bump version to 0.9.6+pve1

[libtpms.git] / CHANGES

CHANGES - changes for libtpms

version 0.9.6:
  - tpm2: Check size of buffer before accessing it (CVE-2023-1017 & -1018)

version 0.9.5:
  - tpm2: Do not set RSA_FLAG_NO_BLINDING on RSA keys anymore
  - tpm2: Fix a potential overflow expression (coverity)
  - tpm2: Fix size check in CryptSecretDecrypt

version 0.9.4:
  - tpm: #undef printf in case it is #define'd (OSS-Fuzz)
  - tpm2: Check return code of BN_div()
  - tpm2: Initialize variables due to gcc complaint (s390x, false positive)
  - tpm12: Initialize variables due to gcc complaint (s390x, false positive)
  - build-sys: Fix configure script to support _FORTIFY_SOURCE=3

version 0.9.3:
  - build-sys: Add probing for -fstack-protector
  - tpm2: Do not call EVP_PKEY_CTX_set0_rsa_oaep_label() for label of size
    (OSSL 3)

version 0.9.2:
  - tpm2: When writing state initialize s_ContextSlotMask if not set

version 0.9.1:
  - tpm2: Do not write permanent state if only clock changed
  - tpm2: Fix "maybe-uninitialized" warning

version 0.9.0:
  - NOTE: Downgrade to previous versions is not possible. See below.
  - The size of the context gap has been adjusted to 0xffff from 0xff.
    As a consequence of this the volatile state's format (STATE_RESET_DATA)
    has changed and cannot be downgraded.
  - Applied work-around for Win 2016 & 2019 server related to
    TPM2_ContextLoad (issue #217)
  - Check for several more compile-time constants
  - Enabled Camellia symmetric key encryption algorithm
  - tpm2: CryptSym: fix AES output IV
  - tpm2: Added a cache for private exponent D and prime Q
  - tpm2: bug fixes related to state marshalling
  - tpm2: Consume padding bytes in TPM2_ContextLoad() (Win2k19, issue #217)
  - tests: Improvements on the fuzzer
  - tpm2: Switch to UINT16 for CONTEXT_SLOT and 64k context gap
  - tpm2: Update to TPM 2 spec rev 164
  - build-sys: Enable building --without-tpm1
  - tpm2: Marshal event sequence objects' hash state
  - tpm2: Fixes for build and runtime when using OpenSSL 3.0

version 0.8.0
  - NOTE: Downgrade to previous versions is not possible. See below.
  - Update to TPM 2 code release 159
    - X509 support is enabled
      - SM2 signing of ceritificates is NOT supported
    - Authenticated timers are disabled
  - Due to fixes in the TPM 2 prime number generation code in rev155 it is not
    possible to downgrade from libtpms version 0.8.0 to some previous version.
    The seeds are now associated with an age so that older seeds use the old
    TPM 2 prime number generation code while newer seed use the newer code.
  - Update to TPM 2 code release 162
    - ECC encryption / decryption is disabled
  - Fix support for elliptic curve due to missing unmarshalling code
  - Runtime filter supported elliptic curves supported by OpenSSL
  - Fix output buffer parameter and size for RSA decryption that could cause
    stack corruption under certain circumstances
  - Set the RSA PSS salt length to the digest length rather than max. possible
  - Fixes to symmetric decryption related to input size check,
    defer padding to the user [EVP_CIPHER_CTX_set_padding(ctx, 0)] and
    to always use a temporary malloc'ed buffer for decryption
  - Fixed the set of PCRs belonging to the TCB group. This affects the
    pcrUpdateCounter in TPM2_Pcrread() responses, thus needs latest `swtpm`
    for test cases to succeed there.

version 0.7.0
  - use OpenSSL crypto for AES, TDES, EC, and RSA operations when possible

version 0.6.0
  - added TPM 2 support (revision 150)

  - New API calls:
    - TPMLIB_CancelCommand
    - TPMLIB_ChooseTPMVersion
    - TPMLIB_SetDebugFD
    - TPMLIB_SetDebugLevel
    - TPMLIB_SetDebugPrefix
    - TPMLIB_SetBufferSize
    - TPMLIB_ValidateState
    - TPMLIB_SetState
    - TPMLIB_GetState

version 0.5.1
  first public release

  - release 7 increased NVRAM area for being able to store more data in
    the TPM's NVRAM areas, i.e., X.509 certificates

  - release 9 added two more APIs:
    - TPM_Free
    - TPMLIB_DecodeBlob