[dpdk-dev] [PATCH] app/test: Remove hard coding for nb_queue_pairs in test_cryptodev

[De Lara Guarch, Pablo]

> -----Original Message-----
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> Sent: Thursday, September 29, 2016 7:25 AM
> To: Trahe, Fiona
> Cc: dev at dpdk.org; De Lara Guarch, Pablo; akhil.goyal at nxp.com
> Subject: Re: [dpdk-dev] [PATCH] app/test: Remove hard coding for
> nb_queue_pairs in test_cryptodev
> 
> 2016-09-29 14:12, Trahe, Fiona:
> > > > From: Akhil Goyal <akhil.goyal at nxp.com>
> > > >
> > > > nb_queue_pairs should not be hard coded with device specific number.
> > > > It should be retrieved from the device infos.
> > > > Also in ut_setup, ts_params->conf.nb_queue_pairs is already set in
> > > > testsuite_setup and we are not modifying it.
> > > >
> > > > Signed-off-by: Akhil Goyal <akhil.goyal at nxp.com>
> > >
> > > Acked-by: Pablo de Lara <pablo.de.lara.guarch at intel.com>
> >
> > The above code is correct, however it exposes a bug in QAT PMD unit tests.
> > And some cleanup needed for unnecessary qp setup code.
> > That cleanup then exposed a bug in aesni_mb PMD which prevents re-
> creating queue pairs of a different size.
> >
> > I have a fix and cleanup patch ready.
> > Just not sure how best to push it?
> > The original patch also needs rebasing, doesn't apply cleanly to the latest
> dpdk-next-crypto
> >
> > Pablo should I push all as a reply to the first patch - waiting first for that to
> be rebased?
> > Or
> > It would save Akhil a rebase and be simpler if I can include the original
> change in my patch and push all as a v2 superceding the original patch?  Is
> this possible?
> > Or
> > should I Nack the original patch and push all instead?
> 
> My preference goes to a v2.

Agree, send a v2, including your name and Akhil's. 

Thanks,
Pablo