[dpdk-dev] [v1] ethdev: support Tx queue used count

[Jerin Jacob]

On Thu, Jan 18, 2024 at 3:47 PM Konstantin Ananyev
<konstantin.ananyev at huawei.com> wrote:
>
>
> Hi Jerin,

Hi Konstantin,


>
> > > > Introduce a new API to retrieve the number of used descriptors
> > > > in a Tx queue. Applications can leverage this API in the fast path to
> > > > inspect the Tx queue occupancy and take appropriate actions based on the
> > > > available free descriptors.
> > > >
> > > > A notable use case could be implementing Random Early Discard (RED)
> > > > in software based on Tx queue occupancy.
> > > >
> > > > Signed-off-by: Jerin Jacob <jerinj at marvell.com>
> >
> > > > @@ -6803,6 +6803,80 @@ rte_eth_recycle_mbufs(uint16_t rx_port_id, uint16_t rx_queue_id,
> > > >  __rte_experimental
> > > >  int rte_eth_buffer_split_get_supported_hdr_ptypes(uint16_t port_id, uint32_t *ptypes, int num);
> > > >
> > > > +/**
> > > > + * @warning
> > > > + * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
> > > > + *
> > > > + * Get the number of used descriptors of a Tx queue
> > > > + *
> > > > + * This function retrieves the number of used descriptors of a transmit queue.
> > > > + * Applications can use this API in the fast path to inspect Tx queue occupancy and take
> > > > + * appropriate actions based on the available free descriptors.
> > > > + * An example action could be implementing the Random Early Discard (RED).
> > >
> > > Sorry, I probably misunderstood your previous mails, but wouldn't it be more convenient
> > > for user to have rte_eth_tx_queue_free_count(...) as fast-op, and
> > > have rte_eth_tx_queue_count(...) {  queue_txd_num - rte_eth_tx_queue_free_count(...);}
> > > as a slow-path function in rte_ethdev.c?
> >
> > The general feedback is to align with the Rx queue API, specifically
> > rte_eth_rx_queue_count,
> > and it's noted that there is no equivalent rte_eth_rx_queue_free_count.
> >
> > Given that the free count can be obtained by subtracting the used
> > count from queue_txd_num,
> > it is considered that either approach is acceptable.
> >
> > The application configures queue_txd_num with tx_queue_setup(), and
> > the application can store that value in its structure.
> > This would enable fast-path usage for both base cases (whether the
> > application needs information about free or used descriptors)
> > with just one API(rte_eth_tx_queue_count())
>
> Right now I don't use these functions, but if I think what most people are interested in:
> - how many packets you can receive immediately (rx_queue_count)
> - how many packets you can transmit immediately (tx_queue_free_count)

Yes. That's why initially I kept the free version.

It seems like other prominent use cases for _used_ version is for QoS
to enable something  like
in positive logic,

if < 98% % used then action Tail drop
else if  < 60% used then action RED

> Sure, I understand that user can store txd_num  somewhere and then do subtraction himself.
> Though it means more effort for the user, and the only reason for that, as I can see,
> is to have RX and TX function naming symmetric.
> Which seems much less improtant to me comparing to user convenience.
> Anyway, as I stated above, I don't use these functions right now,
> so if the majority of users are happy with current approach, I would not insist :)

No strong opinion for me either. Looks like majority users like used version.
Let's go with. I am open for changing to free version, if the majority
of users like that.

> Konstantin
>