core: crypto: add support MD5 hashes in RSA sign/verify/cipher

Introduce support of using MD5 hashes in RSA sign/verify/cipher
operations, which is required by AOSP Keymaster.

This is verified in

core: crypto: add support MD5 hashes in RSA sign/verify/cipher

Introduce support of using MD5 hashes in RSA sign/verify/cipher
operations, which is required by AOSP Keymaster.

This is verified in VerificationOperationsTest.RsaSuccess VTS Test [1],
which checks usage of such digests: NONE, MD5, SHA1, SHA_2_224, SHA_2_256,
SHA_2_384, SHA_2_512.

This patch has been inspired by commit[2]:

Link: [1] https://android.googlesource.com/platform/hardware/interfaces/+/master/keymaster/3.0/vts/functional/keymaster_hidl_hal_test.cpp
Link: [2] https://github.com/OP-TEE/optee_os/commit/199d0b7310d1705661a106358f1f0b46e4c5c587 ("core: crypto: add support MD5 hashes in RSA sign/verify")
Signed-off-by: Julien Masson <jmasson@baylibre.com>
Signed-off-by: Safae Ouajih <souajih@baylibre.com>
Acked-by: Jerome Forissier <jerome.forissier@linaro.org>
Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: crypto: change supported HMAC key size ranges

Change supported HMAC key size range from 64 to 1024 when
CFG_HMAC_64_1024_RANGE config is enabled, This is required
to successfully pass AOSP Key

core: crypto: change supported HMAC key size ranges

Change supported HMAC key size range from 64 to 1024 when
CFG_HMAC_64_1024_RANGE config is enabled, This is required
to successfully pass AOSP Keymaster VTS tests.

From TEE Internal Core API specificaion:
"Table 5-9, support for other sizes or algorithms
is implementation-defined."

Signed-off-by: Igor Opaniuk <igor.opaniuk@linaro.org>
Signed-off-by: Safae Ouajih <souajih@baylibre.com>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: crypto: fix memory leak in Ed25519 support

The software implementation of ED25519 algorithm has a memory leak in the
key and key pair allocation. Upon every public key allocation, a key pair
i

core: crypto: fix memory leak in Ed25519 support

The software implementation of ED25519 algorithm has a memory leak in the
key and key pair allocation. Upon every public key allocation, a key pair
is allocated (public and private components). When freeing the public
key, only the public component is freed. To reproduce the issue:

$ while xtest 4016; do :; done

Until the following error:

* regression_4016 Test TEE Internal API ED25519 sign/verify
E/LD: copy_section_headers:1124 sys_copy_from_ta_bin
E/TC:? 0 ldelf_init_with_ldelf:131 ldelf failed with res: 0xffff000c /usr/src/debug/optee-test/master.imx-r0/host/xtest/regression_4000.c:6062: xtest_teec_open_session(&session, &crypt_user_ta_uuid, ((void *)0), &ret_orig) has an unexpected value: 0xffff000c = TEEC_ERROR_OUT_OF_MEMORY, expected 0x0 = TEEC_SUCCESS
regression_4016 FAILED

To fix the memory leak, a separate public key allocation function must
be defined along a ED25519 public key structure.

Fixes: 0aaad418ac8b ("core: crypto: add Ed25519 support")
Signed-off-by: Clement Faure <clement.faure@nxp.com>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

GP131: Add TA property gpd.ta.doesNotCloseHandleOnCorruptObject

Adds the TA property gpd.ta.doesNotCloseHandleOnCorruptObject.

All syscalls operating on an object handle and can return
TEE_ERROR_CO

GP131: Add TA property gpd.ta.doesNotCloseHandleOnCorruptObject

Adds the TA property gpd.ta.doesNotCloseHandleOnCorruptObject.

All syscalls operating on an object handle and can return
TEE_ERROR_CORRUPT_OBJECT must also do special treatment when
TEE_ERROR_CORRUPT_OBJECT is returned. Prior to
gpd.ta.doesNotCloseHandleOnCorruptObject this meant removing the object
and closing the object handle. With the
gpd.ta.doesNotCloseHandleOnCorruptObject property the object handle
shouldn't be close if this the current TA has the property set to true.

Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: syscall_hash_final() support XOF

Adds support in syscall_hash_final() for Extendable Output Functions,
that is, TEE_ALG_SHAKE128 and TEE_ALG_SHAKE256 for now.

Acked-by: Etienne Carriere <etie

core: syscall_hash_final() support XOF

Adds support in syscall_hash_final() for Extendable Output Functions,
that is, TEE_ALG_SHAKE128 and TEE_ALG_SHAKE256 for now.

Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: crypto: add API for SHA-3

Adds an API for the SHA-3 [1] algorithms SHA3-224, SHA3_256, SHA3_384,
SHA3_512, SHAKE128, and SHAKE256. SHAKE128 and SHAKE256 are
Extendable-Output Functions (XOF),

core: crypto: add API for SHA-3

Adds an API for the SHA-3 [1] algorithms SHA3-224, SHA3_256, SHA3_384,
SHA3_512, SHAKE128, and SHAKE256. SHAKE128 and SHAKE256 are
Extendable-Output Functions (XOF), but handled as hashes in the API.

Adds API for SHA-3 based HMAC functions.

For XOF algorithms crypto_hash_final() can be called multiple times and
generate an arbitrary amount of data. Note that this data will have a
common prefix if done over the same input (see A.2 in [1]).

[1] https://nvlpubs.nist.gov/nistpubs/FIPS/NIST.FIPS.202.pdf

Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

GP131: Update gpd.tee.trustedStorage.antiRollback.protectionLevel

Updates used values for
gpd.tee.trustedStorage.antiRollback.protectionLevel according to TEE
Internal Core API version 1.3.1.

Note

GP131: Update gpd.tee.trustedStorage.antiRollback.protectionLevel

Updates used values for
gpd.tee.trustedStorage.antiRollback.protectionLevel according to TEE
Internal Core API version 1.3.1.

Note that protection level 0 is not valid any longer, so report level
100 when RPMB isn't in use. Only root can tamper with the REE FS storage
due to the REE file system permissions.

Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

GP131: add TA property gpd.client.endian

Adds the new property "gpd.client.endian" to report the endianness used
by a client.

At the moment there is no way for a client to report its endianness to

GP131: add TA property gpd.client.endian

Adds the new property "gpd.client.endian" to report the endianness used
by a client.

At the moment there is no way for a client to report its endianness to
OP-TEE so assume that all are little-endian.

Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

GP131: Update TEE_AsymmetricEncrypt() and TEE_AsymmetricDecrypt()

Updates TEE_AsymmetricEncrypt() and TEE_AsymmetricDecrypt() according to
TEE Internal Core API version 1.3.1. The srcLen and dstLen

GP131: Update TEE_AsymmetricEncrypt() and TEE_AsymmetricDecrypt()

Updates TEE_AsymmetricEncrypt() and TEE_AsymmetricDecrypt() according to
TEE Internal Core API version 1.3.1. The srcLen and dstLen parameters
are changed to use size_t instead of uint32_t

Adds the optional attribute TEE_ATTR_RSA_OAEP_MGF_HASH which if provided
specifies the MGF1 hash function to use. Prior to GP v1.3.1 it was
implementation defined. OP-TEE only supports using the same MGF1 hash
function as the internal hash function of the algorithim, that is, for
the algorithm TEE_ALG_RSAES_PKCS1_OAEP_MGF1_x that hash function
TEE_ALG_x is used.

Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

GP131: Update TEE_CreatePersistentObject()

Update TEE_CreatePersistentObject() according to TEE Internal Core API
version 1.3.1. The objectIDLen parameter is changed to use size_t
instead of uint32_

GP131: Update TEE_CreatePersistentObject()

Update TEE_CreatePersistentObject() according to TEE Internal Core API
version 1.3.1. The objectIDLen parameter is changed to use size_t
instead of uint32_t.

The object parameter is now an optional pointer to a handle. When NULL
the standard says:
If attributes is a handle on an initialized transient object,
the initialized transient object SHALL be transformed to a
persistent object.
So syscall_storage_obj_create() is also updated accordingly.

Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

GP131: Update HKDF defines

OP-TEE provides an HKDF implementation as an extension to the TEE
Internal Core API which predates version 1.3 of the specification where
HKDF was officially introduced.

GP131: Update HKDF defines

OP-TEE provides an HKDF implementation as an extension to the TEE
Internal Core API which predates version 1.3 of the specification where
HKDF was officially introduced.

Update with the official defines and resolve name clashes. With this
OP-TEE supports both the old extension and the new official API.

Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

GP131: Add TEE_ALG_ECDH_DERIVE_SHARED_SECRET and TEE_ALG_ECDSA_SHA*

Adds the defines
- TEE_ALG_ECDH_DERIVE_SHARED_SECRET
- TEE_ALG_ECDSA_SHA1
- TEE_ALG_ECDSA_SHA224
- TEE_ALG_ECDSA_SHA256
- TEE_ALG_

GP131: Add TEE_ALG_ECDH_DERIVE_SHARED_SECRET and TEE_ALG_ECDSA_SHA*

Adds the defines
- TEE_ALG_ECDH_DERIVE_SHARED_SECRET
- TEE_ALG_ECDSA_SHA1
- TEE_ALG_ECDSA_SHA224
- TEE_ALG_ECDSA_SHA256
- TEE_ALG_ECDSA_SHA386
- TEE_ALG_ECDSA_SHA512
from TEE Internal Core API version 1.3.1 to replace the previous now
deprecated defines:
- TEE_ALG_ECDSA_P192
- TEE_ALG_ECDSA_P224
- TEE_ALG_ECDSA_P256
- TEE_ALG_ECDSA_P384
- TEE_ALG_ECDSA_P521
- TEE_ALG_ECDH_P192
- TEE_ALG_ECDH_P224
- TEE_ALG_ECDH_P256
- TEE_ALG_ECDH_P384
- TEE_ALG_ECDH_P521

The new defines have new values and the values from the old defines are
not reused.

The syscall ABI maintains compatibility by handling both the old and new
values in parallel from now on.

Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: crypto: fix TEE_ATTR_EDDSA_PREHASH interpretation

Commit 0aaad418ac8b ("core: crypto: add Ed25519 support") introduced
support for the ED25519 algorithm. This included parsing a
TEE_ATTR_EDDSA

core: crypto: fix TEE_ATTR_EDDSA_PREHASH interpretation

Commit 0aaad418ac8b ("core: crypto: add Ed25519 support") introduced
support for the ED25519 algorithm. This included parsing a
TEE_ATTR_EDDSA_PREHASH parameter that unfortunately was not fully
compliant with the standard. So fix this with a more strict
interpretation of TEE_ATTR_EDDSA_PREHASH as described in the
specification.

Fixes: 0aaad418ac8b ("core: crypto: add Ed25519 support")
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Reviewed-by: Valerii Chubar <valerii_chubar@epam.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: syscall_storage_obj_write(): handle corrupt object

All syscalls operating on an object handle and can return
TEE_ERROR_CORRUPT_OBJECT must also remove the object if the return code
is TEE_ERRO

core: syscall_storage_obj_write(): handle corrupt object

All syscalls operating on an object handle and can return
TEE_ERROR_CORRUPT_OBJECT must also remove the object if the return code
is TEE_ERROR_CORRUPT_OBJECT. This is missing in
syscall_storage_obj_write() so add the missing call to remove the object
if it is corrupt.

Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: crypto: Fix tee_obj_attr_copy_from for X25519

Make sure in tee_obj_attr_copy_from() if the source object type
is TEE_TYPE_X25519_KEYPAIR the destination object type should be
TEE_TYPE_X25519_P

core: crypto: Fix tee_obj_attr_copy_from for X25519

Make sure in tee_obj_attr_copy_from() if the source object type
is TEE_TYPE_X25519_KEYPAIR the destination object type should be
TEE_TYPE_X25519_PUBLIC_KEY. The function should extract public
key attributes if presented with a key pair object.

According to the TEE Internal Core API Specification v1.2 we should
check if the source is of type *_KEYPAIR then the destination is a
subset of type *_PUBLIC_KEY when populating the destination object as
listed in the Table 5-11.

Fixes: 90040fa4c81c ("core: crypto: add X25519 support")
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Sohaib ul Hassan <sohaib.ul.hassan@unikie.com>

show more ...

Fix TEE_ATTR_ECC_EPHEMERAL_PUBLIC_VALUE_*

Commit 5b385b3f835d ("core: crypto: add support for SM2 KEP")defined by
mistake the wrong values for these two. To fix this we're are renaming
these IDs alt

Fix TEE_ATTR_ECC_EPHEMERAL_PUBLIC_VALUE_*

Commit 5b385b3f835d ("core: crypto: add support for SM2 KEP")defined by
mistake the wrong values for these two. To fix this we're are renaming
these IDs alternative IDs which OP-TEE will recognize in addition to the
correct official values when deriving a key using the TEE_ALG_SM2_KEP
algorithm.

TEE_ATTR_ECC_EPHEMERAL_PUBLIC_VALUE_X and
TEE_ATTR_ECC_EPHEMERAL_PUBLIC_VALUE_Y are only used as input parameters
so there is no need to translate back to the old invalid values.

Fixes: 5b385b3f835d ("core: crypto: add support for SM2 KEP")
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core, libutee: introduce TEE_ALG_SM4_XTS

In this patch, The sm4-xts algorithm is supported in the GP process.

Signed-off-by: Pingan Xie <xiepingan3@huawei.com>
Reviewed-by: Xiaoxu Zeng <zengxiaoxu@

core, libutee: introduce TEE_ALG_SM4_XTS

In this patch, The sm4-xts algorithm is supported in the GP process.

Signed-off-by: Pingan Xie <xiepingan3@huawei.com>
Reviewed-by: Xiaoxu Zeng <zengxiaoxu@huawei.com>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>

show more ...

core: tee_svc_storage: remove unused local variable

Removes file local variable from syscall_storage_obj_open() that is
not used.

Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewe

core: tee_svc_storage: remove unused local variable

Removes file local variable from syscall_storage_obj_open() that is
not used.

Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

Use struct utee_object_info in TA syscall abi

TEE_ObjectInfo was until now used in the syscall ABI provided to TAs.
TEE_ObjectInfo changes in later versions of the TEE Internal Core API so
add an in

Use struct utee_object_info in TA syscall abi

TEE_ObjectInfo was until now used in the syscall ABI provided to TAs.
TEE_ObjectInfo changes in later versions of the TEE Internal Core API so
add an independent definition with struct utee_object_info in order to
preserve a stable ABI.

Acked-by: Jerome Forissier <jerome.forissier@linaro.org>
Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: update to new field names in TEE_ObjectInfo

With GP TEE Internal Core API version 1.1.1 the two fields keySize and
maxKeySize was renamed to objectSize and maxObjectSize respectively.
Update t

core: update to new field names in TEE_ObjectInfo

With GP TEE Internal Core API version 1.1.1 the two fields keySize and
maxKeySize was renamed to objectSize and maxObjectSize respectively.
Update the rest of the core code to reflect that change as a preparation
for future updates.

Acked-by: Jerome Forissier <jerome.forissier@linaro.org>
Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: fix potential integer overflow in syscall_log()

Fixes a potential integer overflow in syscall_log(). Note that an
eventual overflow would still be caught by copy_from_user(), but it's
preferab

core: fix potential integer overflow in syscall_log()

Fixes a potential integer overflow in syscall_log(). Note that an
eventual overflow would still be caught by copy_from_user(), but it's
preferable to catch this earlier.

Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Suggested-by: Asaf Modelevsky <amodele@amazon.com>
Reported-by: Asaf Modelevsky <amodele@amazon.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: avoid TA panic when secure storage is corrupt

init_head_from_data() triggers a TA panic in case corrupt data is read
from the secure storage, for instance by request from the PKCS#11
trusted a

core: avoid TA panic when secure storage is corrupt

init_head_from_data() triggers a TA panic in case corrupt data is read
from the secure storage, for instance by request from the PKCS#11
trusted application.
"Every Trusted Storage implementation is expected to return
TEE_ERROR_CORRUPT_OBJECT if a Trusted Application attempts to open
an object and the TEE determines that its contents (or those of the
storage itself) have been tampered with or rolled back."
See TEE Internal Core API Specification v1.1.2, section 5.7.1.

Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Zengxx <zengxiaoxu@huawei.com>
Signed-off-by: Joakim Nordell <joakim.nordell@axis.com>
Acked-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

crypto: se050: updates to the crypto object deletion interface

Keys created on the Secure Element NVM via the PKCS#11 TA are removed
by scanning the data buffer holding the reference to the key duri

crypto: se050: updates to the crypto object deletion interface

Keys created on the Secure Element NVM via the PKCS#11 TA are removed
by scanning the data buffer holding the reference to the key during
the release of the object.

The storage allocated to hold those keys (ECC/RSA) is always below the
page size length which seems like a reasonable figure to use for future
extensions.

- This commit avoids scanning objects larger than that length.

This commit also updates the interface to delegate the actual handling
of the object to the crypto driver instead of passing just the raw data
contained in the object.

The cryptographic layer is also being allowed to block the deletion of
the object. This is to cover the scenario where the I2C device is not
accessible while a reference to the key is being removed from the secure
storage in the filesystem.

Incidentally also fixes regression 6018: this test releases an object
of size 0xA0000 which can't be scanned due to this part of the code
hitting an Out of Memory condition.

Signed-off-by: Jorge Ramirez-Ortiz <jorge@foundries.io>
Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: crypto: remove TEE_ATTR_ECC_CURVE as an attribute of TEE_TYPE_ED25519_KEYPAIR

The ECC curve is not an attribute of an Ed25519 key pair, let alone a
mandatory one for key generation. It was mis

core: crypto: remove TEE_ATTR_ECC_CURVE as an attribute of TEE_TYPE_ED25519_KEYPAIR

The ECC curve is not an attribute of an Ed25519 key pair, let alone a
mandatory one for key generation. It was mistakenly added by commit
03e07432b68f ("ta: pkcs11: Add Ed25519 support"), thus breaking xtest
regression_4007_ed25519 (subcase .1 Generate Ed25519 key).
Remove that attribute from the key type definition.

Fixes: 03e07432b68f ("ta: pkcs11: Add Ed25519 support")
Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: mmu: Fix wrong input argument of tee_mm_init()

Since commit [1], tee_mm_init() take pool size instead of end address.
This change corrects the input arg of caller which still use old
definitio

core: mmu: Fix wrong input argument of tee_mm_init()

Since commit [1], tee_mm_init() take pool size instead of end address.
This change corrects the input arg of caller which still use old
definition.

Link: [1] 2380d70 ("core: mmu: fix overflow with high address in tee_mm_pool_t")
Signed-off-by: james.jiang <james.jiang@mediatek.com>
Signed-off-by: Mark-PK Tsai <mark-pk.tsai@mediatek.com>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...