Re: [dpdk-dev] [PATCH v4 1/4] ethdev: add tm support for shaper config in pkt mode

[Thomas Monjalon <thomas@monjalon.net>]

27/04/2020 18:28, Dumitrescu, Cristian:
> From: Yigit, Ferruh <ferruh.yigit@intel.com>
> > On 4/27/2020 10:19 AM, Dumitrescu, Cristian wrote:
> > > From: Yigit, Ferruh <ferruh.yigit@intel.com>
> > >> Hi Nithin,
> > >>
> > >> It looks like patch is causing ABI break, I am getting following warning [1],
> > >> can you please check?
> > >>
> > >> [1]
> > >> https://pastebin.com/XYNFg14u
> > >
> > >
> > > Hi Ferruh,
> > >
> > > The RTE_TM API is marked as experimental, but it looks that this was not
> > correctly marked when __rte_experimental ABI checker was introduced.
> > >
> > > It is marked as experimental at the top of the rte_tm.h, similarly to other
> > APIs introduced around same time, but it was not correctly picked up by the
> > ABI check procedure when later introduced, so __rte_experimental was not
> > added to every function.
> > >
> > 
> > :(
> > 
> > Is it time to mature them?
> > 
> > As you said they are not marked as experimental both in header file
> > (function
> > declarations) and .map file.
> > 
> > The problem is, they are not marked as experimental in DPDK_20.0 ABI
> > (v19.11),
> > so marking them as experimental now will break the ABI. Not sure what to
> > do,
> > cc'ed a few ABI related names for comment.
> > 
> > For me, we need to proceed as the experimental tag removed and APIs
> > become
> > mature starting from v19.11, since this is what happened in practice, and
> > remove
> > a few existing being experimental references in the doxygen comments.
> > 
> > Ray, Neil, David, Luca, Kevin, what do you think?
> 
> Hi Ferruh,
> 
> IMO your proposed approach is fixing the wrong problem and is
> probably not the right way of doing things.
> 
> This API is correctly marked as experimental in the header
> file rte_tm.h and in the MAINTAINERS file,

in rte_tm.h:
	* @warning
	* @b EXPERIMENTAL: this API may change without prior notice

in MAINTAINERS:
	Traffic Management API - EXPERIMENTAL
	M: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
	T: git://dpdk.org/next/dpdk-next-qos
	F: lib/librte_ethdev/rte_tm*


> therefore it should remain experimental,

in rte_ethdev_version.map:
	before 19.11: DPDK_17.08 {
	since 19.11: DPDK_20.0 {

When adding rte_tm in 17.08:
http://git.dpdk.org/dpdk/diff/lib/librte_ether/rte_ether_version.map?id=5d109deffa

In 17.08, early July 2017, the first EXPERIMENTAL section was declared in EAL:
http://git.dpdk.org/dpdk/diff/lib/librte_eal/linuxapp/eal/rte_eal_version.map?id=a3ee360f444

When adding rte_mtr in 17.11, all functions are made experimental:
http://git.dpdk.org/dpdk/diff/lib/librte_ether/rte_ethdev_version.map?id=6613ffe1

In 18.02, the tag __rte_experimental was introduced:
	http://git.dpdk.org/dpdk/commit/?id=7d540a3e735
and functions are marked (including rte_mtr but not rte_tm):
	http://git.dpdk.org/dpdk/commit/?id=77b7b81e32e


> as more changes are expected from the people like Nithin and others
> currently upstreaming drivers for it.

They are doing changes in the API introduced 3 years ago.


> For some reason, the __rte_experimental tags were not added to
> this file when the ABI checker was introduced,
> and this is the real problem that should be fixed.

The mistake was done 2 years ago.
As maintainer of rte_tm code, you are expected to notice the issue.
As maintainer of rte_mtr code, you were expected to review the change
on the mtr functions being marked as experimental.
I am sorry that it took 2 years to discover the gap.

We can fix the ABI in the ABI-breakage window: in 20.11.