Re: [dpdk-dev] [dpdk-dev v2 1/2] fips_validation: add SGL support

["Zhang, Roy Fan" <roy.fan.zhang@intel.com>]

Hi Suanming Mou,

Thanks for the review.

> -----Original Message-----
> From: Suanming Mou <suanmingm@nvidia.com>
> Sent: Monday, September 7, 2020 2:32 PM
> To: Zhang, Roy Fan <roy.fan.zhang@intel.com>; dev@dpdk.org
> Cc: akhil.goyal@nxp.com; Trahe, Fiona <fiona.trahe@intel.com>; Kusztal,
> ArkadiuszX <arkadiuszx.kusztal@intel.com>; Dybkowski, AdamX
> <adamx.dybkowski@intel.com>
> Subject: Re: [dpdk-dev] [dpdk-dev v2 1/2] fips_validation: add SGL support
> 
> Hi,
> 
...
> > @@ -50,8 +55,10 @@ cryptodev_fips_validate_app_int(void)
> >   {
> >   	struct rte_cryptodev_config conf = {rte_socket_id(), 1, 0};
> >   	struct rte_cryptodev_qp_conf qp_conf = {128, NULL, NULL};
> > +	struct rte_cryptodev_info dev_info;
> Better to initialize the dev_info here?

[Fan] dev_info is initialized by rte_cryptodev_info_get() after the device is
successfully configured. So at least we can be sure the device ID is a valid
one and the buffer will be properly set by the driver. Plus the
the implementation of rte_cryptodev_info_get() will do a memset to
the dev_info buffer so it is not necessary to be initialized there.

> >   	uint32_t sess_sz = rte_cryptodev_sym_get_private_session_size(
> >   			env.dev_id);
> > +	uint32_t nb_mbufs = UINT16_MAX / env.mbuf_data_room + 1;
> >   	int ret;
> >

Regards,
Fan