[dpdk-dev] [PATCH v8 2/2] app/test_pmd: add tests for new API's

Iremonger, Bernard bernard.iremonger at intel.com
Wed Oct 12 17:48:11 CEST 2016


Hi Thomas,

<snip>

> Subject: Re: [dpdk-dev] [PATCH v8 2/2] app/test_pmd: add tests for new
> API's
> 
> 2016-10-12 15:27, Iremonger, Bernard:
> > Hi Thomas,
> >
> > <snip>
> >
> > > Subject: Re: [dpdk-dev] [PATCH v8 2/2] app/test_pmd: add tests for
> > > new API's
> > >
> > > 2016-10-12 16:03, Bernard Iremonger:
> > > > --- a/app/test-pmd/Makefile
> > > > +++ b/app/test-pmd/Makefile
> > > > @@ -58,6 +58,17 @@ SRCS-y += csumonly.c  SRCS-y += icmpecho.c
> > > >  SRCS-$(CONFIG_RTE_LIBRTE_IEEE1588) += ieee1588fwd.c
> > > >
> > > > +ifeq ($(CONFIG_RTE_BUILD_SHARED_LIB),n)
> > > > +LDLIBS-$(CONFIG_RTE_LIBRTE_IXGBE_PMD) += -lrte_pmd_ixgbe
> endif
> > > > +
> > > > +ifeq ($(CONFIG_RTE_BUILD_SHARED_LIB),y)
> > > > +ifeq ($(CONFIG_RTE_LIBRTE_IXGBE_PMD),y)
> > > > +LDLIBS += -lrte_pmd_ixgbe
> > > > +endif
> > > > +endif
> > >
> > > Sorry if I miss something, but I thought it was enough to do:
> > > LDLIBS-$(CONFIG_RTE_LIBRTE_IXGBE_PMD) += -lrte_pmd_ixgbe
> >
> > No unfortunately not, the above line does not work for all scenarios .
> >
> > There are 4 scenarios as follows:
> >
> > CONFIG_RTE_LIBRTE_IXGBE_PMD=y  with
> CONFIG_RTE_BUILD_SHARED_LIB=n
> >
> > CONFIG_RTE_LIBRTE_IXGBE_PMD=y  with
> CONFIG_RTE_BUILD_SHARED_LIB=y
> >
> > CONFIG_RTE_LIBRTE_IXGBE_PMD=n  with
> CONFIG_RTE_BUILD_SHARED_LIB=y
> >
> > CONFIG_RTE_LIBRTE_IXGBE_PMD=n  with
> CONFIG_RTE_BUILD_SHARED_LIB=n
> >
> > I have been doing quite a bit of building today to get it to work in all 4
> scenarios.
> 
> I have a doubt about the tests because LDLIBS-y does not exist.
> There is _LDLIBS-y and LDLIBS.
> But in the static case, you wrote:
> 	LDLIBS-$(CONFIG_RTE_LIBRTE_IXGBE_PMD) += -lrte_pmd_ixgbe
> endif Please could you check?

LDLIBS-y exists in some of the scenarios but not all.
Do you want me to check the four scenarios  with just the line below.

LDLIBS-$(CONFIG_RTE_LIBRTE_IXGBE_PMD) += -lrte_pmd_ixgbe

Regards,

Bernard.



More information about the dev mailing list