DPDK patches and discussions
 help / color / mirror / Atom feed
From: Wan Junjie <wanjunjie@bytedance.com>
To: "Min Hu (Connor)" <humin29@huawei.com>
Cc: dev@dpdk.org, Chas Williams <chas3@att.com>
Subject: Re: [External] Re: [PATCH 2/2] net/bonding: fix slaves initilizing on mtu setting
Date: Tue, 15 Feb 2022 18:32:44 +0800	[thread overview]
Message-ID: <CALYSL_De_fVU+jHSN9Mm8RVq=VGG40xe2TC-s_X0Ld5b8qJQqw@mail.gmail.com> (raw)
In-Reply-To: <bf2d6fa1-013d-b04d-6c8b-aa0b43d07b0c@huawei.com>

Hi Min Hu,

On Tue, Feb 15, 2022 at 5:26 PM Min Hu (Connor) <humin29@huawei.com> wrote:
>
> Hi, wan,
> two comments:
> 1. wrong spelling in headline

Will fix it in v2.

> 2. comment info should be detailed, such as,
> show the testpmd commands about how to operate
> in this two situations.
>
Testpmd did not show the issue here.
to reproduce, we need call the api by the sequence:
rte_eth_dev_configure
rte_eth_dev_set_mtu
queue setup and start ...

for dpdk application, pass the device
--vdev="net_bonding0,mode=2,slave=0000:af:00.0"

Then application will complain for the slave device, that
"Port 0 must be configured before MTU set"

Will add it to commit msg in v2


Regards,
Wan Junjie

> Thanks.
>
> 在 2022/2/14 16:13, Wan Junjie 写道:
> > When initializing a bonding device with all slaves in vdev string,
> > which means bonding device initilizes slaves, instead of initializing
> > slaves before creating the bonding device, it will fail to call
> > set_mtu api for the bonding device.
> >
> > Fixes:b26bee1 ("ethdev: forbid MTU set before device configure")
> >
> > Signed-off-by: Wan Junjie <wanjunjie@bytedance.com>
> > ---
> >   app/test/test_link_bonding.c           |  4 +++
> >   app/test/test_link_bonding_rssconf.c   |  4 +++
> >   drivers/net/bonding/eth_bond_private.h |  4 +++
> >   drivers/net/bonding/rte_eth_bond_api.c |  6 ++++
> >   drivers/net/bonding/rte_eth_bond_pmd.c | 43 ++++++++++++++++++++++----
> >   5 files changed, 55 insertions(+), 6 deletions(-)
> >
> > diff --git a/app/test/test_link_bonding.c b/app/test/test_link_bonding.c
> > index dc6fc46b9c..12c50ef393 100644
> > --- a/app/test/test_link_bonding.c
> > +++ b/app/test/test_link_bonding.c
> > @@ -181,6 +181,10 @@ configure_ethdev(uint16_t port_id, uint8_t start, uint8_t en_isr)
> >                       test_params->nb_tx_q, &default_pmd_conf),
> >                       "rte_eth_dev_configure for port %d failed", port_id);
> >
> > +     int ret = rte_eth_dev_set_mtu(port_id, 1550);
> > +     RTE_TEST_ASSERT(ret == 0 || ret == -ENOTSUP,
> > +                     "rte_eth_dev_set_mtu for port %d failed", port_id);
> > +
> >       for (q_id = 0; q_id < test_params->nb_rx_q; q_id++)
> >               TEST_ASSERT_SUCCESS(rte_eth_rx_queue_setup(port_id, q_id, RX_RING_SIZE,
> >                               rte_eth_dev_socket_id(port_id), &rx_conf_default,
> > diff --git a/app/test/test_link_bonding_rssconf.c b/app/test/test_link_bonding_rssconf.c
> > index f9eae93973..7228965ced 100644
> > --- a/app/test/test_link_bonding_rssconf.c
> > +++ b/app/test/test_link_bonding_rssconf.c
> > @@ -128,6 +128,10 @@ configure_ethdev(uint16_t port_id, struct rte_eth_conf *eth_conf,
> >                       RXTX_QUEUE_COUNT, eth_conf) == 0, "Failed to configure device %u",
> >                       port_id);
> >
> > +     int ret = rte_eth_dev_set_mtu(port_id, 1550);
> > +     RTE_TEST_ASSERT(ret == 0 || ret == -ENOTSUP,
> > +                     "rte_eth_dev_set_mtu for port %d failed", port_id);
> > +
> >       for (rxq = 0; rxq < RXTX_QUEUE_COUNT; rxq++) {
> >               TEST_ASSERT(rte_eth_rx_queue_setup(port_id, rxq, RXTX_RING_SIZE,
> >                               rte_eth_dev_socket_id(port_id), NULL,
> > diff --git a/drivers/net/bonding/eth_bond_private.h b/drivers/net/bonding/eth_bond_private.h
> > index 156335c425..8222e3cd38 100644
> > --- a/drivers/net/bonding/eth_bond_private.h
> > +++ b/drivers/net/bonding/eth_bond_private.h
> > @@ -246,6 +246,10 @@ int
> >   slave_configure(struct rte_eth_dev *bonded_eth_dev,
> >               struct rte_eth_dev *slave_eth_dev);
> >
> > +int
> > +slave_start(struct rte_eth_dev *bonded_eth_dev,
> > +             struct rte_eth_dev *slave_eth_dev);
> > +
> >   void
> >   slave_remove(struct bond_dev_private *internals,
> >               struct rte_eth_dev *slave_eth_dev);
> > diff --git a/drivers/net/bonding/rte_eth_bond_api.c b/drivers/net/bonding/rte_eth_bond_api.c
> > index b78867b125..4ac191c468 100644
> > --- a/drivers/net/bonding/rte_eth_bond_api.c
> > +++ b/drivers/net/bonding/rte_eth_bond_api.c
> > @@ -566,6 +566,12 @@ __eth_bond_slave_add_lock_free(uint16_t bonded_port_id, uint16_t slave_port_id)
> >                                       slave_port_id);
> >                       return -1;
> >               }
> > +             if (slave_start(bonded_eth_dev, slave_eth_dev) != 0) {
> > +                     internals->slave_count--;
> > +                     RTE_BOND_LOG(ERR, "rte_bond_slaves_start: port=%d",
> > +                                     slave_port_id);
> > +                     return -1;
> > +             }
> >       }
> >
> >       /* Update all slave devices MACs */
> > diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c
> > index aa6519f83c..c31b723071 100644
> > --- a/drivers/net/bonding/rte_eth_bond_pmd.c
> > +++ b/drivers/net/bonding/rte_eth_bond_pmd.c
> > @@ -1678,14 +1678,10 @@ int
> >   slave_configure(struct rte_eth_dev *bonded_eth_dev,
> >               struct rte_eth_dev *slave_eth_dev)
> >   {
> > -     struct bond_rx_queue *bd_rx_q;
> > -     struct bond_tx_queue *bd_tx_q;
> >       uint16_t nb_rx_queues;
> >       uint16_t nb_tx_queues;
> >
> >       int errval;
> > -     uint16_t q_id;
> > -     struct rte_flow_error flow_error;
> >
> >       struct bond_dev_private *internals = bonded_eth_dev->data->dev_private;
> >
> > @@ -1758,6 +1754,19 @@ slave_configure(struct rte_eth_dev *bonded_eth_dev,
> >                               slave_eth_dev->data->port_id, errval);
> >               return errval;
> >       }
> > +     return 0;
> > +}
> > +
> > +int
> > +slave_start(struct rte_eth_dev *bonded_eth_dev,
> > +             struct rte_eth_dev *slave_eth_dev)
> > +{
> > +     int errval = 0;
> > +     struct bond_rx_queue *bd_rx_q;
> > +     struct bond_tx_queue *bd_tx_q;
> > +     uint16_t q_id;
> > +     struct rte_flow_error flow_error;
> > +     struct bond_dev_private *internals = bonded_eth_dev->data->dev_private;
> >
> >       /* Setup Rx Queues */
> >       for (q_id = 0; q_id < bonded_eth_dev->data->nb_rx_queues; q_id++) {
> > @@ -1806,10 +1815,13 @@ slave_configure(struct rte_eth_dev *bonded_eth_dev,
> >                       return errval;
> >               }
> >
> > -             if (internals->mode4.dedicated_queues.flow[slave_eth_dev->data->port_id] != NULL)
> > -                     rte_flow_destroy(slave_eth_dev->data->port_id,
> > +             if (internals->mode4.dedicated_queues.flow[slave_eth_dev->data->port_id] != NULL) {
> > +                     errval = rte_flow_destroy(slave_eth_dev->data->port_id,
> >                                       internals->mode4.dedicated_queues.flow[slave_eth_dev->data->port_id],
> >                                       &flow_error);
> > +                     RTE_BOND_LOG(ERR, "bond_ethdev_8023ad_flow_destroy: port=%d, err (%d)",
> > +                             slave_eth_dev->data->port_id, errval);
> > +             }
> >
> >               errval = bond_ethdev_8023ad_flow_set(bonded_eth_dev,
> >                               slave_eth_dev->data->port_id);
> > @@ -2001,6 +2013,13 @@ bond_ethdev_start(struct rte_eth_dev *eth_dev)
> >                               internals->slaves[i].port_id);
> >                       goto out_err;
> >               }
> > +             if (slave_start(eth_dev, slave_ethdev) != 0) {
> > +                     RTE_BOND_LOG(ERR,
> > +                             "bonded port (%d) failed to start slave device (%d)",
> > +                             eth_dev->data->port_id,
> > +                             internals->slaves[i].port_id);
> > +                     goto out_err;
> > +             }
> >               /* We will need to poll for link status if any slave doesn't
> >                * support interrupts
> >                */
> > @@ -3849,6 +3868,18 @@ bond_ethdev_configure(struct rte_eth_dev *dev)
> >               return -1;
> >       }
> >
> > +     /* configure slaves so we can pass mtu setting */
> > +     for (i = 0; i < internals->slave_count; i++) {
> > +             struct rte_eth_dev *slave_ethdev =
> > +                             &(rte_eth_devices[internals->slaves[i].port_id]);
> > +             if (slave_configure(dev, slave_ethdev) != 0) {
> > +                     RTE_BOND_LOG(ERR,
> > +                             "bonded port (%d) failed to configure slave device (%d)",
> > +                             dev->data->port_id,
> > +                             internals->slaves[i].port_id);
> > +                     return -1;
> > +             }
> > +     }
> >       return 0;
> >   }
> >
> >

  reply	other threads:[~2022-02-15 10:32 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-02-14  8:13 [PATCH 1/2] net/bonding: add error hint for invald args Wan Junjie
2022-02-14  8:13 ` [PATCH 2/2] net/bonding: fix slaves initilizing on mtu setting Wan Junjie
2022-02-15  9:26   ` Min Hu (Connor)
2022-02-15 10:32     ` Wan Junjie [this message]
2022-02-15 10:59   ` [PATCH v2 2/2] net/bonding: fix slaves initializing " Wan Junjie
2022-02-16 13:22     ` Ferruh Yigit
2022-02-17  8:55       ` Min Hu (Connor)
2022-02-15  9:12 ` [PATCH 1/2] net/bonding: add error hint for invald args Min Hu (Connor)
2022-02-15  9:50   ` [External] " Wan Junjie
2022-02-15 10:56 ` [PATCH v2 1/2] net/bonding: add error hint for invalid args Wan Junjie
2022-02-15 11:56   ` Ferruh Yigit
2022-02-17 11:57     ` Ferruh Yigit
2022-02-16  0:29   ` Min Hu (Connor)

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CALYSL_De_fVU+jHSN9Mm8RVq=VGG40xe2TC-s_X0Ld5b8qJQqw@mail.gmail.com' \
    --to=wanjunjie@bytedance.com \
    --cc=chas3@att.com \
    --cc=dev@dpdk.org \
    --cc=humin29@huawei.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).