Re: [dpdk-dev] [PATCH v2 3/6] lib/cryptodev: add asymmetric crypto capability in cryptodev

["Verma, Shally" <Shally.Verma@cavium.com>]

>-----Original Message-----
>From: Trahe, Fiona [mailto:fiona.trahe@intel.com]
>Sent: 08 May 2018 20:30
>To: Verma, Shally <Shally.Verma@cavium.com>; De Lara Guarch, Pablo <pablo.de.lara.guarch@intel.com>
>Cc: Doherty, Declan <declan.doherty@intel.com>; Athreya, Narayana Prasad <NarayanaPrasad.Athreya@cavium.com>; Sahu, Sunila
><Sunila.Sahu@cavium.com>; Gupta, Ashish <Ashish.Gupta@cavium.com>; dev@dpdk.org; Sahu, Sunila <Sunila.Sahu@cavium.com>;
>Gupta, Ashish <Ashish.Gupta@cavium.com>
>Subject: RE: [PATCH v2 3/6] lib/cryptodev: add asymmetric crypto capability in cryptodev
>
>Hi Shally,
>
>> -----Original Message-----
>> From: Shally Verma [mailto:shally.verma@caviumnetworks.com]
>> Sent: Thursday, April 5, 2018 12:25 PM
>> To: De Lara Guarch, Pablo <pablo.de.lara.guarch@intel.com>
>> Cc: Doherty, Declan <declan.doherty@intel.com>; Trahe, Fiona <fiona.trahe@intel.com>;
>> pathreya@caviumnetworks.com; ssahu@caviumnetworks.com; agupta@caviumnetworks.com;
>> dev@dpdk.org; Sunila Sahu <sunila.sahu@caviumnetworks.com>; Ashish Gupta
>> <ashish.gupta@caviumnetworks.com>
>> Subject: [PATCH v2 3/6] lib/cryptodev: add asymmetric crypto capability in cryptodev
>>
>> Extend cryptodev with asymmetric capability APIs and
>> definitions.
>>
>> Signed-off-by: Shally Verma <shally.verma@caviumnetworks.com>
>> Signed-off-by: Sunila Sahu <sunila.sahu@caviumnetworks.com>
>> Signed-off-by: Ashish Gupta <ashish.gupta@caviumnetworks.com>
>>
>> ---
>/// snip ///
>> +int __rte_experimental
>> +rte_cryptodev_asym_xfrm_capability_check_modlen(
>> +	const struct rte_cryptodev_asymmetric_xfrm_capability *capability,
>> +	uint16_t modlen)
>> +{
>> +	/* handle special case of 0 which mean PMD define no limit defined */
>[Fiona] grammar. Maybe "which means PMD doesn't define any limit"
>
>> +	if ((capability->modlen.min != 0) &&
>> +		((modlen < capability->modlen.min) ||
>> +		(capability->modlen.increment != 0 &&
>> +		(modlen % (capability->modlen.increment)))))
>> +		return -1;
>> +	if ((capability->modlen.max != 0) &&
>> +		((modlen > capability->modlen.max) ||
>> +		(capability->modlen.increment != 0 &&
>> +		(modlen % (capability->modlen.increment)))))
>> +		return -1;
>> +
>> +	return 0;
>> +}
>> +
>>
>>  const char *
>>  rte_cryptodev_get_feature_name(uint64_t flag)
>> diff --git a/lib/librte_cryptodev/rte_cryptodev.h b/lib/librte_cryptodev/rte_cryptodev.h
>> index 68d1ae1..deae3d6 100644
>> --- a/lib/librte_cryptodev/rte_cryptodev.h
>> +++ b/lib/librte_cryptodev/rte_cryptodev.h
>> @@ -178,6 +178,37 @@ struct rte_cryptodev_symmetric_capability {
>>  	};
>>  };
>>
>> +/**
>> + * Asymmetric Xform Crypto Capability
>> + *
>> + */
>> +struct rte_cryptodev_asymmetric_xfrm_capability {
>> +	enum rte_crypto_asym_xform_type xform_type;
>> +	/**< Transform type: RSA/MODEXP/DH/DSA/MODINV */
>> +
>> +	uint32_t op_types;
>> +	/**< bitmask for supported rte_crypto_asym_op_type */
>> +
>> +	__extension__
>> +	union {
>> +		struct rte_crypto_param_range modlen;
>> +		/**< Range of modulus length supported by modulus based xform.
>> +		 * Value 0 mean implementation default
>> +		 */
>> +	};
>> +};
>> +
>> +/**
>> + * Asymmetric Crypto Capability
>> + *
>> + */
>> +struct rte_cryptodev_asymmetric_capability {
>> +	enum rte_crypto_asym_xform_type xform_type;
>> +	/**< Transform type: RSA/MODEXP/DH/DSA/MODINV */
>> +	struct rte_cryptodev_asymmetric_xfrm_capability xfrm_capa;
>> +};
>[Fiona] Is it necessary to have xform_type in both above structures?
>Seems like duplication. Or would it be better if both are combined into 1 struct?
>
[Shally] Ok.
>> +
>> +
>>  /** Structure used to capture a capability of a crypto device */
>>  struct rte_cryptodev_capabilities {
>>  	enum rte_crypto_op_type op;
>> @@ -187,6 +218,8 @@ struct rte_cryptodev_capabilities {
>>  	union {
>>  		struct rte_cryptodev_symmetric_capability sym;
>>  		/**< Symmetric operation capability parameters */
>> +		struct rte_cryptodev_asymmetric_capability asym;
>> +		/**< Asymmetric operation capability parameters */
>>  	};
>>  };
>/// snip ///