- mbed-tls - lists.trustedfirmware.org

by arman_ilmak＠yahoo.com

Hey friends Im trying to do a secure connection between my stm32 board and server. I wrote the code based on GitHub - eziya/STM32F4_HAL_ETH_MBEDTLS: STM32 mbedTLS library testing (SSL/TLS client) that i found. My board is based on stm32h7 series. And im using google to test my app.(ip 142.250.74.196 port 443) This is the debug section: https://aws1.discourse-cdn.com/standard17/uploads/mbed/original/2X/e/ee636c… How should i solve the problem?

3 years

2
2
0 0

Issues with mbedtls_rsa_import_raw + mbedtls_rsa_pkcs1_verify

by m4D

Greetings! I'm having an issue while veryfing signature with imported RSA2048 public key, generated with Win7 CryptoAPI (PUBLICKEYBLOB) into latest mbedtls 2.28.x. The blob contains RSA modulus N (256 bytes) and public exponent E (4 bytes) - I do extract them succesfully, then provide into mbedtls_rsa_import_raw. It all goes smth. like this: u32 errval = mbedtls_rsa_init(ctx, MBEDTLS_RSA_PKCS_V15, 0); // errval == 0 here errval = mbedtls_rsa_import_raw(ctx, n, nlen, 0l, 0, 0l, 0, 0l, 0, e, elen); // errval == 0 here errval = mbedtls_rsa_complete(ctx); // errval == 0 here errval = mbedtls_rsa_check_pubkey(ctx); // errval == 0 here Then i ran: errval = mbedtls_rsa_pkcs1_verify(ctx, 0l, 0l, MBEDTLS_RSA_PUBLIC, MBEDTLS_MD_SHA512, 0, _src, _sign); and get -0x4380 (verify failed) _src - is sha512 hash of data to be verified (64 bytes) _sign - is 256 bytes of signature, provided by win7 cryptoapi P.S. just in case, i did tried messing with endianess in every way for e AND n, it didn't help. I added a little debugging inside library/rsa.cpp, turned out we do call mbedtls_rsa_rsassa_pkcs1_v15_verify, and there is a memcmp between 'encoded' and 'encoded_expected' bufs. 'encoded' is derived from signature (_sign), and 'encoded_expected' is derived from hash (_src) printhex for 'encoded' looks like this: 1a1da83b 14be17a2 c8401d41 1d453909 ... total 16 lines (256 bytes) ... 7fb37ea2 719a5562 aebdb3ed 296e0ed1 but printhex for 'encoded_expected' looks like this: ffff0100 ffffffff ffffffff ffffffff ffffffff ffffffff ffffffff ffffffff ... wtf??? padding ??? ... ffffffff ffffffff ffffffff ffffffff ffffffff ffffffff ffffffff 30513000 6009060d 65014886 03020403 40040005 b190b45a a40b354f 32271b34 f022abd3 ... sha512-derived data here, 64 bytes 557abf2b e2cc4e0f 0b77bdfc b45688b0 So, there is no way these two bufs match. I wonded if there is some issue in parsing _sign, or I didn't prepared input data good enough. Any ideas? B.R., m4D.

3 years

1
0
0 0

How to verify in a server that session cache resumption was used for a given connection

by eoin.mcmahon.dev＠gmail.com

I am trying to modify the dtls_server.c example to keep track of whether session caching was used for a given connection. Ideally I would have an interget value i.e `session_resumed = #1 or 0` One way I tried to do this was by reading the value of the mbedtls_ssl_context struct `ssl`: ``` /* * 5. Handshake */ do ret = mbedtls_ssl_handshake( &ssl ); while( ret == MBEDTLS_ERR_SSL_WANT_READ || ret == MBEDTLS_ERR_SSL_WANT_WRITE ); if( ret == MBEDTLS_ERR_SSL_HELLO_VERIFY_REQUIRED ) { printf( " hello verification requested\n" ); ret = 0; goto reset; } else if( ret != 0 ) { printf( " failed\n ! mbedtls_ssl_handshake returned -0x%x\n\n", (unsigned int) -ret ); goto reset; } printf( " session cache status: %d\n", ssl.handshake.resume ); ``` The issue with this is that the ssl struct is set to private, so the code fails to compile with the error: 'struct mbedtls_ssl_context' has no member named 'handshake' Can somebody help me with some example code that would make this possible?

3 years

3
3
0 0

Mbed TLS GitHub migration

by Dave Rodgman

Hi all, Please note that in the next couple of weeks, we will migrate Mbed TLS to a new GitHub organisation. Your existing scripts, links etc for accessing Mbed TLS on GitHub should not be affected. This will change the url from https://github.com/ARMmbed/mbedtls to https://github.com/Mbed-TLS/mbedtls . GitHub will redirect any accesses to the old URL for the foreseeable future, but we would recommend updating your links once the migration is complete. All of the Mbed TLS repositories will migrate to this new organisation, i.e.: mbedtls mbedtls-docs mbedtls-test Thanks Dave Rodgman

3 years

1
1
0 0

ECDSA Help

by Deep Patel

Hello, mbedTLS has ECDSA module that takes Signature with ASN1 encoding as input mbedtls_pk_verify() The Signature I receive is without ASN1 encoded. Trying to find an implementation within mbedTLS that can add ASN1 to signature before I feed into the verify function. Any help ? [RF IDeas]<http://www.rfideas.com/> Deep Patel Sr. Embedded Software Engineer D: 224-333-2084 P: 847-870-1723 Ext 437 E: ddpatel(a)rfideas.com<mailto:ddpatel@rfideas.com> A: <https://www.rfideas.com/> 425 North Martingale Road, Suite 1680, Schaumburg, IL 60173 <https://www.rfideas.com/>

3 years

2
1
0 0

TrustedFirmware Code of Conduct

by Don Harbin

Hi All, Please find the link to the TrustedFirmware Community Code of Conduct here: https://developer.trustedfirmware.org/w/collaboration/community_guidelines/… Trusted Firmware has a very diverse and global developer community. It is important that we adhere to the code of conduct in all our interactions. For some of you all this may be new and for others just a gentle reminder. In either case, if you have any questions, please feel free to reach out to me directly. And thanks to you all for your contributions to the TrustedFirmware community! Best regards, Don Harbin TrustedFirmware Community Manager don.harbin(a)linaro.org

3 years, 1 month

1
0
0 0

Official release with mutual authentication support

by Tove Rumar

Hello! Is there a time plan for when there will be an official release with TLS 1.3 (Client) that supports mutual authentication? Kind Regards Tove Rumar Software Engineer u-blox Malmo Östra Varvsgatan 4 SE- Malmö www.u-blox.com<https://www.u-blox.com> Reliable. Smart. Secure.

3 years, 1 month

1
0
0 0

AES Memory Leak

by Subramanian Gopi Krishnan

Hi, I have developed a program to send encrypted data continuously. The program is working fine for few hours. After 6 to 8 Hrs, I get returned MBEDTLS_ERR_CIPHER_ALLOC_FAILED -0x6180 Do I need to free the gcm context for every time I send data? What would be the root cause. #define AEAD_KEY_SIZE (32) /**< AEAD Key Size(in bytes). */ #define AEAD_IV_SIZE (12) /**< AEAD IV Size(in bytes). */ #define AEAD_AD_SIZE (16) /**< AEAD AAD Size(in bytes). */ #define AEAD_TAG_SIZE (16) /**< AEAD Tag Size(in bytes). */ #define AEAD_ID_SIZE (4) /**< AEAD Tag Size(in bytes). */ #define MAX_CIPHER_SIZE (2048)/**< AEAD Cipher/Plaintext Segent Size unoperation(in bytes). */ typedef struct { mbedtls_gcm_context CtxAead; /**< aes-gcm 3rd party library context */ uint8_t pu8Key[AEAD_KEY_SIZE]; /**< Key - Randomly Generated */ uint8_t pu8InitVector[AEAD_IV_SIZE]; /**< Randomly Generated */ uint8_t pu8AddData[AEAD_AD_SIZE]; /**< Additional Associated Data - md5sum() */ uint8_t pu8Tag[AEAD_TAG_SIZE]; /**< Hold Digest/Tag of encryped data */ uint8_t pu8Id[AEAD_ID_SIZE]; /**< KeyID is generated */ uint8_t pu8Data[MAX_CIPHER_SIZE]; /**< Buffer to hold plain-text after decipher */ uint8_t pu8Cipher[MAX_CIPHER_SIZE]; /**< Buffer to hold cipher-text after cipher */ uint16_t u16Length; /**< Current length of buffer to encrypt/decrypt */ }S_Aead; int32_t i32Aead_EncryptionCycle(S_Aead *ctx ,uint8_t *pu8AddData, uint16_t u16Length ) { int32_t i32Ret = 0; if(NULL == ctx) { return (NULL_AEAD_CTX); } /* Initilaze GCM Context **/ mbedtls_gcm_init ( &(ctx->CtxAead) ); /*SETTING ADDITIONAL DATA as MD5 of incoming data **/ if( AEAD_AD_SIZE < u16Length ) { return (ERROR_INPUT_BUFFER_TOO_LONG); } memset( ctx->pu8AddData, 0x00, AEAD_AD_SIZE ); i32Ret = mbedtls_md5_ret(pu8AddData, u16Length, ctx->pu8AddData); if (i32Ret) { return (i32Ret); } /* symetric key for testing */ memset( ctx->pu8Key, 0xAB, AEAD_KEY_SIZE ) /* Setting the key */ if( i32Ret = mbedtls_gcm_setkey ( &(ctx->CtxAead), MBEDTLS_CIPHER_ID_AES, ctx->pu8Key, AEAD_KEY_SIZE*OCTETS ) ) { return (i32Ret); } while(1) { /* OS Wait for 1000ms for data to get ready */ if( OS_WaitSingleEventTimed( TX_DATA_BUFFER_READY, WAIT_1000MS ) ) { if( ctx->u16Length == 0 ) { /* No data ready for sending (or) timed-out */; continue; } /* New Initial Vector */ if( i32Ret = i32NewRandomByte( ctx->pu8InitVector, AEAD_IV_SIZE, "AEAD_NEW_IV_GENERATION" ) ) { return(i32Ret); } ctx->u16Length = u16Length; if( i32Ret = mbedtls_gcm_crypt_and_tag( &(ctx->CtxAead), MBEDTLS_GCM_ENCRYPT, ctx->u16Length, ctx->pu8InitVector, AEAD_IV_SIZE, ctx->pu8AddData, AEAD_AD_SIZE, ctx->pu8Data, ctx->pu8Cipher, AEAD_TAG_SIZE, ctx->pu8Tag)) { return(i32Ret); } if( ctx->u16Length == send_data(ctx->pu8Cipher, ctx->u16Length) ) { /* tx success*/ memset( ctx->pu8Data, 0x00, ctx->u16Length ); memset( ctx->pu8Cipher, 0x00, ctx->u16Length ); ctx->u16Length = 0; } } } /* Free GCM Context */ mbedtls_gcm_free( &(ctx->CtxAead) ); /* Reset context memory */ memset(ctx, 0, sizeof(S_Aead)); }

3 years, 1 month

1
0
0 0

MBed TLS Technical Forum - reminder

by Don Harbin

Hi All, A gentle reminder that the US-Europe timezone-friendly MBed TLS Tech forum is next Monday. If you have any topics, please let Dave Rodgman know. :) Best regards, Don Harbin TrustedFirmware Community Manager don.harbin(a)linaro.org

3 years, 1 month

1
0
0 0

signature verification of ECC type

by Sunil Jain

Hello, we are using mbedTLS version 2.16. We are facing a problem in verifying the signed message for ECDSA type of algorithms. Do you have any sample code for this as given for the RSA type algorithm in rsa_verify.c. We have derived the R and S values and their length, but we are not sure which context to use to verify the signature. Please help urgently. -- Regards, Sunil Jain

3 years, 1 month

2
1
0 0

2025

2024

2023

2022

2021

2020

mbed-tls