[3/3] app/test: add more cryptodev tests to meson suite

  The meson driver-tests suite did not include some ipsec_mb SW Crypto
PMD tests, and QAT tests. These are now added to avoid them being missed
if the user runs tests only using the meson suite infrastructure.

Signed-off-by: Ciara Power <ciara.power@intel.com>
---
 app/test/meson.build | 5 +++++
 1 file changed, 5 insertions(+)
  

> Subject: [EXT] [PATCH 3/3] app/test: add more cryptodev tests to meson suite
> 
> The meson driver-tests suite did not include some ipsec_mb SW Crypto
> PMD tests, and QAT tests. These are now added to avoid them being missed
> if the user runs tests only using the meson suite infrastructure.
> 
> Signed-off-by: Ciara Power <ciara.power@intel.com>
> ---
This patch is ok for now, but it is not scalable.
We should look for a way to consolidate all cryptodev autotests under one name.
  

Hi Akhil,

> -----Original Message-----
> From: Akhil Goyal <gakhil@marvell.com>
> Sent: Tuesday 7 March 2023 17:23
> To: Power, Ciara <ciara.power@intel.com>; dev@dpdk.org
> Cc: Ji, Kai <kai.ji@intel.com>
> Subject: RE: [EXT] [PATCH 3/3] app/test: add more cryptodev tests to meson
> suite
> 
> > Subject: [EXT] [PATCH 3/3] app/test: add more cryptodev tests to meson
> > suite
> >
> > The meson driver-tests suite did not include some ipsec_mb SW Crypto
> > PMD tests, and QAT tests. These are now added to avoid them being
> > missed if the user runs tests only using the meson suite infrastructure.
> >
> > Signed-off-by: Ciara Power <ciara.power@intel.com>
> > ---
> This patch is ok for now, but it is not scalable.
> We should look for a way to consolidate all cryptodev autotests under one
> name.

I actually had started out to add all the missing cryptodev tests (bcmfs, caam_jr, ccp, nitrox etc.),
But the docs mention it's up to the maintainer/developer to decide whether the tests should be included in the meson test suite or not.
So, because I was unaware of the reasoning behind leaving them out, I just stuck to adding QAT + ipsec_mb ones in.

I guess, if we have them all under one name, it would remove that level of granularity for which tests are added.
Not sure if that is needed or not?

Thanks,
Ciara
  

Hi Ciara,
> Hi Akhil,
> 
> > -----Original Message-----
> > From: Akhil Goyal <gakhil@marvell.com>
> > Sent: Tuesday 7 March 2023 17:23
> > To: Power, Ciara <ciara.power@intel.com>; dev@dpdk.org
> > Cc: Ji, Kai <kai.ji@intel.com>
> > Subject: RE: [EXT] [PATCH 3/3] app/test: add more cryptodev tests to meson
> > suite
> >
> > > Subject: [EXT] [PATCH 3/3] app/test: add more cryptodev tests to meson
> > > suite
> > >
> > > The meson driver-tests suite did not include some ipsec_mb SW Crypto
> > > PMD tests, and QAT tests. These are now added to avoid them being
> > > missed if the user runs tests only using the meson suite infrastructure.
> > >
> > > Signed-off-by: Ciara Power <ciara.power@intel.com>
> > > ---
> > This patch is ok for now, but it is not scalable.
> > We should look for a way to consolidate all cryptodev autotests under one
> > name.
> 
> I actually had started out to add all the missing cryptodev tests (bcmfs, caam_jr,
> ccp, nitrox etc.),
> But the docs mention it's up to the maintainer/developer to decide whether the
> tests should be included in the meson test suite or not.
> So, because I was unaware of the reasoning behind leaving them out, I just stuck
> to adding QAT + ipsec_mb ones in.
> 
> I guess, if we have them all under one name, it would remove that level of
> granularity for which tests are added.
> Not sure if that is needed or not?
> 
From the past some time, we have added a lot of capability checks in the test cases
So that the same case may be run on all the PMDs.

I believe it should be doable at some point going forward.
This will help in reducing the unnecessary bloating of the code.
We should put in some effort to make it similar to other device test cases.
PMD specific suites are only in case of cryptodev. This should be fixed.
Moreover, it would be beneficial for all the PMDs to be tested at similar levels
Of granularity.

Regards,
Akhil