/********************************************************************************/ /* */ /* Math functions performed with canonical integers in byte buffers */ /* Written by Ken Goldman */ /* IBM Thomas J. Watson Research Center */ /* */ /* Licenses and Notices */ /* */ /* 1. Copyright Licenses: */ /* */ /* - Trusted Computing Group (TCG) grants to the user of the source code in */ /* this specification (the "Source Code") a worldwide, irrevocable, */ /* nonexclusive, royalty free, copyright license to reproduce, create */ /* derivative works, distribute, display and perform the Source Code and */ /* derivative works thereof, and to grant others the rights granted herein. */ /* */ /* - The TCG grants to the user of the other parts of the specification */ /* (other than the Source Code) the rights to reproduce, distribute, */ /* display, and perform the specification solely for the purpose of */ /* developing products based on such documents. */ /* */ /* 2. Source Code Distribution Conditions: */ /* */ /* - Redistributions of Source Code must retain the above copyright licenses, */ /* this list of conditions and the following disclaimers. */ /* */ /* - Redistributions in binary form must reproduce the above copyright */ /* licenses, this list of conditions and the following disclaimers in the */ /* documentation and/or other materials provided with the distribution. */ /* */ /* 3. Disclaimers: */ /* */ /* - THE COPYRIGHT LICENSES SET FORTH ABOVE DO NOT REPRESENT ANY FORM OF */ /* LICENSE OR WAIVER, EXPRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE, WITH */ /* RESPECT TO PATENT RIGHTS HELD BY TCG MEMBERS (OR OTHER THIRD PARTIES) */ /* THAT MAY BE NECESSARY TO IMPLEMENT THIS SPECIFICATION OR OTHERWISE. */ /* Contact TCG Administration (admin@trustedcomputinggroup.org) for */ /* information on specification licensing rights available through TCG */ /* membership agreements. */ /* */ /* - THIS SPECIFICATION IS PROVIDED "AS IS" WITH NO EXPRESS OR IMPLIED */ /* WARRANTIES WHATSOEVER, INCLUDING ANY WARRANTY OF MERCHANTABILITY OR */ /* FITNESS FOR A PARTICULAR PURPOSE, ACCURACY, COMPLETENESS, OR */ /* NONINFRINGEMENT OF INTELLECTUAL PROPERTY RIGHTS, OR ANY WARRANTY */ /* OTHERWISE ARISING OUT OF ANY PROPOSAL, SPECIFICATION OR SAMPLE. */ /* */ /* - Without limitation, TCG and its members and licensors disclaim all */ /* liability, including liability for infringement of any proprietary */ /* rights, relating to use of information in this specification and to the */ /* implementation of this specification, and TCG disclaims all liability for */ /* cost of procurement of substitute goods or services, lost profits, loss */ /* of use, loss of data or any incidental, consequential, direct, indirect, */ /* or special damages, whether under contract, tort, warranty or otherwise, */ /* arising in any way out of use or reliance upon this specification or any */ /* information herein. */ /* */ /* (c) Copyright IBM Corp. and others, 2016 - 2023 */ /* */ /********************************************************************************/ /*(Auto-generated) * Created by TpmPrototypes; Version 3.0 July 18, 2017 * Date: Mar 28, 2019 Time: 08:25:19PM */ #ifndef _MATH_ON_BYTE_BUFFERS_FP_H_ #define _MATH_ON_BYTE_BUFFERS_FP_H_ //*** UnsignedCmpB // This function compare two unsigned values. The values are byte-aligned, // big-endian numbers (e.g, a hash). // Return Type: int // 1 if (a > b) // 0 if (a = b) // -1 if (a < b) LIB_EXPORT int UnsignedCompareB(UINT32 aSize, // IN: size of a const BYTE* a, // IN: a UINT32 bSize, // IN: size of b const BYTE* b // IN: b ); //***SignedCompareB() // Compare two signed integers: // Return Type: int // 1 if a > b // 0 if a = b // -1 if a < b int SignedCompareB(const UINT32 aSize, // IN: size of a const BYTE* a, // IN: a buffer const UINT32 bSize, // IN: size of b const BYTE* b // IN: b buffer ); //*** ModExpB // This function is used to do modular exponentiation in support of RSA. // The most typical uses are: 'c' = 'm'^'e' mod 'n' (RSA encrypt) and // 'm' = 'c'^'d' mod 'n' (RSA decrypt). When doing decryption, the 'e' parameter // of the function will contain the private exponent 'd' instead of the public // exponent 'e'. // // If the results will not fit in the provided buffer, // an error is returned (CRYPT_ERROR_UNDERFLOW). If the results is smaller // than the buffer, the results is de-normalized. // // This version is intended for use with RSA and requires that 'm' be // less than 'n'. // // Return Type: TPM_RC // TPM_RC_SIZE number to exponentiate is larger than the modulus // TPM_RC_NO_RESULT result will not fit into the provided buffer // TPM_RC ModExpB(UINT32 cSize, // IN: the size of the output buffer. It will // need to be the same size as the modulus BYTE* c, // OUT: the buffer to receive the results // (c->size must be set to the maximum size // for the returned value) const UINT32 mSize, const BYTE* m, // IN: number to exponentiate const UINT32 eSize, const BYTE* e, // IN: power const UINT32 nSize, const BYTE* n // IN: modulus ); //*** DivideB() // Divide an integer ('n') by an integer ('d') producing a quotient ('q') and // a remainder ('r'). If 'q' or 'r' is not needed, then the pointer to them // may be set to NULL. // // Return Type: TPM_RC // TPM_RC_NO_RESULT 'q' or 'r' is too small to receive the result // LIB_EXPORT TPM_RC DivideB(const TPM2B* n, // IN: numerator const TPM2B* d, // IN: denominator TPM2B* q, // OUT: quotient TPM2B* r // OUT: remainder ); //*** AdjustNumberB() // Remove/add leading zeros from a number in a TPM2B. Will try to make the number // by adding or removing leading zeros. If the number is larger than the requested // size, it will make the number as small as possible. Setting 'requestedSize' to // zero is equivalent to requesting that the number be normalized. UINT16 AdjustNumberB(TPM2B* num, UINT16 requestedSize); //*** ShiftLeft() // This function shifts a byte buffer (a TPM2B) one byte to the left. That is, // the most significant bit of the most significant byte is lost. TPM2B* ShiftLeft(TPM2B* value // IN/OUT: value to shift and shifted value out ); #endif // _MATH_ON_BYTE_BUFFERS_FP_H_