On Wed, Apr 04, 2018 at 11:14:35AM +0100, Ferruh Yigit wrote:
> On 4/4/2018 3:25 AM, Lu, Wenzhuo wrote:
> > Hi Shweta,
> >
> >
> >> -----Original Message-----
> >> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Shweta Choudaha
> >> Sent: Friday, February 23, 2018 7:59 PM
> >> To: dev@dpdk.org
> >> Cc: shweta.choud...@att.com
> >> Subject: [dpdk-dev] [PATCH 1/1] net/ixgbe: Add API to update SBP bit
> >>
> >> From: Shweta Choudaha <shweta.choud...@att.com>
> >>
> >> Add ixgbe API to enable SBP bit in FCTRL register to allow receiving
> >> packets
> >> that may otherwise be considered length errors by ixgbe and dropped
> >>
> >> Signed-off-by: Shweta Choudaha <shweta.choud...@att.com>
> >> Reviewed-by: Chas Williams <ch...@att.com>
> >> Reviewed-by: Luca Boccassi <bl...@debian.org>
> >> ---
> >
> >
> >> --- a/drivers/net/ixgbe/rte_pmd_ixgbe_version.map
> >> +++ b/drivers/net/ixgbe/rte_pmd_ixgbe_version.map
> >> @@ -52,3 +52,9 @@ DPDK_17.08 {
> >> rte_pmd_ixgbe_bypass_wd_timeout_show;
> >> rte_pmd_ixgbe_bypass_wd_timeout_store;
> >> } DPDK_17.05;
> >> +
> >> +EXPERIMENTAL {
> >> + global:
> >> +
> >> + rte_pmd_ixgbe_upd_fctrl_sbp;
> >> +} DPDK_17.08;
> > The patch is fine except we should use 18.05 here.
>
> We have a rule to have an API as experimental for one release.
>
> Not sure about this rule covering the PMD specific APIs but by default it is,
> we
> can discuss to have an exception for PMD APIs or not.
>
> After one release developer should send patch to remove experimental from API
>
+1 for following the process.
I don't see why pmd-specific APIs should be a general exception. Any
exceptions should be on an individual case-by-case basis IMHO
