Keystore for the key characteristics will find this tag in the
sw_enforced list and the secure hardware will not enforce the
expiration requirement. However, attempts to use the key after expiration will
be rejected by Keystore.
If the device is then upgraded with secure hardware that does support
expiration, then a request for key characteristics will find
KM_TAG_ORIGINATION_EXPIRE_DATETIME in the
list, and attempts to use the key after expiration will fail even if the
keystore is somehow subverted or bypassed.
Cryptographic message construction authorizations
The following tags are used to define the cryptographic characteristics of
operations using the associated key:
are repeatable, meaning that multiple values may be associated with a single
key, and the value to be used will be specified at operation time.
Keys have an associated set of purposes, expressed as one or more authorization
entries with tag
KM_TAG_PURPOSE, which defines how they can be used. The purposes are:
Any key can have any subset of these purposes. Note that some combinations create security problems. For example, an RSA key that can be used to both encrypt and to sign allows an attacker who can convince the system to decrypt arbitrary data to generate signatures.
Import and export
Keymaster supports export of public keys only, in X.509 format, and import of:
- Public and private key pairs in DER-encoded PKCS#8 format, without password-based encryption, and
- Symmetric keys as raw bytes
To ensure that imported keys can be distinguished from securely-generated
KM_TAG_ORIGIN is included in the appropriate key
authorization list. For example, if a key
was generated in secure hardware,
KM_ORIGIN_GENERATED will be found in
hw_enforced list of the key characteristics, while a key
that was imported into secure
hardware will have the value
Secure Keymaster implementations do not implement user authentication, but depend on other trusted apps which do. For the interface that must be implemented by these apps, see the Gatekeeper page.
User authentication requirements are specified via two sets of tags. The first set indicate which user can use the key:
KM_TAG_ALL_USERSindicates the key is usable by all users. If present,
KM_TAG_SECURE_USER_IDmust not be present.
KM_TAG_USER_IDhas a numeric value specifying the ID of the authorized user. Note that this is the Android user ID (for multi-user), not the application UID, and it is enforced by non-secure software only. If present,
KM_TAG_ALL_USERSmust not be present.
KM_TAG_SECURE_USER_IDhas a 64-bit numeric value specifying the secure user ID that must be provided in a secure authentication token to unlock use of the key. If repeated, the key may be used if any of the values is provided in a secure authentication token.
The second set indicate whether and when the user must be authenticated. If
neither of these tags is present, but
KM_TAG_SECURE_USER_ID is, authentication is
required for every use of the key.
KM_NO_AUTHENTICATION_REQUIREDindicates no user authentication is required, though the key still may only be used by apps running as the user(s) specified by
KM_TAG_AUTH_TIMEOUTis a numeric value specifying, in seconds, how fresh the user authentication must be to authorize key usage. This applies only to private/secret key operations. Public key operations don't require authentication. Timeouts do not cross reboots; after a reboot, all keys are "never authenticated." The timeout may be set to a large value to indicate that authentication is required once per boot (2^32 seconds is ~136 years; presumably Android devices are rebooted more often than that).
Client binding, the association of a key with a particular client
application, is done via an optional client ID and some optional client data
respectively). Keystore treats these values as opaque blobs, only ensuring that
the same blobs presented during key generation/import are presented for every
use and are byte-for-byte identical. The client binding data is not returned by
Keymaster. The caller must know it in order to use the key.
This feature is not exposed to applications.
Keystore supports restricting key usage by date. Key start of validity and
key expirations can be associated with a key and Keymaster will refuse to
perform key operations if the current date/time is outside of the valid
range. The key validity range is specified with the tags
KM_TAG_USAGE_EXPIRE_DATETIME. The distinction between
"origination" and "usage" is based on whether the key is being used to
"originate" a new ciphertext/signature/etc., or to "use" an existing
ciphertext/signature/etc. Note that this distinction is not exposed to
KM_TAG_USAGE_EXPIRE_DATETIME tags are optional. If the tags are absent, it is
assumed that the key in
question can always be used to decrypt/verify messages.
Because wall-clock time is provided by the non-secure world, it's unlikely that the expiration-related tags will be in the hardware-enforced list. Hardware enforcement of expiry would require that the secure world somehow obtain trusted time and data, for example via a challenge response protocol with a trusted remote timeserver.
Root of trust binding
Keystore requires keys to be bound to a root of trust, which is a bitstring provided to the Keymaster secure hardware during startup, preferably by the bootloader. This bitstring must be cryptographically bound to every key managed by Keymaster.
The root of trust consists of the public key used to verify the signature on the boot image and the lock state of the device. If the public key is changed to allow a different system image to be used or if the lock state is changed, then none of the Keymaster-protected keys created by the previous system will be usable, unless the previous root of trust is restored and a system that is signed by that key is booted. The goal is to increase the value of the software-enforced key access controls by making it impossible for an attacker-installed operating system to use Keymaster keys.
Some Keymaster secure hardware may choose to store key material internally
and return handles rather than encrypted key material. Or there may be other
cases in which keys cannot be used until some other non-secure or secure world
system component is available. The Keymaster 1.0 HAL allows the caller to
request that a key be "standalone," via the
meaning that no resources other than the blob and the running Keymaster system
are required. The tags associated with a key may be inspected to see whether a
key is standalone. At present, only two values are defined:
This feature is not exposed to applications.
When it's created, the maximum usage velocity can be specified
TrustZone implementations will refuse to perform cryptographic operations
with that key if an operation was performed less
KM_TAG_MIN_SECONDS_BETWEEN_OPS seconds earlier.
The simple approach to implementing velocity limits is a table of key IDs and last-use timestamps. This table will likely be of limited size, but must accommodate at least 16 entries. In the event that the table is full and no entries may be updated or discarded, secure hardware implementations must "fail safe," preferring to refuse all velocity-limited key operations until one of the entries expires. It is acceptable for all entries to expire upon reboot.
Keys can also be limited to n uses per boot with
KM_TAG_MAX_USES_PER_BOOT. This also requires a tracking table,
which must accommodate at least four keys, and must also fail safe. Note that
applications will be unable to create per-boot limited keys. This feature will
not be exposed through Keystore and will be reserved for system operations.
This feature is not exposed to applications.
Random number generator re-seeding
Because secure hardware must generate random numbers for key material and Initialization Vectors (IVs), and because hardware random number generators may not always be fully trustworthy, the Keymaster HAL provides an interface to allow the client to provide additional entropy which will be mixed into the random numbers generated.
A hardware random-number generator must be used as the primary seed source, and the seed data provided through the external API must not be the sole source of randomness used for number generation. Further, the mixing operation used must ensure that the random output is unpredictable if any one of the seed sources is unpredictable.
This feature is not exposed to applications but is used by the framework, which regularly provides additional entropy, retrieved from a Java SecureRandom instance, to the secure hardware.