[dpdk-dev] [PATCH v2] mempool/dpaa2: add DPAA2 hardware offloaded mempool

[Olivier MATZ]

On Tue, 11 Apr 2017 14:50:14 +0200
Thomas Monjalon <thomas.monjalon at 6wind.com> wrote:

> 2017-04-11 09:39, Ferruh Yigit:
> > On 4/11/2017 8:50 AM, Thomas Monjalon wrote:  
> > > 2017-04-11 11:28, Hemant Agrawal:  
> > >> On 4/11/2017 1:28 AM, Olivier MATZ wrote:  
> > >>> Hemant Agrawal <hemant.agrawal at nxp.com> wrote:  
> > >>>> --- a/drivers/bus/Makefile
> > >>>> +++ b/drivers/bus/Makefile
> > >>>> @@ -33,6 +33,10 @@ include $(RTE_SDK)/mk/rte.vars.mk
> > >>>>
> > >>>>  core-libs := librte_eal librte_mbuf librte_mempool librte_ring librte_ether
> > >>>>
> > >>>> +ifeq ($(CONFIG_RTE_LIBRTE_DPAA2_MEMPOOL),y)
> > >>>> +CONFIG_RTE_LIBRTE_FSLMC_BUS = $(CONFIG_RTE_LIBRTE_DPAA2_MEMPOOL)
> > >>>> +endif
> > >>>> +
> > >>>>  DIRS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += fslmc
> > >>>>  DEPDIRS-fslmc = ${core-libs}
> > >>>>  
> > >>>
> > >>> What's the purpose of this?
> > >>> Not sure we are allowed to modify the configs in the Makefiles.  
> > >>
> > >> DPAA2_MEMPOOL will not work without the DPAA2 mempool hw instance 
> > >> detected on FSLMC_BUS.
> > >> So, it is required that if you are enabling DPAA2_MEMPOOL, FSLMC_BUS is 
> > >> to be enabled.
> > >>
> > >> Currently the config structure do not provide such dependency definitions.
> > >>
> > >> This was done based on the suggestions on the initial patches from 
> > >> Ferruh and Jerin.  
> > > 
> > > Please do not do that.
> > > We do not change the configuration in the back of the user.
> > > This kind of dependency should be managed in the configuration step
> > > which do not exist yet.
> > > 
> > > You can use $(error) to stop the compilation instead.  
> > 
> > As Hemant mentioned, this was my suggestion. There is a configuration
> > dependency here, that we don't have a way to resolve in dpdk.
> > 
> > If one of the end leaf selected, it makes sense to me to auto select
> > dependent pieces.  
> 
> A dependency must be solved at configuration time with appropriate
> user notification.
> For now, we just check them at compilation time and throw an error.

Yes, a good reason for not doing this is because the "make config"
generates a rte_config.h file. Changing a configuration option at
one place in a Makefile makes configuration inconsistent.

I don't think it's a blocker issue for the patch integration.

Regards,
Olivier