Re: [dpdk-dev] [PATCH v3 1/2] mbuf: add function returning default buffer address

[Olivier Matz <olivier.matz@6wind.com>]

Hi,

On Fri, Jan 11, 2019 at 01:48:12PM +0100, David Marchand wrote:
> On Fri, Jan 11, 2019 at 12:57 PM Bruce Richardson <
> bruce.richardson@intel.com> wrote:
> 
> > On Fri, Jan 11, 2019 at 02:17:04PM +0300, Andrew Rybchenko wrote:
> > > Olivier, David,
> > >
> > > could you take a look at naming suggested below and share your thoughts.
> > > My fear is that rte_mbuf_buf_addr() is too generic and true for direct
> > mbuf
> > > only. That's why I'd like to highlight it in the function name.
> > >
> >
> > I would tend to agree with that concern.
> >
> 
> I understand your concern as well.
> 
> The only usecase we have so far is for drivers on the rx side, so
> implicitely direct mbufs.
> But from a api user pov, explicit is always better.
> 
> I will let Olivier have the last word :-)

Thanks Andrew for pointing this out.

However I agree with Yongseok: we already have many functions that
applies to direct mbufs that don't have "direct" in their names.

In my opinion, rte_mbuf_buf_addr() is a good name, but I think the
doxygen comment could be improved a bit to state that it returns the
pointer to the embedded data. I also think that a small comment
explaining why the mp arg is required would be helpful.

Thanks,
Olivier