BR
Rongwei
> -----Original Message-----
> From: Stephen Hemminger <step...@networkplumber.org>
> Sent: Thursday, December 1, 2022 23:10
> To: Rongwei Liu <rongw...@nvidia.com>
> Cc: Matan Azrad <ma...@nvidia.com>; Slava Ovsiienko
> <viachesl...@nvidia.com>; Ori Kam <or...@nvidia.com>; NBU-Contact-
> Thomas Monjalon (EXTERNAL) <tho...@monjalon.net>; Ferruh Yigit
> <ferruh.yi...@amd.com>; Andrew Rybchenko
> <andrew.rybche...@oktetlabs.ru>; dev@dpdk.org; Raslan Darawsheh
> <rasl...@nvidia.com>
> Subject: Re: [RFC 2/2] ethdev: add API to set process to primary or secondary
>
> External email: Use caution opening links or attachments
>
>
> On Thu, 1 Dec 2022 10:20:05 +0200
> Rongwei Liu <rongw...@nvidia.com> wrote:
>
> > +/**
> > + * @warning
> > + * @b EXPERIMENTAL: this API may change without prior notice
> > + *
> > + * Set the rte_eth process to the primary or secondary role which
> > +affects
> > + * the flow rules offloading. It doesn't allow multiple processes to
> > +be the
> > + * same role unless no offload rules are set.
> > + * The primary process's flow rules are effective immediately while
> > +the secondary
> > + * process's rules will be queued in hardware until it becomes
> > +primary or no
> > + * primary process is alive.
> > + * The primary application will always receive traffic while the
> > +secondary
> > + * application will receive traffic when no matching rules are
> > +present from
> > + * the primary application.
> > + *
> > + * The application is primary by default if this API is not called.
> > + *
> > + * When a process transforms from a secondary to a primary role, all
> > +preceding
> > + * flow rules which are queued by hardware will be effective immediately.
> > + * Before role transition, all the rules set by the primary process
> > +should be
> > + * flushed first.
> > + *
> > + * When role flag "RTE_ETH_PROCESS_NIC_DUP_WITH_SECONDARY" is set,
> > +NIC domain
> > + * flow rules are effective immediately even if a process is secondary.
> > + *
> > + * @param active
> > + * Process primary (role) or not (secondary).
> > + * @param flag
> > + * The role flag.
> > + * @return
> > + * - (>=0) Number of rte devices which have been switched successfully.
> > + * - (-EINVAL) if bad parameter.
> > + */
> > +__rte_experimental
> > +int rte_eth_process_set_primary(bool primary, uint32_t flag);
>
> The state of the devices and the system is really unstable if this fails.
> There is
> no rollback here.
>
Assume application is calling rte_eth_process_set_primary(false);
Once failed, call all preceding successful ports as
rte_eth_process_set_primary(true);
What do you think?
> I think this should have a PMD capability flag so that application can check
> that device supports doing this. And it would have to be opt-in so that
> existing
> devices would always fail.
If device doesn't support it, it can set the ethdev callback to NULL or return
failure for all devices.
Then the devices' state will be consistent.
