Wrong SHA256 hash value from hashcrypt

TestNXP — Tue, 17 Oct 2023 15:17:53 GMT

Hello,

when I am calculating a SHA256 hash via hashcrypt functions for messages longer than 120000 Bytes I get wrong hash results (see screenshot). You can see the python results on the left and the NXP MCU calculated values on the right.

I am caluclating the hash for byte arrays filled with "0xAB".

Is there some limitation to the message size that can be hashed by the hashcrypt periphery?

I am using a LPC55S28 MCU. I am testing this with following function based on the hashcrypt example:

void TestSha256(void) { static uint8_t message[MESSAGE_SIZE]; for (int i = 0; i < MESSAGE_SIZE; i++) { message[i] = 0xab; } status_t status; size_t outLength; unsigned int length; unsigned char output[32]; length = sizeof(message); outLength = sizeof(output); memset(&output, 0, outLength); /************************ SHA-256 **************************/ status = HASHCRYPT_SHA(HASHCRYPT, kHASHCRYPT_Sha256, message, 100000, output, &outLength); TEST_ASSERT(kStatus_Success == status); TEST_ASSERT(outLength == 32u); PRINTF("100kiB: "); for (int i = 0; i < 32; i++) { PRINTF("%x", output[i]); } PRINTF("\n"); status = HASHCRYPT_SHA(HASHCRYPT, kHASHCRYPT_Sha256, message, 120000, output, &outLength); TEST_ASSERT(kStatus_Success == status); TEST_ASSERT(outLength == 32u); PRINTF("120kiB: "); for (int i = 0; i < 32; i++) { PRINTF("%x", output[i]); } PRINTF("\n"); status = HASHCRYPT_SHA(HASHCRYPT, kHASHCRYPT_Sha256, message, 130000, output, &outLength); TEST_ASSERT(kStatus_Success == status); TEST_ASSERT(outLength == 32u); PRINTF("130kiB: "); for (int i = 0; i < 32; i++) { PRINTF("%x", output[i]); } PRINTF("\n"); status = HASHCRYPT_SHA(HASHCRYPT, kHASHCRYPT_Sha256, message, 140000, output, &outLength); TEST_ASSERT(kStatus_Success == status); TEST_ASSERT(outLength == 32u); PRINTF("140kiB: "); for (int i = 0; i < 32; i++) { PRINTF("%x", output[i]); } PRINTF("\n"); status = HASHCRYPT_SHA(HASHCRYPT, kHASHCRYPT_Sha256, message, 150000, output, &outLength); TEST_ASSERT(kStatus_Success == status); TEST_ASSERT(outLength == 32u); PRINTF("150kiB: "); for (int i = 0; i < 32; i++) { PRINTF("%x", output[i]); } PRINTF("\n"); }

Re: Wrong SHA256 hash value from hashcrypt

Alice_Yang — Wed, 18 Oct 2023 08:53:07 GMT

Hello @TestNXP

Please use latest version SDK v2.14.0.

Alice

Re: Wrong SHA256 hash value from hashcrypt

TestNXP — Fri, 20 Oct 2023 07:01:24 GMT

Hello @Alice_Yang

is there a difference according this problem between v2.13.0 and v.2.14.0. ?

I did also get the information that hardware hashing is limited to 2047 blocks each 512 Bit which does indeed result to ~130000 Bytes. Thats also the point where hashcrypt generates wrong hash values for me. (See screenshot from LPC55xx user guide)

Can you tell me where this limitation comes from? From my understanding hashing should be possible for as big data as you want.

LPC MicrocontrollersのトピックRe: Wrong SHA256 hash value from hashcrypt

Wrong SHA256 hash value from hashcrypt

Re: Wrong SHA256 hash value from hashcrypt

Re: Wrong SHA256 hash value from hashcrypt