From: Gregory Etelson <getelson@nvidia.com>
To: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>,
"dev@dpdk.org" <dev@dpdk.org>
Cc: Matan Azrad <matan@nvidia.com>,
Raslan Darawsheh <rasland@nvidia.com>, Ori Kam <orika@nvidia.com>,
"NBU-Contact-Thomas Monjalon (EXTERNAL)" <thomas@monjalon.net>,
Ferruh Yigit <ferruh.yigit@amd.com>
Subject: RE: [PATCH v8 1/2] ethdev: add query_update sync and async function calls
Date: Thu, 2 Feb 2023 10:24:22 +0000 [thread overview]
Message-ID: <IA1PR12MB6332C9F3CB7546DC41EC6197A5D69@IA1PR12MB6332.namprd12.prod.outlook.com> (raw)
In-Reply-To: <0a24a07c-6d36-2f5e-5f24-1501bcc41c1e@oktetlabs.ru>
Hello Andrew,
Please check out my comments below.
Regards,
Gregory
>
>
> query_update is bad for summary, consider something like
> ethdev: add query and update sync and async API
>
Will fix in v9.
> On 2/1/23 18:16, Gregory Etelson wrote:
> > Current API allows either query or update indirect flow action.
> > If indirect action must be conditionally updated according to it's
> > present state application must first issue action query then
> > analyze returned data and if needed issue update request.
> > When the update will be processed, action state can change and
> > the update can invalidate the action.
> >
> > Add `rte_flow_action_handle_query_update` function call,
> > and it's async version `rte_flow_async_action_handle_query_update`
> > to atomically query and update flow action.
> >
> > Application can control query and update order, if that is supported
> > by port hardware, by setting `qu_mode` parameter to
> > RTE_FLOW_QU_QUERY_FIRST or RTE_FLOW_QU_UPDATE_FIRST.
> >
> > Signed-off-by: Gregory Etelson <getelson@nvidia.com>
>
> [snip]
>
> > diff --git a/lib/ethdev/rte_flow.c b/lib/ethdev/rte_flow.c
> > index 7d0c24366c..ece40e7877 100644
> > --- a/lib/ethdev/rte_flow.c
> > +++ b/lib/ethdev/rte_flow.c
> > @@ -1883,3 +1883,48 @@ rte_flow_async_action_handle_query(uint16_t
> port_id,
> > action_handle, data, user_data, error);
> > return flow_err(port_id, ret, error);
> > }
> > +
> > +int
> > +rte_flow_action_handle_query_update(uint16_t port_id,
> > + struct rte_flow_action_handle *handle,
> > + const void *update, void *query,
> > + enum rte_flow_query_update_mode mode,
> > + struct rte_flow_error *error)
> > +{
> > + int ret;
> > + struct rte_eth_dev *dev;
> > + const struct rte_flow_ops *ops;
> > +
> > + RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
> > + dev = &rte_eth_devices[port_id];
>
> (-EINVAL) if *handle* invalid or both *query* and *update* are NULL.
>
> must be handled here. It is generic checks and we should do it
> in a generic place to avoid duplication in PMD callbacks.
>
Will fix in v9.
> > + ops = rte_flow_ops_get(port_id, error);
> > + if (!ops || !ops->action_handle_query_update)
> > + return -ENOTSUP;
>
> May be it makes sense to use single-operation callbacks if
> another operation input is NULL. I guess it could be
> convenient in some cases. Just an idea.
>
There is a plan to deprecate indirect query and update functions
and replace them with a single query_update.
> > + ret = ops->action_handle_query_update(dev, handle, update, query,
> > + mode, error);
> > + return flow_err(port_id, ret, error);
> > +}
> > +
> > +int
> > +rte_flow_async_action_handle_query_update(uint16_t port_id, uint32_t
> queue_id,
> > + const struct rte_flow_op_attr *attr,
> > + struct rte_flow_action_handle *handle,
> > + const void *update, void *query,
> > + enum rte_flow_query_update_mode mode,
> > + void *user_data,
> > + struct rte_flow_error *error)
> > +{
> > + int ret;
> > + struct rte_eth_dev *dev;
> > + const struct rte_flow_ops *ops;
> > +
> > + RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
> > + dev = &rte_eth_devices[port_id];
>
> same here
>
Will fix in v9.
> > + ops = rte_flow_ops_get(port_id, error);
> > + if (!ops || !ops->async_action_handle_query_update)
> > + return -ENOTSUP;
>
> same here
>
Will fix in v9.
> > + ret = ops->async_action_handle_query_update(dev, queue_id, attr,
> > + handle, update, query, mode,
> > + user_data, error);
> > + return flow_err(port_id, ret, error);
> > +}
> > diff --git a/lib/ethdev/rte_flow.h b/lib/ethdev/rte_flow.h
> > index b60987db4b..d3fbba5b99 100644
> > --- a/lib/ethdev/rte_flow.h
> > +++ b/lib/ethdev/rte_flow.h
> > @@ -5622,6 +5622,109 @@
> rte_flow_async_action_handle_query(uint16_t port_id,
> > void *user_data,
> > struct rte_flow_error *error);
> >
> > +/**
> > + * @warning
> > + * @b EXPERIMENTAL: this API may change without prior notice.
> > + *
> > + * Query and update operational mode.
> > + *
> > + * RTE_FLOW_QU_QUERY_FIRST
> > + * Force port to query action before update.
> > + * RTE_FLOW_QU_UPDATE_FIRST
> > + * Force port to update action before query.
>
> I'm sorry, but I strongly dislike enum members duplication
> here. I don't understand why we need to duplicate it and why
> we can't document it in a right way below.
>
I don't understand where the duplication is.
Query and update operations are atomic for application,
but in hardware these are 2 separate operations that reference hardware object.
The operations execution order can have different results on object state.
When application asks both actions it must explicitly specify execution order for hardware.
> > + *
> > + * @see rte_flow_action_handle_query_update()
> > + * @see rte_flow_async_action_handle_query_update()
> > + */
> > +enum rte_flow_query_update_mode {
> > + RTE_FLOW_QU_QUERY_FIRST, /**< Query before update. */
> > + RTE_FLOW_QU_UPDATE_FIRST, /**< Query after update. */
> > +};
> > +
> > +/**
> > + * @warning
> > + * @b EXPERIMENTAL: this API may change without prior notice.
> > + *
> > + * Query and/or update indirect flow action.
> > + * If both query and update not NULL, the function atomically
> > + * queries and updates indirect action. Query and update are carried in
> order
> > + * specified in the mode parameter.
> > + *
> > + * @param port_id
> > + * Port identifier of Ethernet device.
> > + * @param handle
> > + * Handle for the indirect action object to be updated.
> > + * @param update
> > + * If not NULL, update profile specification used to modify the action
> > + * pointed by handle.
> > + * @param query
> > + * If not NULL pointer to storage for the associated query data type.
> > + * @param mode
> > + * Operational mode.
> > + * Required if both *update* and *query* are not NULL.
>
> It should be a part of generic checks in
> rte_flow_action_handle_query_update() implementation.
>
Will fix in v9.
> > + * @param error
> > + * Perform verbose error reporting if not NULL.
> > + * PMDs initialize this structure in case of error only.
> > + *
> > + * @return
> > + * 0 on success, a negative errno value otherwise and rte_errno is set.
> > + * - (-ENODEV) if *port_id* invalid.
> > + * - (-ENOTSUP) if underlying device does not support this functionality.
> > + * - (-EINVAL) if *handle* invalid or both *query* and *update* are
> NULL.
> > + */
> > +__rte_experimental
> > +int
> > +rte_flow_action_handle_query_update(uint16_t port_id,
> > + struct rte_flow_action_handle *handle,
> > + const void *update, void *query,
> > + enum rte_flow_query_update_mode mode,
> > + struct rte_flow_error *error);
> > +
> > +/**
> > + * @warning
> > + * @b EXPERIMENTAL: this API may change without prior notice.
> > + *
> > + * Enqueue async indirect flow action query and/or update
> > + *
> > + * @param port_id
> > + * Port identifier of Ethernet device.
> > + * @param queue_id
> > + * Flow queue which is used to update the rule.
> > + * @param attr
> > + * Indirect action update operation attributes.
> > + * @param handle
> > + * Handle for the indirect action object to be updated.
> > + * @param update
> > + * If not NULL, update profile specification used to modify the action
> > + * pointed by handle.
> > + * @param query
> > + * If not NULL, pointer to storage for the associated query data type.
> > + * Query result returned on async completion event.
> > + * @param mode
> > + * Operational mode.
> > + * Required if both *update* and *query* are not NULL.
>
> It should be a part of generic checks in
> rte_flow_action_handle_query_update() implementation.
>
Will fix in v9.
> > + * @param user_data
> > + * The user data that will be returned on async completion event.
> > + * @param error
> > + * Perform verbose error reporting if not NULL.
> > + * PMDs initialize this structure in case of error only.
> > + *
> > + * @return
> > + * 0 on success, a negative errno value otherwise and rte_errno is set.
> > + * - (-ENODEV) if *port_id* invalid.
> > + * - (-ENOTSUP) if underlying device does not support this functionality.
> > + * - (-EINVAL) if *handle* invalid or both *update* and *query* are
> NULL.
> > + */
> > +__rte_experimental
> > +int
> > +rte_flow_async_action_handle_query_update(uint16_t port_id, uint32_t
> queue_id,
> > + const struct rte_flow_op_attr *attr,
> > + struct rte_flow_action_handle *handle,
> > + const void *update, void *query,
> > + enum rte_flow_query_update_mode mode,
> > + void *user_data,
> > + struct rte_flow_error *error);
> > +
> > #ifdef __cplusplus
> > }
> > #endif
>
> [snip]
>
next prev parent reply other threads:[~2023-02-02 10:24 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-12-21 7:35 [PATCH " Gregory Etelson
2022-12-21 7:35 ` [PATCH 2/2] ethdev: add quota flow action and item Gregory Etelson
2023-01-08 13:47 ` Ori Kam
2023-01-04 9:56 ` [PATCH 1/2] ethdev: add query_update sync and async function calls Ori Kam
2023-01-11 9:28 ` Gregory Etelson
2023-01-11 9:22 ` [PATCH v2 " Gregory Etelson
2023-01-11 9:22 ` [PATCH v2 2/2] ethdev: add quota flow action and item Gregory Etelson
2023-01-11 12:20 ` [PATCH v3 1/2] ethdev: add query_update sync and async function calls Gregory Etelson
2023-01-11 12:20 ` [PATCH v3 2/2] ethdev: add quota flow action and item Gregory Etelson
2023-01-18 10:31 ` [PATCH v4 1/2] ethdev: add query_update sync and async function calls Gregory Etelson
2023-01-18 10:31 ` [PATCH v4 2/2] ethdev: add quota flow action and item Gregory Etelson
2023-01-18 14:03 ` Thomas Monjalon
2023-01-19 9:13 ` Gregory Etelson
2023-01-19 9:31 ` Thomas Monjalon
2023-01-19 9:39 ` Gregory Etelson
2023-01-18 13:56 ` [PATCH v4 1/2] ethdev: add query_update sync and async function calls Thomas Monjalon
2023-01-18 17:34 ` Gregory Etelson
2023-01-19 8:44 ` Thomas Monjalon
2023-01-19 13:25 ` [PATCH v5 " Gregory Etelson
2023-01-19 13:25 ` [PATCH v5 2/2] ethdev: add quota flow action and item Gregory Etelson
2023-01-19 16:47 ` [PATCH v6 1/2] ethdev: add query_update sync and async function calls Gregory Etelson
2023-01-19 16:47 ` [PATCH v6 2/2] ethdev: add quota flow action and item Gregory Etelson
2023-01-20 8:52 ` Andrew Rybchenko
2023-01-24 9:26 ` Gregory Etelson
2023-01-20 8:35 ` [PATCH v6 1/2] ethdev: add query_update sync and async function calls Andrew Rybchenko
2023-01-20 10:46 ` Gregory Etelson
2023-01-20 11:22 ` Andrew Rybchenko
2023-01-20 16:50 ` Gregory Etelson
2023-02-01 11:00 ` Andrew Rybchenko
2023-02-01 14:03 ` Gregory Etelson
2023-01-24 9:37 ` [PATCH v7 " Gregory Etelson
2023-01-24 9:37 ` [PATCH v7 2/2] ethdev: add quota flow action and item Gregory Etelson
2023-02-01 11:22 ` Andrew Rybchenko
2023-02-01 15:09 ` Gregory Etelson
2023-02-01 11:16 ` [PATCH v7 1/2] ethdev: add query_update sync and async function calls Andrew Rybchenko
2023-02-01 14:52 ` Gregory Etelson
2023-02-01 15:16 ` [PATCH v8 " Gregory Etelson
2023-02-01 15:16 ` [PATCH v8 2/2] ethdev: add quota flow action and item Gregory Etelson
2023-02-02 9:17 ` Andrew Rybchenko
2023-02-01 17:30 ` [PATCH v8 1/2] ethdev: add query_update sync and async function calls Ori Kam
2023-02-02 9:15 ` Andrew Rybchenko
2023-02-02 10:24 ` Gregory Etelson [this message]
2023-02-02 10:30 ` Andrew Rybchenko
2023-02-02 10:44 ` Gregory Etelson
2023-02-02 10:47 ` Andrew Rybchenko
2023-02-02 11:54 ` [PATCH v9 1/2] ethdev: add query and update " Gregory Etelson
2023-02-02 11:54 ` [PATCH v9 2/2] ethdev: add quota flow action and item Gregory Etelson
2023-02-02 12:13 ` [PATCH v9 1/2] ethdev: add query and update sync and async function calls Andrew Rybchenko
2023-02-09 19:16 ` Gregory Etelson
2023-02-02 13:47 ` [PATCH v10 " Gregory Etelson
2023-02-02 13:47 ` [PATCH v10 2/2] ethdev: add quota flow action and item Gregory Etelson
2023-02-19 17:04 ` Thomas Monjalon
2023-02-07 16:03 ` [PATCH v10 1/2] ethdev: add query and update sync and async function calls Gregory Etelson
2023-02-09 15:13 ` Ferruh Yigit
2023-02-09 16:10 ` Gregory Etelson
2023-02-12 11:13 ` Gregory Etelson
2023-02-13 12:30 ` Gregory Etelson
2023-02-16 15:43 ` 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=IA1PR12MB6332C9F3CB7546DC41EC6197A5D69@IA1PR12MB6332.namprd12.prod.outlook.com \
--to=getelson@nvidia.com \
--cc=andrew.rybchenko@oktetlabs.ru \
--cc=dev@dpdk.org \
--cc=ferruh.yigit@amd.com \
--cc=matan@nvidia.com \
--cc=orika@nvidia.com \
--cc=rasland@nvidia.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).