Re: [dpdk-dev] [PATCH] doc: announce malloc virt2phys symbol removal

[Thomas Monjalon <thomas@monjalon.net>]

05/08/2019 17:05, David Marchand:
> On Mon, Aug 5, 2019 at 4:39 PM Thomas Monjalon <thomas@monjalon.net> wrote:
> > 02/08/2019 15:29, David Marchand:
> > > This symbol has been deprecated for quite some time.
> > > Let's drop it in the next release.
> > > ---
> > > --- a/doc/guides/rel_notes/deprecation.rst
> > > +++ b/doc/guides/rel_notes/deprecation.rst
> > > +* eal: The ``rte_malloc_virt2phy`` function has been deprecated and replaced
> > > +  by ``rte_malloc_virt2iova`` since v17.11 and will be removed in DPDK 19.11.
> >
> > For this patch and another one about removing rte_cpu_check_supported(),
> > I have a general comment on the date of removal.
> >
> > As was stated recently in the contribution guide:
> >         http://git.dpdk.org/dpdk/commit/?id=7abe4a24cc
> >         "Deprecated APIs are removed completely just after the next LTS."
> >
> > The idea behind this policy is to avoid removals during LTS releases,
> > in order to have at least one release before X.11 LTS for end users
> > to prepare replacing the usage of the removed API.
> >
> > Does it make sense to postpone any API removal after 19.11?
> 
> Those symbols have been marked as deprecated for a long time.
> Users had to either disable Werror or they actually migrated to the new apis.
> If they chose the lazy way of not migrating to the new apis, I suspect
> they forgot about it and/or they won't look at the release notes.

Yes I agree.
That's why they can be surprised when hitting the removal.
Avoiding this removal in LTS release is one more care for
the lazy users. The question: is too much caution?

> I don't particularly have a problem with waiting for 20.02, those are
> easy to remove anyway.