[dpdk-dev] [PATCH 1/3] test/event_crypto_adapter: return error with unsupported mode

[Gujjar, Abhinandan S]

Hi Ankur,

> -----Original Message-----
> From: Ankur Dwivedi <adwivedi at marvell.com>
> Sent: Wednesday, September 23, 2020 11:02 AM
> To: Gujjar, Abhinandan S <abhinandan.gujjar at intel.com>; dev at dpdk.org
> Cc: Jerin Jacob Kollanukkaran <jerinj at marvell.com>; Anoob Joseph
> <anoobj at marvell.com>
> Subject: RE: [PATCH 1/3] test/event_crypto_adapter: return error with
> unsupported mode
> 
> Hi Abhinandan,
> 
> Please see my comments inline.
> 
> Thanks
> Ankur
> 
> >-----Original Message-----
> >From: Ankur Dwivedi
> >Sent: Friday, September 18, 2020 8:34 PM
> >To: 'Gujjar, Abhinandan S' <abhinandan.gujjar at intel.com>; dev at dpdk.org
> >Cc: Jerin Jacob Kollanukkaran <jerinj at marvell.com>; Anoob Joseph
> ><anoobj at marvell.com>
> >Subject: RE: [PATCH 1/3] test/event_crypto_adapter: return error with
> >unsupported mode
> >
> >Hi Abhinandan,
> >
> >Please see my comments inline.
> >
> >Thanks
> >Ankur
> >
> >>-----Original Message-----
> >>From: Gujjar, Abhinandan S <abhinandan.gujjar at intel.com>
> >>Sent: Friday, September 18, 2020 8:13 PM
> >>To: Ankur Dwivedi <adwivedi at marvell.com>; dev at dpdk.org
> >>Cc: Jerin Jacob Kollanukkaran <jerinj at marvell.com>; Anoob Joseph
> >><anoobj at marvell.com>
> >>Subject: [EXT] RE: [PATCH 1/3] test/event_crypto_adapter: return error
> >>with unsupported mode
> >>
> >>External Email
> >>
> >>----------------------------------------------------------------------
> >>Hi Ankur,
> >>
> >>Please see comments inline.
> >>
> >>Regards
> >>Abhinandan
> >>
> >>> -----Original Message-----
> >>> From: Ankur Dwivedi <adwivedi at marvell.com>
> >>> Sent: Friday, September 18, 2020 8:05 PM
> >>> To: Gujjar, Abhinandan S <abhinandan.gujjar at intel.com>;
> dev at dpdk.org
> >>> Cc: Jerin Jacob Kollanukkaran <jerinj at marvell.com>; Anoob Joseph
> >>> <anoobj at marvell.com>
> >>> Subject: RE: [PATCH 1/3] test/event_crypto_adapter: return error
> >>> with unsupported mode
> >>>
> >>> Hi Abhinandan,
> >>>
> >>> Please see my comments inline.
> >>>
> >>> Thanks
> >>> Ankur
> >>>
> >>> >-----Original Message-----
> >>> >From: Gujjar, Abhinandan S <abhinandan.gujjar at intel.com>
> >>> >Sent: Friday, September 18, 2020 12:41 PM
> >>> >To: Ankur Dwivedi <adwivedi at marvell.com>; dev at dpdk.org
> >>> >Cc: Jerin Jacob Kollanukkaran <jerinj at marvell.com>; Anoob Joseph
> >>> ><anoobj at marvell.com>
> >>> >Subject: [EXT] RE: [PATCH 1/3] test/event_crypto_adapter: return
> >>> >error with unsupported mode
> >>> >
> >>> >External Email
> >>> >
> >>> >-------------------------------------------------------------------
> >>> >-
> >>> >-
> >>> >-
> >>> >Hi Ankur,
> >>> >
> >>> >This patch requires minor change. Please find the comment inline.
> >>> >
> >>> >> -----Original Message-----
> >>> >> From: Gujjar, Abhinandan S
> >>> >> Sent: Friday, September 11, 2020 1:07 PM
> >>> >> To: Ankur Dwivedi <adwivedi at marvell.com>; dev at dpdk.org
> >>> >> Cc: jerinj at marvell.com; anoobj at marvell.com
> >>> >> Subject: RE: [PATCH 1/3] test/event_crypto_adapter: return error
> >>> >> with unsupported mode
> >>> >>
> >>> >> Acked-by: abhinandan.gujjar at intel.com (For series)
> >>> >>
> >>> >> > -----Original Message-----
> >>> >> > From: Ankur Dwivedi <adwivedi at marvell.com>
> >>> >> > Sent: Tuesday, September 8, 2020 2:46 PM
> >>> >> > To: dev at dpdk.org
> >>> >> > Cc: Gujjar, Abhinandan S <abhinandan.gujjar at intel.com>;
> >>> >> > jerinj at marvell.com; anoobj at marvell.com; Ankur Dwivedi
> >>> >> > <adwivedi at marvell.com>
> >>> >> > Subject: [PATCH 1/3] test/event_crypto_adapter: return error
> >>> >> > with unsupported mode
> >>> >> >
> >>> >> > The capability of a event device should be checked before
> >>> >> > creating a event crypto adapter in a particular mode. The test
> >>> >> > case returns error if the mode is not supported.
> >>> >> >
> >>> >> > Signed-off-by: Ankur Dwivedi <adwivedi at marvell.com>
> >>> >> > ---
> >>> >> >  app/test/test_event_crypto_adapter.c | 30
> >>> >> > ++++++++++++++++++----------
> >>> >> >  1 file changed, 20 insertions(+), 10 deletions(-)
> >>> >> >
> >>> >> > diff --git a/app/test/test_event_crypto_adapter.c
> >>> >> > b/app/test/test_event_crypto_adapter.c
> >>> >> > index 8d42462d8..930c2a9bc 100644
> >>> >> > --- a/app/test/test_event_crypto_adapter.c
> >>> >> > +++ b/app/test/test_event_crypto_adapter.c
> >>> >> > @@ -750,15 +750,23 @@ configure_event_crypto_adapter(enum
> >>> >> > rte_event_crypto_adapter_mode mode)
> >>> >> >  	uint32_t cap;
> >>> >> >  	int ret;
> >>> >> >
> >>> >> > +	ret = rte_event_crypto_adapter_caps_get(evdev,
> >TEST_CDEV_ID,
> >>> >> > &cap);
> >>> >> > +	TEST_ASSERT_SUCCESS(ret, "Failed to get adapter
> >>> >> > +capabilities\n");
> >>> >> > +
> >>> >> > +	if ((mode == RTE_EVENT_CRYPTO_ADAPTER_OP_FORWARD)
> >&&
> >>> >> > +	    !(cap &
> >>> >> > RTE_EVENT_CRYPTO_ADAPTER_CAP_INTERNAL_PORT_OP_FWD))
> >>> >> > +		return -ENOTSUP;
> >>> >> > +
> >>> >> > +	if ((mode == RTE_EVENT_CRYPTO_ADAPTER_OP_NEW) &&
> >>> >> > +	    !(cap &
> >>> >> > RTE_EVENT_CRYPTO_ADAPTER_CAP_INTERNAL_PORT_OP_NEW))
> >>> >> > +		return -ENOTSUP;
> >>> >> > +
> >>> >SW adapter does not have *_INTERNAL_PORT_* caps. Just noticed test
> >>> >was failing here.
> >>> >Could you please bypass this check for "event_sw" and repost the
> patch?
> >>> [Ankur] I was thinking of handling this by reading
> >>> rte_eventdevs[evdev].data->name.
> >>> If the name matches event_sw0(which is the default device name), I
> >>> will bypass if checks.
> >>> Can the event_sw device name be anything else?
> >>> Or, if there is a better way of handling this please let me know.
> >>Yes. You can also use rte_event_dev_info_get(evdev, &info) and check
> >>info.driver_name for "event_sw".
> >[Ankur] Thanks for the suggestion. I will try this and make the
> >required changes in v2.
> [Ankur] I thought of an alternative to implement the above. If the capability
> does not have any *_INTERNAL_PORT_* caps, it will be a SW PMD. Is this
> correct?
> In that case capability will be used to skip the if checks instead of the pmd
> specific name comparison. Please let me know your thoughts on this.
Right now, SW has only " RTE_EVENT_CRYPTO_ADAPTER_CAP_SESSION_PRIVATE_DATA " as caps.
See, if you can use this info and continue to execute the test without returning error.
> >>
> >>> >
> >>> >> >  	/* Create adapter with default port creation callback */
> >>> >> >  	ret = rte_event_crypto_adapter_create(TEST_ADAPTER_ID,
> >>> >> > -					      TEST_CDEV_ID,
> >>> >> > +					      evdev,
> >>> >> >  					      &conf, mode);
> >>> >> >  	TEST_ASSERT_SUCCESS(ret, "Failed to create event crypto
> >>> >> adapter\n");
> >>> >> >
> >>> >> > -	ret =
> rte_event_crypto_adapter_caps_get(TEST_ADAPTER_ID,
> >>> >> > evdev, &cap);
> >>> >> > -	TEST_ASSERT_SUCCESS(ret, "Failed to get adapter
> >capabilities\n");
> >>> >> > -
> >>> >> >  	if (cap &
> >>> >> >
> RTE_EVENT_CRYPTO_ADAPTER_CAP_INTERNAL_PORT_QP_EV_BIND) {
> >>> >> >  		ret =
> >>> >> > rte_event_crypto_adapter_queue_pair_add(TEST_ADAPTER_ID,
> >>> >> >  				TEST_CDEV_ID, TEST_CDEV_QP_ID,
> >>> &response_info); @@ -813,6
> >>> >> > +821,8 @@ test_crypto_adapter_conf(enum
> >>> >> > rte_event_crypto_adapter_mode mode)
> >>> >> >  			TEST_ASSERT(ret >= 0, "Failed to link queue
> >%d "
> >>> >> >  					"port=%u\n", qid,
> >>> >> >
> >	params.crypto_event_port_id);
> >>> >> > +		} else {
> >>> >> > +			return ret;
> >>> >> >  		}
> >>> >> >  		crypto_adapter_setup_done = 1;
> >>> >> >  	}
> >>> >> > @@ -845,24 +855,24 @@ static int
> >>> >> >  test_crypto_adapter_conf_op_forward_mode(void)
> >>> >> >  {
> >>> >> >  	enum rte_event_crypto_adapter_mode mode;
> >>> >> > +	int ret;
> >>> >> >
> >>> >> >  	mode = RTE_EVENT_CRYPTO_ADAPTER_OP_FORWARD;
> >>> >> > -	TEST_ASSERT_SUCCESS(test_crypto_adapter_conf(mode),
> >>> >> > -				"Failed to config crypto adapter");
> >>> >> > +	ret = test_crypto_adapter_conf(mode);
> >>> >> >
> >>> >> > -	return TEST_SUCCESS;
> >>> >> > +	return ret;
> >>> >> >  }
> >>> >> >
> >>> >> >  static int
> >>> >> >  test_crypto_adapter_conf_op_new_mode(void)
> >>> >> >  {
> >>> >> >  	enum rte_event_crypto_adapter_mode mode;
> >>> >> > +	int ret;
> >>> >> >
> >>> >> >  	mode = RTE_EVENT_CRYPTO_ADAPTER_OP_NEW;
> >>> >> > -	TEST_ASSERT_SUCCESS(test_crypto_adapter_conf(mode),
> >>> >> > -				"Failed to config crypto adapter");
> >>> >> > +	ret = test_crypto_adapter_conf(mode);
> >>> >> >
> >>> >> > -	return TEST_SUCCESS;
> >>> >> > +	return ret;
> >>> >> >  }
> >>> >> >
> >>> >> >
> >>> >> > --
> >>> >> > 2.28.0