Here is a full code sample which should work:

int main(void) {
    int i;
    int j;
    byte input[] = {255};
    byte hash[48];
    wc_Sha384 sha384;
    for (j = 0; j < 10; j++) {
        wc_InitSha384(&sha384);
        wc_Sha384Update(&sha384, input, sizeof(input));
        wc_Sha384Final(&sha384, hash);
        printf("Buffer: \n");
        for (i = 0; i < 48; i++)
            printf("%02x ", hash[i]);
        printf("\n");
    }
    return 0; 
}

Yes you are correct sizeof(input); is correct

No hardware acceleration. I'm using wolfssl-5.7.4-gplv3-fips-ready with the fipsv5 user_settings.h

Are you using any kind of hardware acceleration for SHA384?  Can you share your version and build settings?

Thanks for the followup.  Looks like you are initializing a different variable "sha" here: wc_InitSha384(&sha);
sha384 must be initialized before you can use it, please try replacing this line with: wc_InitSha384(&sha384);

Thanks,
Kareem

byte input[] = {255};
byte hash[48];
wc_Sha384 sha384;
wc_InitSha384(&sha);
wc_Sha384Update(&sha384, input, sizeof(data));
wc_Sha384Final(&sha384, hash);

and then I'm checking the first 3 indices of hash against some hardcoded values.

That is strange, I would definitely expect the same input to create the same output from SHA384.  Please share your code.

We don't have a SHA384 example but we do have a SHA512 example, does your code follow the same general flow?  https://github.com/wolfSSL/wolfssl-exam … 512-hash.c

Thanks,
Kareem