Re: [dpdk-dev] [PATCH 0/2] add MRVL MVPP2 PMD to meson

[Bruce Richardson <bruce.richardson@intel.com>]

On Thu, Apr 19, 2018 at 10:55:33AM +0200, Tomasz Duszynski wrote:
> On Wed, Apr 18, 2018 at 04:02:30PM +0100, Bruce Richardson wrote:
> > On Fri, Apr 13, 2018 at 05:12:19PM +0100, Bruce Richardson wrote:
> > > On Wed, Apr 11, 2018 at 01:45:05PM +0200, Tomasz Duszynski wrote:
> > > > This patchseries adds MRVL MVPP2 PMD to meson build system.
> > > >
> > > > Tomasz Duszynski (2):
> > > >   net/mvpp2: rename the version file to standard
> > > >   net/mvpp2: add meson build file
> > > >
> > >
> > > The patches look ok to me as far as the meson code is concerned, but I have
> > > no way to test compilation etc. It doesn't cause issues with other x86 or
> > > arm builds though, so:
> > >
> > > Series Acked-by: Bruce Richardson <bruce.richardson@intel.com>
> >
> > +Pablo, who is looking at the crypto driver which is similar.
> >
> > I've just realised at this stage - while looking at something similar with
> > the turbo_sw baseband driver - that the use of environmental variables is
> > probably going to cause us problems down the line here. In the case of
> > cross-compilation, the meson build is going to pull the environment
> > variable of the host, and use that, even in cases where there is no
> > cross-compile library available.
> >
> > I think that for cases like this, using a build option is a better
> > solution. It explicitly can be set for each independent build, avoiding the
> > cross-build issues I refer to, and also prevents us having issues with
> > changing the path in the environment and meson not recognising the change
> > (environment variables are not tracked for reconfigure, unlike options).
> >
> > So, would you be ok with changing this to take the MUSDK path from a meson
> > option rather than the environment?
> >
> > /Bruce
> 
> I am okay with the proposed change. As for crypto patch I'll prepare v2.
> 
> --
> - Tomasz Duszyński

Thanks.