From: Konstantin Ananyev <konstantin.ananyev@huawei.com>
To: Dariusz Sosnowski <dsosnowski@nvidia.com>,
Thomas Monjalon <thomas@monjalon.net>,
Ferruh Yigit <ferruh.yigit@amd.com>,
Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Cc: "dev@dpdk.org" <dev@dpdk.org>
Subject: RE: [PATCH v2 2/4] ethdev: add get restore flags driver callback
Date: Fri, 11 Oct 2024 12:54:11 +0000 [thread overview]
Message-ID: <36888a4e9c074e06a822ca84a2eda09b@huawei.com> (raw)
In-Reply-To: <20241011093351.187191-3-dsosnowski@nvidia.com>
> Before this patch, ethdev layer assumed that all drivers require that
> it has to forcefully restore:
>
> - MAC addresses
> - promiscuous mode setting
> - all multicast mode setting
>
> upon rte_eth_dev_start().
>
> This patch introduces a new callback to eth_dev_ops -
> get_restore_flags().
> Drivers implementing this callback can explicitly enable/disable
> certain parts of config restore procedure.
>
> In order to minimize the changes to all the drivers and
> preserve the current behavior, it is assumed that
> if this callback is not defined, all configuration should be restored.
>
> Signed-off-by: Dariusz Sosnowski <dsosnowski@nvidia.com>
LGTM, just few nits/suggestions, pls see below.
Series-Acked-by: Konstantin Ananyev <konstantin.ananyev@huawei.com>
> ---
> lib/ethdev/ethdev_driver.c | 11 +++++++
> lib/ethdev/ethdev_driver.h | 64 ++++++++++++++++++++++++++++++++++++++
> lib/ethdev/version.map | 1 +
> 3 files changed, 76 insertions(+)
>
> diff --git a/lib/ethdev/ethdev_driver.c b/lib/ethdev/ethdev_driver.c
> index c335a25a82..f78f9fb5c1 100644
> --- a/lib/ethdev/ethdev_driver.c
> +++ b/lib/ethdev/ethdev_driver.c
> @@ -958,3 +958,14 @@ rte_eth_switch_domain_free(uint16_t domain_id)
>
> return 0;
> }
> +
> +void
> +rte_eth_get_restore_flags(struct rte_eth_dev *dev,
> + enum rte_eth_dev_operation op,
> + uint32_t *flags)
I would go with uint64_t for flags (to avoid limitations in future).
Also, If it is void anyway, might be just return flags?
i.e:
uint64_t +rte_eth_get_restore_flags(struct rte_eth_dev *dev, enum rte_eth_dev_operation op);
Another thing - do we need to do update docs?
PMD or PG sections, release notes?
> +{
> + if (dev->dev_ops->get_restore_flags != NULL)
> + dev->dev_ops->get_restore_flags(dev, op, flags);
> + else
> + *flags = RTE_ETH_RESTORE_ALL;
> +}
> diff --git a/lib/ethdev/ethdev_driver.h b/lib/ethdev/ethdev_driver.h
> index ae00ead865..8ac5328521 100644
> --- a/lib/ethdev/ethdev_driver.h
> +++ b/lib/ethdev/ethdev_driver.h
> @@ -1235,6 +1235,47 @@ typedef int (*eth_count_aggr_ports_t)(struct rte_eth_dev *dev);
> typedef int (*eth_map_aggr_tx_affinity_t)(struct rte_eth_dev *dev, uint16_t tx_queue_id,
> uint8_t affinity);
>
> +/**
> + * @internal
> + * Defines types of operations which can be executed by the application.
> + */
> +enum rte_eth_dev_operation {
> + RTE_ETH_START,
> +};
> +
> +/**@{@name Restore flags
> + * Flags returned by get_restore_flags() callback.
> + * They indicate to ethdev layer which configuration is required to be restored.
> + */
> +/** If set, ethdev layer will forcefully restore default and any other added MAC addresses. */
> +#define RTE_ETH_RESTORE_MAC_ADDR RTE_BIT32(0)
> +/** If set, ethdev layer will forcefully restore current promiscuous mode setting. */
> +#define RTE_ETH_RESTORE_PROMISC RTE_BIT32(1)
> +/** If set, ethdev layer will forcefully restore current all multicast mode setting. */
> +#define RTE_ETH_RESTORE_ALLMULTI RTE_BIT32(2)
> +/**@}*/
> +
> +/** All configuration which can be restored by ethdev layer. */
> +#define RTE_ETH_RESTORE_ALL (RTE_ETH_RESTORE_MAC_ADDR | \
> + RTE_ETH_RESTORE_PROMISC | \
> + RTE_ETH_RESTORE_ALLMULTI)
> +
> +/**
> + * @internal
> + * Fetch from the driver what kind of configuration must be restored by ethdev layer,
> + * after certain operations are performed by the application (such as rte_eth_dev_start()).
> + *
> + * @param dev
> + * Port (ethdev) handle.
> + * @param op
> + * Type of operation executed by the application.
> + * @param flags
> + * Flags indicating what configuration must be restored by ethdev layer.
> + */
> +typedef void (*eth_get_restore_flags_t)(struct rte_eth_dev *dev,
> + enum rte_eth_dev_operation op,
> + uint32_t *flags);
> +
> /**
> * @internal A structure containing the functions exported by an Ethernet driver.
> */
> @@ -1474,6 +1515,9 @@ struct eth_dev_ops {
> eth_count_aggr_ports_t count_aggr_ports;
> /** Map a Tx queue with an aggregated port of the DPDK port */
> eth_map_aggr_tx_affinity_t map_aggr_tx_affinity;
> +
> + /** Get configuration which ethdev should restore */
> + eth_get_restore_flags_t get_restore_flags;
> };
>
> /**
> @@ -2131,6 +2175,26 @@ struct rte_eth_fdir_conf {
> struct rte_eth_fdir_flex_conf flex_conf;
> };
>
> +/**
> + * @internal
> + * Fetch from the driver what kind of configuration must be restored by ethdev layer,
> + * using get_restore_flags() callback.
> + *
> + * If callback is not defined, it is assumed that all supported configuration must be restored.
> + *
> + * @param dev
> + * Port (ethdev) handle.
> + * @param op
> + * Type of operation executed by the application.
> + * @param flags
> + * Flags indicating what configuration must be restored by ethdev layer.
> + */
> +__rte_internal
For new function we probably need 'rte_experimental'...
But from other side - it is internal.
Not sure what are rules here for such case.
> +void
> +rte_eth_get_restore_flags(struct rte_eth_dev *dev,
> + enum rte_eth_dev_operation op,
> + uint32_t *flags);
> +
> #ifdef __cplusplus
> }
> #endif
> diff --git a/lib/ethdev/version.map b/lib/ethdev/version.map
> index 1669055ca5..fa0469e602 100644
> --- a/lib/ethdev/version.map
> +++ b/lib/ethdev/version.map
> @@ -358,4 +358,5 @@ INTERNAL {
> rte_eth_switch_domain_alloc;
> rte_eth_switch_domain_free;
> rte_flow_fp_default_ops;
> + rte_eth_get_restore_flags;
> };
> --
> 2.39.5
next prev parent reply other threads:[~2024-10-11 12:54 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-10-11 9:20 [PATCH 0/4] ethdev: rework config restore Dariusz Sosnowski
2024-10-11 9:21 ` [PATCH 1/4] " Dariusz Sosnowski
2024-10-11 9:21 ` [PATCH 2/4] ethdev: add get restore flags driver callback Dariusz Sosnowski
2024-10-11 9:21 ` [PATCH 3/4] ethdev: restore config only when requested Dariusz Sosnowski
2024-10-11 9:21 ` [PATCH 4/4] net/mlx5: disable config restore Dariusz Sosnowski
2024-10-11 9:33 ` [PATCH v2 0/4] ethdev: rework " Dariusz Sosnowski
2024-10-11 9:33 ` [PATCH v2 1/4] " Dariusz Sosnowski
2024-10-11 9:33 ` [PATCH v2 2/4] ethdev: add get restore flags driver callback Dariusz Sosnowski
2024-10-11 12:54 ` Konstantin Ananyev [this message]
2024-10-11 15:30 ` Dariusz Sosnowski
2024-10-11 9:33 ` [PATCH v2 3/4] ethdev: restore config only when requested Dariusz Sosnowski
2024-10-11 9:33 ` [PATCH v2 4/4] net/mlx5: disable config restore Dariusz Sosnowski
2024-10-11 18:51 ` [PATCH v3 0/4] ethdev: rework " Dariusz Sosnowski
2024-10-11 18:51 ` [PATCH v3 1/4] " Dariusz Sosnowski
2024-10-11 18:51 ` [PATCH v3 2/4] ethdev: add get restore flags driver callback Dariusz Sosnowski
2024-10-11 18:51 ` [PATCH v3 3/4] ethdev: restore config only when requested Dariusz Sosnowski
2024-10-11 18:51 ` [PATCH v3 4/4] net/mlx5: disable config restore Dariusz Sosnowski
2024-10-11 21:25 ` [PATCH v3 0/4] ethdev: rework " Ferruh Yigit
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=36888a4e9c074e06a822ca84a2eda09b@huawei.com \
--to=konstantin.ananyev@huawei.com \
--cc=andrew.rybchenko@oktetlabs.ru \
--cc=dev@dpdk.org \
--cc=dsosnowski@nvidia.com \
--cc=ferruh.yigit@amd.com \
--cc=thomas@monjalon.net \
/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).