Topic: WolfSSL sha384 vs OpenSSL sha384
I'm using a simple sha digest to get my feet wet w/ WolfSSL API. I've created a program to exercise the OpenSSL compatibility layer. I can create a sha384 digest w/ OpenSSL, but not w/ Wolf. Simply changing the TARGET_DIGEST to SHA1/NID_sha1 and both libraries generate the same sha.
Here's the setup
WolfSSL 3.9.8
OpenSSL 1.0.1f
Wolf Configure:
./configure --enable-rsa --enable-aes --enable-pkcs7 --enable-examples --enable-ocsp --enable-ocspstapling \
--enable-ocspstapling2 --enable-openssh --enable-opensslextra --enable-keygen --prefix=/home/username/workspace/wolf/x86/install && make all install
#include <stdio.h>
#ifdef OPENSSL
#include <openssl/ssl.h>
#include <openssl/err.h>
#include <openssl/hmac.h>
#else
#include "wolfssl/ssl.h"
#include "wolfssl/openssl/hmac.h"
#endif
/*
g++ -fpermissive -DOPENSSL ssl_client.c -lssl -lcrypto -o ssl && ./ssl
g++ -fpermissive ggdb -I../x86/install/include/ -L../x86/install/lib ssl_client.c -lwolfssl -o wolfssl && LD_LIBRARY_PATH=../x86/install/lib ./wolfssl
*/
void hmac_test()
{
//const EVP_MD *md = EVP_get_digestbyname("SHA256");
#ifdef OPENSSL
#define TARGET_DIGEST "SHA384"
const EVP_MD *md = EVP_get_digestbyname(TARGET_DIGEST);
#else
#define TARGET_DIGEST NID_sha384
const WOLFSSL_EVP_MD *md = wolfSSL_EVP_get_digestbynid(TARGET_DIGEST);
#endif
//const char *key="this_is_the_key";
//const char *msg="This is a test string to hash. We make things.";
const char *key="";
const char *msg="";
int keyLen = strlen(key);
int resultLen = 0;
unsigned char result[512+1];
char r2[512];
char *p2 = r2;
int offset = 0;
HMAC(md, key, keyLen, msg, strlen(msg), &result[0], &resultLen);
for ( int z = 0; z < resultLen; z++ )
{
p2 += sprintf(p2, "%x", result[z]);
}
*p2++ = 0;
printf("Result: %d: %s\n", resultLen, r2);
}
int main(int argc, const char **argv)
{
#ifdef OPENSSL
SSL_library_init(); /* load encryption & hash algorithms for SSL */
SSL_load_error_strings();
ERR_load_BIO_strings();
ERR_load_crypto_strings();
#else
wolfSSL_Init();
#endif
hmac_test();
}Segfault from WolfSSL:
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Program received signal SIGSEGV, Segmentation fault.
0x00007ffff7bacb41 in wolfSSL_HMAC () from ../x86/install/lib/libwolfssl.so.3
(gdb) bt
#0 0x00007ffff7bacb41 in wolfSSL_HMAC () from ../x86/install/lib/libwolfssl.so.3
#1 0x000000000040094c in hmac_test () at ssl_client.c:78
#2 0x0000000000400a04 in main (argc=1, argv=0x7fffffffdd28) at ssl_client.c:100
(gdb) quit
A debugging session is active.
It seems reasonable this should work. I have not tried to create a sha384 using native WolfAPI, but some of our assumptions <ahem> are that a lot of the effort to port over to Wolf will be simplified w/ the OpenSSL API layer. What am I missing?