Re: [PATCH] net/tls: support SM4 GCM/CCM algorithm

[Tianjia Zhang <tianjia.zhang@linux.alibaba.com>]

Hi Jakub,

On 9/16/21 4:06 AM, Jakub Kicinski wrote:
> On Wed, 15 Sep 2021 19:12:42 +0800 Tianjia Zhang wrote:
>> +		memcpy(sm4_gcm_info->iv,
>> +		       cctx->iv + TLS_CIPHER_AES_GCM_128_SALT_SIZE,
>> +		       TLS_CIPHER_AES_GCM_128_IV_SIZE);
>> +		memcpy(sm4_gcm_info->rec_seq, cctx->rec_seq,
>> +		       TLS_CIPHER_AES_GCM_128_REC_SEQ_SIZE);
>> +		release_sock(sk);
>> +		if (copy_to_user(optval, sm4_gcm_info, sizeof(*sm4_gcm_info)))
>> +			rc = -EFAULT;
>> +		break;
>> +	}
>> +	case TLS_CIPHER_SM4_CCM: {
>> +		struct tls12_crypto_info_sm4_ccm *sm4_ccm_info =
>> +			container_of(crypto_info,
>> +				struct tls12_crypto_info_sm4_ccm, info);
>> +
>> +		if (len != sizeof(*sm4_ccm_info)) {
>> +			rc = -EINVAL;
>> +			goto out;
>> +		}
>> +		lock_sock(sk);
>> +		memcpy(sm4_ccm_info->iv,
>> +		       cctx->iv + TLS_CIPHER_AES_GCM_128_SALT_SIZE,
>> +		       TLS_CIPHER_AES_GCM_128_IV_SIZE);
>> +		memcpy(sm4_ccm_info->rec_seq, cctx->rec_seq,
>> +		       TLS_CIPHER_AES_GCM_128_REC_SEQ_SIZE);
> 
> Doesn't matter from the functional perspective but perhaps use the SM4
> defines rather than the AES ones, since they exist, anyway?
> 
> With that fixed feel free to add my ack.
> 

Thanks for pointing it out, I forgot to modify the macro name, this is 
not my intention, I was careless. will fix it in v2.

Best regards,
Tianjia