core: fs_htree: fix incorrect data type in create_and_sync

Fixes an incorrect data type used in the create_and_sync function.
When this function calls tee_fs_htree_sync_to_storage, it may invoke
tee

core: fs_htree: fix incorrect data type in create_and_sync

Fixes an incorrect data type used in the create_and_sync function.
When this function calls tee_fs_htree_sync_to_storage, it may invoke
tee_fs_htree_close to free the input hash tree in error cases.
Therefore, the input type should be struct tee_fs_htree **ht_arg
to handle these scenarios correctly.

Link: https://github.com/OP-TEE/optee_os/issues/7512
Fixes: 5a9d570a8d7f ("core: fs_htree: re-init hash tree when both tag and counter are zero")

Signed-off-by: Ox Yeh <ox.yeh@mediatek.com>
Reviewed-by: Menson Chen <menson.chen@mediatek.com>
Reviewed-by: Rayan Hu <rayan.hu@mediatek.com>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@foss.st.com>

show more ...

core: support dynamic protected memory lending

With CFG_CORE_DYN_PROTMEM=y support dynamic protected memory lending.

A new internal struct mobj_ffa_rsm is added to handle dynamic protected
memory f

core: support dynamic protected memory lending

With CFG_CORE_DYN_PROTMEM=y support dynamic protected memory lending.

A new internal struct mobj_ffa_rsm is added to handle dynamic protected
memory for FF-A.

A new internal struct mobj_protmem is add to handle dynamic protected
memory without FF-A.

Lending non-secure memory to OP-TEE to use it as protected memory means
that it should to become inaccessible by the normal world as part of the
process. This part is currently not supported, since it must be done in
a platform specific way for platforms that support that. QEMU don't
support that.

Adding two platform specific functions, plat_get_protmem_config() and
plat_set_protmem_range() for dynamic protected memory. The functions has
__weak implementation to allow easier testing. However,
plat_set_protmem_range() requires CFG_INSECURE=y since it doesn't change
memory protection.

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

show more ...

core: fs_htree: re-init hash tree when both tag and counter are zero

Creating and update a hash tree involves several RPC commands.
If a power loss occurs during the creation flow, it may result
in

core: fs_htree: re-init hash tree when both tag and counter are zero

Creating and update a hash tree involves several RPC commands.
If a power loss occurs during the creation flow, it may result
in a hash tree with an incomplete header and a counter value of 0.
If attempting to read this file subsequently leads to a
TEE_ERROR_CORRUPT_OBJECT error.

Instead of returning TEE_ERROR_CORRUPT_OBJECT, continue the
initialization flow to support subsequent functionality.

Link: https://github.com/OP-TEE/optee_os/issues/7512
Fixes: 50a814981d8b ("core: provide a hash tree for secure storage")

Signed-off-by: Ox Yeh <ox.yeh@mediatek.com>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>
Acked-by: Etienne Carriere <etienne.carriere@foss.st.com>

show more ...

core: fs_htree: Fix AAD length when CFG_REE_FS_HTREE_HASH_SIZE_COMPAT=y

Correct the hash size declared in AAD length declared in REE FS hash
tree authentication sequence when CFG_REE_FS_HTREE_HASH_S

core: fs_htree: Fix AAD length when CFG_REE_FS_HTREE_HASH_SIZE_COMPAT=y

Correct the hash size declared in AAD length declared in REE FS hash
tree authentication sequence when CFG_REE_FS_HTREE_HASH_SIZE_COMPAT is
enabled in which case the hash is truncated to the size of the
FEK key (TEE_FS_HTREE_FEK_SIZE).

The issue has currently no impact since REE FS hash tree authentication
is based on AES-GCM but it would be of importance if, for example, one
moves to an AES-CCM scheme while still enabling
CFG_REE_FS_HTREE_HASH_SIZE_COMPAT (even if unlikely to happen).
To prevent such issue in the future, let's declare the effectively
used hash size.

Suggested-by: Jens Wiklander <jens.wiklander@linaro.org>
Link: https://github.com/OP-TEE/optee_os/pull/7340/commits/087325faec7c057a638cca80f0549e9abe49f190#r2024716984
Signed-off-by: Etienne Carriere <etienne.carriere@foss.st.com>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: fs_htree: Fix wrong AAD length in authenc_init()

In authenc_init(), AAD length field passed to
crypto_authenc_init() does not match with the total
AAd data passed via crypto_authenc_update_aad

core: fs_htree: Fix wrong AAD length in authenc_init()

In authenc_init(), AAD length field passed to
crypto_authenc_init() does not match with the total
AAd data passed via crypto_authenc_update_aad() for
lower layer crypto computation.

To fix this issue and to support the legacy without
breaking existing REE file system content, introduce
a config CFG_REE_FS_HTREE_HASH_SIZE_COMPAT.

By default this config is enabled to support the legacy
REE FS hash tree that uses truncated hash implementation.

Link: https://github.com/OP-TEE/optee_os/issues/7331
Signed-off-by: Anil Kumar Reddy <areddy3@marvell.com>
Reviewed-by: Etienne Carriere <etienne.carriere@foss.st.com>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: fail build when CFG_INSECURE=n used with unsafe RNG seed

A production build needs a proper RNG seed. Fail the build when
CFG_INSECURE=n is used in conjunction with the default weak stub
implem

core: fail build when CFG_INSECURE=n used with unsafe RNG seed

A production build needs a proper RNG seed. Fail the build when
CFG_INSECURE=n is used in conjunction with the default weak stub
implementation of plat_rng_init().

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@foss.st.com>

show more ...

core: rpmb: get RPMB storage space allocation stats

Add rpmb_mem_stats() to get the RPMB secure storage space allocation
statistics.

Signed-off-by: Etienne Carriere <etienne.carriere@foss.st.com>
A

core: rpmb: get RPMB storage space allocation stats

Add rpmb_mem_stats() to get the RPMB secure storage space allocation
statistics.

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

show more ...

core: rpmb: fix mutex in directory populate

Fix mutex unlocking in rpmb_fs_dir_populate() that should protect
fat_entry_dir_deinit() execution.

Fixes: 5f68d7848fe8 ("core: RPMB FS: Caching for FAT

core: rpmb: fix mutex in directory populate

Fix mutex unlocking in rpmb_fs_dir_populate() that should protect
fat_entry_dir_deinit() execution.

Fixes: 5f68d7848fe8 ("core: RPMB FS: Caching for FAT FS entries")
Signed-off-by: Etienne Carriere <etienne.carriere@foss.st.com>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: rpmb: handle not-implemented error code on device probing

Consider that non-secure world may report TEE_ERROR_NOT_IMPLEMENTED
error code on OPTEE_RPC_CMD_RPMB_PROBE_RESET RPC command when that

core: rpmb: handle not-implemented error code on device probing

Consider that non-secure world may report TEE_ERROR_NOT_IMPLEMENTED
error code on OPTEE_RPC_CMD_RPMB_PROBE_RESET RPC command when that
RPC command is not supported. This is needed to support U-Boot
that provides this return code, at least up to its release tag
v2025.01 [1].

Without this change, OP-TEE fails to communicate with U-Boot RPMB
RPC service since RPMB probe capability was merged in OP-TEE
(Fixes: tag below).

Link: https://source.denx.de/u-boot/u-boot/-/blame/v2025.01/drivers/tee/optee/supplicant.c?ref_type=tags#L96 [1]
Reported-by: Gavin Liu <gavin.liu@mediatek.com>
Closes: https://github.com/OP-TEE/optee_os/issues/7200
Fixes: 8dfdf3927214 ("core: rpmb: probe for kernel RPMB driver")
Signed-off-by: Etienne Carriere <etienne.carriere@foss.st.com>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

tree-wide: use ROUNDUP_DIV() where applicable

Use ROUNDUP_DIV() instead of ROUNDUP(..., size) / size where applicable.

Signed-off-by: Etienne Carriere <etienne.carriere@foss.st.com>
Reviewed-by: Je

tree-wide: use ROUNDUP_DIV() where applicable

Use ROUNDUP_DIV() instead of ROUNDUP(..., size) / size where applicable.

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

show more ...

core: define plat_rng_init() as a weak alias of __plat_rng_init()

Rename the function plat_rng_init() to __plat_rng_init() and define
plat_rng_init() as a weak alias of __plat_rng_init().
This allow

core: define plat_rng_init() as a weak alias of __plat_rng_init()

Rename the function plat_rng_init() to __plat_rng_init() and define
plat_rng_init() as a weak alias of __plat_rng_init().
This allow to use the default RNG initialization in the platform
plat_rng_init() function.

Signed-off-by: Thomas Bourgoin <thomas.bourgoin@foss.st.com>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>
Acked-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@foss.st.com>

show more ...

core: rpmb: probe for kernel RPMB driver

Three RPC functions are added to support RPMB probing and properly align
RPMB frames, OPTEE_RPC_CMD_RPMB_PROBE_RESET,
OPTEE_RPC_CMD_RPMB_PROBE_NEXT, and OPTE

core: rpmb: probe for kernel RPMB driver

Three RPC functions are added to support RPMB probing and properly align
RPMB frames, OPTEE_RPC_CMD_RPMB_PROBE_RESET,
OPTEE_RPC_CMD_RPMB_PROBE_NEXT, and OPTEE_RPC_CMD_RPMB_FRAMES.

OPTEE_RPC_CMD_RPMB_PROBE_RESET resets probing to a well known state and
returns the shared memory type needed when allocating shared memory for
communication with later RPMB functions.

OPTEE_RPC_CMD_RPMB_PROBE_NEXT selects the next RPMB device and returns
its device information. Later calls to OPTEE_RPC_CMD_RPMB will use this
selected device.

OPTEE_RPC_CMD_RPMB_FRAMES sends the raw RPMB frames to normal world for
further routing to the RPMB device.

tee_rpmb_reinit() is added to allow re-initializing the RPMB FS if a
boot stage has used RPMB.

Backwards compatibility is maintained by falling back to the old type of
initialization if OPTEE_RPC_CMD_RPMB_PROBE_RESET returns
TEE_ERROR_NOT_SUPPORTED.

Whether RPMB devices are probed by the kernel or tee-supplicant is
decided by the kernel driver where the shared memory type returned by
OPTEE_RPC_CMD_RPMB_PROBE_RESET plays a vital role.

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

show more ...

core: rpmb: use thread_rpc_shm_cache_alloc()

Use thread_rpc_shm_cache_alloc() to cache shared memory allocations to
simplify and make the code slightly more efficient.

Signed-off-by: Jens Wiklander

core: rpmb: use thread_rpc_shm_cache_alloc()

Use thread_rpc_shm_cache_alloc() to cache shared memory allocations to
simplify and make the code slightly more efficient.

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

show more ...

core: rpmb: set device ID in tee_rpmb_init()

Until now, when initializing RPMB the devices ID has been passed to
tee_rpmb_init(), but there isn't any flexibility since it's always
CFG_RPMB_FS_DEV_ID

core: rpmb: set device ID in tee_rpmb_init()

Until now, when initializing RPMB the devices ID has been passed to
tee_rpmb_init(), but there isn't any flexibility since it's always
CFG_RPMB_FS_DEV_ID passed along. So simplify the code a bit and
use CFG_RPMB_FS_DEV_ID directly in tee_rpmb_init().

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

show more ...

core: fix error handling in tee_svc_storage_read_head()

Prior to this all errors except TEE_ERROR_OUT_OF_MEMORY from
fops->read() was reported as TEE_ERROR_CORRUPT_OBJECT leading
to removal of the o

core: fix error handling in tee_svc_storage_read_head()

Prior to this all errors except TEE_ERROR_OUT_OF_MEMORY from
fops->read() was reported as TEE_ERROR_CORRUPT_OBJECT leading
to removal of the object.
We should not treat all errors as corrupt, so remove the error
code translation.

Signed-off-by: Pengguang Zhu <pengguang.zhu@amlogic.com>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: add and use service_init_crypto() macro

Add the service_init_crypto() used to initialize the internal crypto API
before the rest of the registered service_init functions are called.

Signed-of

core: add and use service_init_crypto() macro

Add the service_init_crypto() used to initialize the internal crypto API
before the rest of the registered service_init functions are called.

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

show more ...

core: pass TEE_ATTR_RSA_OAEP_MGF_HASH to RSA-OAEP implementations

OP-TEE currently doesn't support using a different hash for MGF1
with RSA-OAEP. However, this is required for AOSP compatibility
(e.

core: pass TEE_ATTR_RSA_OAEP_MGF_HASH to RSA-OAEP implementations

OP-TEE currently doesn't support using a different hash for MGF1
with RSA-OAEP. However, this is required for AOSP compatibility
(e.g. in EncryptionOperationsTest.RsaOaepWithMGFDigestSuccess [1]).

Pass the MGF1 attribute to crypto implementations. Note that
only libtomcrypt supports this feature at the moment, so other
implementations will either fail or fall back to libtomcrypt when
passed a different MGF1 hash.

Link: https://android.googlesource.com/platform/hardware/interfaces/+/refs/heads/main/security/keymint/aidl/vts/functional/KeyMintTest.cpp#5552 [1]
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

core: ree_fs: fix dirfile handle refcount

The counter ree_fs_dirh_refcount is used to determine when ree_fs_dirh
should be free, not as a guarantee that ree_fs_dirh is still valid. This
wasn't the a

core: ree_fs: fix dirfile handle refcount

The counter ree_fs_dirh_refcount is used to determine when ree_fs_dirh
should be free, not as a guarantee that ree_fs_dirh is still valid. This
wasn't the assumption in ree_fs_readdir_rpc(), ree_fs_closedir_rpc(),
and ree_fs_opendir_rpc(). So fix that by using get_dirh() in
ree_fs_readdir_rpc as needed.

Reported-by: Gavin Liu <gavin.liu@mediatek.com>
Closes: https://github.com/OP-TEE/optee_os/issues/6895
Fixes: ace6039fd434 ("core: REE_FS: refcount dirfile handle")
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>

show more ...

core: add CFG_RSA_PUB_EXPONENT_3

When generating RSA key pairs, OP-TEE currently enforces a minimum public
exponent size of 65537 per NIST SP800-56B recommendations. However, AOSP
KeyMint VTS (Encry

core: add CFG_RSA_PUB_EXPONENT_3

When generating RSA key pairs, OP-TEE currently enforces a minimum public
exponent size of 65537 per NIST SP800-56B recommendations. However, AOSP
KeyMint VTS (EncryptionOperationsTest.RsaNoPaddingSuccess [1]) requires
implementations to support public exponent 3 for backwards compatibility.
Add CFG_RSA_PUB_EXPONENT_3 to allow public exponents >= 3.

Link: https://android.googlesource.com/platform/hardware/interfaces/+/refs/heads/main/security/keymint/aidl/vts/functional/KeyMintTest.cpp#5258 [1]
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@foss.st.com>

show more ...

core: pta: drop benchmark

Drop Benchmark PTA as current implementation is non-function
and obsolete, and it's not supported anymore.

Acked-by: Jens Wiklander <jens.wiklander@linaro.org>
Acked-by: E

core: pta: drop benchmark

Drop Benchmark PTA as current implementation is non-function
and obsolete, and it's not supported anymore.

Acked-by: Jens Wiklander <jens.wiklander@linaro.org>
Acked-by: Etienne Carriere <etienne.carriere@foss.st.com>
Signed-off-by: Igor Opaniuk <igor.opaniuk@foundries.io>

show more ...

core: tee: initialize dirfile|tadb_entry objects

Coverity reports many errors where dirfile_entry{} and tadb_entry{}
objects are used un-initialized.
Resolve these errors by setting these objects to

core: tee: initialize dirfile|tadb_entry objects

Coverity reports many errors where dirfile_entry{} and tadb_entry{}
objects are used un-initialized.
Resolve these errors by setting these objects to zero on declaration.

Signed-off-by: Clement Faure <clement.faure@nxp.com>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@foss.st.com>

show more ...

core: check if binary to bignum conversion succeed

Check return value of crypto_bignum_bin2bn().

Signed-off-by: Clement Faure <clement.faure@nxp.com>
Reviewed-by: Jerome Forissier <jerome.forissier

core: check if binary to bignum conversion succeed

Check return value of crypto_bignum_bin2bn().

Signed-off-by: Clement Faure <clement.faure@nxp.com>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@foss.st.com>

show more ...

core: remove unnecessary includes

Remove unnecessary includes.

Signed-off-by: Clement Faure <clement.faure@nxp.com>
Acked-by: Jerome Forissier <jerome.forissier@linaro.org>
Acked-by: Jens Wiklander

core: remove unnecessary includes

Remove unnecessary includes.

Signed-off-by: Clement Faure <clement.faure@nxp.com>
Acked-by: Jerome Forissier <jerome.forissier@linaro.org>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>
Acked-by: Etienne Carriere <etienne.carriere@foss.st.com>

show more ...

core: rpmb: cleanup useless write_counter in struct rpmb_fat_entry

The write_counter field of the rpmb_fat_entry struct is set from the
RPMB device but never used. Rename it to 'unused' and cleanup

core: rpmb: cleanup useless write_counter in struct rpmb_fat_entry

The write_counter field of the rpmb_fat_entry struct is set from the
RPMB device but never used. Rename it to 'unused' and cleanup the
associated code.

Tested on QEMUv8 with:

$ make CFG_RPMB_FS=y CFG_RPMB_WRITE_KEY=y CFG_RPMB_TESKEY=y check

Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>
Acked-by: Etienne Carriere <etienne.carriere@foss.st.com>
Tested-by: Etienne Carriere <etienne.carriere@foss.st.com> (stm32mp157c-ev1 w/ RPMB)

show more ...

core: fix objectUsage handling for persistent objects

Before this patch was the object usage field stored in the
non-persistent part of an object handle, regardless of whether the
storage object was

core: fix objectUsage handling for persistent objects

Before this patch was the object usage field stored in the
non-persistent part of an object handle, regardless of whether the
storage object was persistent. This prevents updates to this field from
being restored the next time the persistent object is opened. Updates to
the field are also not replicated to eventual other open handles for the
object. Fix this by storing the "usage" bits in a new obj_info_usage
field in struct tee_pobj for persistent objects. Updates to the field
are also written into secure storage to preserve the content the next
time the object is opened.

Fixes: b01047730e77 ("Open-source the TEE Core")
Closes: https://github.com/OP-TEE/optee_os/issues/6495
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Jerome Forissier <jerome.forissier@linaro.org>
Acked-by: Etienne Carriere <etienne.carriere@foss.st.com>

show more ...