DPDK patches and discussions
 help / color / mirror / Atom feed
From: Pavan Nikhilesh Bhagavatula <pbhagavatula@marvell.com>
To: Nitin Saxena <nsaxena@marvell.com>,
	Nithin Kumar Dabilpuram <ndabilpuram@marvell.com>,
	Robin Jarry <rjarry@redhat.com>,
	Christophe Fontaine <cfontain@redhat.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>, Jerin Jacob <jerinj@marvell.com>,
	Nitin Saxena <nsaxena16@gmail.com>
Subject: RE: [PATCH v5 1/2] node: add global node mbuf dynfield
Date: Fri, 18 Apr 2025 19:03:41 +0000	[thread overview]
Message-ID: <CO6PR18MB4084AB7880F9D37F0862C636DEBF2@CO6PR18MB4084.namprd18.prod.outlook.com> (raw)
In-Reply-To: <20250409135554.2180390-2-nsaxena@marvell.com>



> -----Original Message-----
> From: Nitin Saxena <nsaxena@marvell.com>
> Sent: Wednesday, April 9, 2025 7:26 PM
> To: Nithin Kumar Dabilpuram <ndabilpuram@marvell.com>; Pavan Nikhilesh
> Bhagavatula <pbhagavatula@marvell.com>; Robin Jarry
> <rjarry@redhat.com>; Christophe Fontaine <cfontain@redhat.com>
> Cc: dev@dpdk.org; Jerin Jacob <jerinj@marvell.com>; Nitin Saxena
> <nsaxena16@gmail.com>
> Subject: [PATCH v5 1/2] node: add global node mbuf dynfield
> 
> This patch defines rte_node specific dynamic field structure
> (rte_node_mbuf_dynfield_t)
> 
> rte_node_mbuf_dynfield_t structure holds two types of fields
> - Persistent data fields which are preserved across graph walk.
>   Currently size of persistent data fields is zero.
> - Overloadable data fields which are used by any two adjacent nodes.
>   Same fields can be repurposed by any other adjacent nodes
> 
> This dynfield can be also be used by out-of-tree nodes.
> 
> Signed-off-by: Nitin Saxena <nsaxena@marvell.com>
> ---
>  doc/api/doxy-api-index.md              |   3 +-
>  doc/guides/rel_notes/release_25_07.rst |   6 ++
>  lib/node/meson.build                   |   2 +
>  lib/node/node_mbuf_dynfield.c          |  57 +++++++++++
>  lib/node/rte_node_mbuf_dynfield.h      | 132
> +++++++++++++++++++++++++
>  5 files changed, 199 insertions(+), 1 deletion(-)
>  create mode 100644 lib/node/node_mbuf_dynfield.c
>  create mode 100644 lib/node/rte_node_mbuf_dynfield.h
> 
> diff --git a/doc/api/doxy-api-index.md b/doc/api/doxy-api-index.md
> index 5c425a2cb9..763cfb3f3c 100644
> --- a/doc/api/doxy-api-index.md
> +++ b/doc/api/doxy-api-index.md
> @@ -219,7 +219,8 @@ The public API headers are grouped by topics:
>      [ip4_node](@ref rte_node_ip4_api.h),
>      [ip6_node](@ref rte_node_ip6_api.h),
>      [udp4_input_node](@ref rte_node_udp4_input_api.h)
> -
> +  * graph_nodes_mbuf:
> +    [node_mbuf_dynfield](@ref rte_node_mbuf_dynfield.h)

Missing blank line here.

>  - **basic**:
>    [bitops](@ref rte_bitops.h),
>    [approx fraction](@ref rte_approx.h),
> diff --git a/doc/guides/rel_notes/release_25_07.rst
> b/doc/guides/rel_notes/release_25_07.rst
> index 093b85d206..2dc888e65d 100644
> --- a/doc/guides/rel_notes/release_25_07.rst
> +++ b/doc/guides/rel_notes/release_25_07.rst
> @@ -55,6 +55,12 @@ New Features
>       Also, make sure to start the actual text at the margin.
>       =======================================================
> 
> +* **Added rte_node specific global mbuf dynamic field.**
> +
> +  Instead each node registering mbuf dynamic field for its own purpose, a
> +  global structure is added which can be used/overloaded by all nodes
> +  (including out-of-tree nodes). This minimizes footprint of node specific
> mbuf
> +  dynamic field.
> 
>  Removed Items
>  -------------
> diff --git a/lib/node/meson.build b/lib/node/meson.build
> index 0bed97a96c..152fe41129 100644
> --- a/lib/node/meson.build
> +++ b/lib/node/meson.build
> @@ -8,6 +8,7 @@ if is_windows
>  endif
> 
>  sources = files(
> +        'node_mbuf_dynfield.c',
>          'ethdev_ctrl.c',
>          'ethdev_rx.c',
>          'ethdev_tx.c',
> @@ -30,6 +31,7 @@ headers = files(
>          'rte_node_ip4_api.h',
>          'rte_node_ip6_api.h',
>          'rte_node_udp4_input_api.h',
> +        'rte_node_mbuf_dynfield.h',
>  )
> 
>  # Strict-aliasing rules are violated by uint8_t[] to context size casts.
> diff --git a/lib/node/node_mbuf_dynfield.c b/lib/node/node_mbuf_dynfield.c
> new file mode 100644
> index 0000000000..6005e72ed6
> --- /dev/null
> +++ b/lib/node/node_mbuf_dynfield.c
> @@ -0,0 +1,57 @@
> +/* SPDX-License-Identifier: BSD-3-Clause
> + * Copyright(C) 2025 Marvell International Ltd.
> + */
> +#include <rte_memzone.h>
> +#include <rte_node_mbuf_dynfield.h>
> +#include <node_private.h>
> +#include <eal_export.h>
> +
> +#define NODE_MBUF_DYNFIELD_MEMZONE_NAME
> "__rte_node_mbuf_dynfield"
> +
> +struct node_mbuf_dynfield_mz {
> +	int dynfield_offset;
> +};
> +
> +static const struct rte_mbuf_dynfield node_mbuf_dynfield_desc = {
> +	.name = "rte_node_mbuf_dynfield",
> +	.size = sizeof(rte_node_mbuf_dynfield_t),
> +	.align = alignof(rte_node_mbuf_dynfield_t),
> +};
> +
> +RTE_EXPORT_EXPERIMENTAL_SYMBOL(rte_node_mbuf_dynfield_register,
> 25.07);
> +int rte_node_mbuf_dynfield_register(void)
> +{
> +	struct node_mbuf_dynfield_mz *f = NULL;
> +	const struct rte_memzone *mz = NULL;
> +	int dyn_offset;
> +
> +	RTE_BUILD_BUG_ON(sizeof(rte_node_mbuf_dynfield_t) <
> RTE_NODE_MBUF_DYNFIELD_SIZE);
> +	RTE_BUILD_BUG_ON(sizeof(rte_node_mbuf_overload_fields_t) <
> +			 RTE_NODE_MBUF_OVERLOADABLE_FIELDS_SIZE);
> +
> +	mz =
> rte_memzone_lookup(NODE_MBUF_DYNFIELD_MEMZONE_NAME);
> +
> +	if (!mz) {
> +		mz =
> rte_memzone_reserve_aligned(NODE_MBUF_DYNFIELD_MEMZONE_NAME,
> +						 sizeof(struct
> node_mbuf_dynfield_mz),
> +						 SOCKET_ID_ANY, 0,
> +						 RTE_CACHE_LINE_SIZE);
> +		if (!mz) {
> +			node_err("node_mbuf_dyn",
> "rte_memzone_reserve_aligned failed");
> +			return -1;

Please return proper error codes, or set rte_errno.

> +		}
> +		dyn_offset =
> rte_mbuf_dynfield_register(&node_mbuf_dynfield_desc);
> +		if (dyn_offset < 0) {
> +			node_err("node_mbuf_dyn",
> "rte_mbuf_dynfield_register failed");
> +			return -1;

Please return proper error codes, or set rte_errno.

> +		}
> +		f = (struct node_mbuf_dynfield_mz *)mz->addr;
> +		f->dynfield_offset = dyn_offset;
> +
> +		node_dbg("node_mbuf_dyn", "memzone: %s of size: %zu at
> offset : %d",
> +			 mz->name, mz->len, f->dynfield_offset);
> +	} else {
> +		f = (struct node_mbuf_dynfield_mz *)mz->addr;
> +	}
> +	return f->dynfield_offset;
> +}
> diff --git a/lib/node/rte_node_mbuf_dynfield.h
> b/lib/node/rte_node_mbuf_dynfield.h
> new file mode 100644
> index 0000000000..b5a3d6db3f
> --- /dev/null
> +++ b/lib/node/rte_node_mbuf_dynfield.h
> @@ -0,0 +1,132 @@
> +/* SPDX-License-Identifier: BSD-3-Clause
> + * Copyright(C) 2025 Marvell International Ltd.
> + */
> +
> +#ifndef _RTE_GRAPH_MBUF_DYNFIELD_H_
> +#define _RTE_GRAPH_MBUF_DYNFIELD_H_
> +
> +#include <rte_common.h>
> +#include <rte_mbuf.h>
> +#include <rte_mbuf_dyn.h>
> +
> +/**
> + * @file: rte_node_mbuf_dynfield.h
> + *

Please mark the file as experimental
i.e., 
 * @warning
 * @b EXPERIMENTAL:

> + * Defines rte_node specific mbuf dynamic field region
> [rte_node_mbuf_dynfield_t] which
> + * can used by both DPDK in-built and out-of-tree nodes for storing per-mbuf
> + * fields for graph walk
> + */
> +#ifdef __cplusplus
> +extern "C" {
> +#endif
> +
> +/** Size of persistent mbuf fields */
> +#define RTE_NODE_MBUF_PERSISTENT_FIELDS_SIZE          (0)
> +/** Size of overloadable mbuf fields */
> +#define RTE_NODE_MBUF_OVERLOADABLE_FIELDS_SIZE        (8)
> +/** Size of node mbuf dynamic field */
> +#define RTE_NODE_MBUF_DYNFIELD_SIZE     \
> +	(RTE_NODE_MBUF_PERSISTENT_FIELDS_SIZE +
> RTE_NODE_MBUF_OVERLOADABLE_FIELDS_SIZE)

New line here.

> +/**
> + * Node mbuf overloadable data.
> + *
> + * Out-of-tree nodes can repurpose overloadable fields via
> + * rte_node_mbuf_overload_fields_get(mbuf). Overloadable fields are not
> + * preserved and typically can be used with-in two adjacent nodes in the
> graph.
> + */
> +typedef struct rte_node_mbuf_overload_fields {
> +	union {
> +		uint8_t
> data[RTE_NODE_MBUF_OVERLOADABLE_FIELDS_SIZE];
> +	};
> +} rte_node_mbuf_overload_fields_t;
> +
> +/**
> + * rte_node specific mbuf dynamic field structure
> [rte_node_mbuf_dynfield_t]
> + *
> + * It holds two types of fields:
> + * 1. Persistent fields: Fields which are preserved across nodes during graph
> walk.
> + *    - Eg: rx/tx interface etc
> + * 2. Overloadable fields: Fields which can be repurposed by two adjacent
> nodes.
> + */
> +typedef struct rte_node_mbuf_dynfield {
> +	/**
> +	 * Persistent mbuf region across nodes in graph walk
> +	 *
> +	 * These fields are preserved across graph walk and can be accessed by
> +	 * rte_node_mbuf_dynfield_get() in fast path.
> +	 */
> +	union {
> +		uint8_t
> persistent_data[RTE_NODE_MBUF_PERSISTENT_FIELDS_SIZE];
> +	};
> +
> +	/**
> +	 * Overloadable mbuf fields across graph walk. Fields which can
> change
> +	 *
> +	 * Two adjacent nodes (producer, consumer) can use this memory
> region for
> +	 * sharing per-mbuf specific information. Once mbuf leaves a
> consumer node,
> +	 * this region can be repurposed by another sets of [producer,
> consumer] node.
> +	 *
> +	 * In fast path, this region can be accessed by
> rte_node_mbuf_overload_fields_get()
> +	 */
> +	rte_node_mbuf_overload_fields_t overloadable_data;
> +} rte_node_mbuf_dynfield_t;
> +
> +/**
> + * For a given mbuf and dynamic offset, return pointer to
> rte_node_mbuf_dynfield_t *
> + *
> + * @param m
> + *   Mbuf
> + * @param offset
> + *   Dynamic offset returned by @ref rte_node_mbuf_dynfield_register()

Missing return field in doxygen.

> + */
> +__rte_experimental
> +static __rte_always_inline rte_node_mbuf_dynfield_t *
> +rte_node_mbuf_dynfield_get(struct rte_mbuf *m, const int offset)
> +{
> +	return RTE_MBUF_DYNFIELD(m, offset, struct
> rte_node_mbuf_dynfield *);
> +}
> +
> +/**
> + * For a given mbuf and dynamic offset, return pointer to overloadable fields
> + * Nodes can typecast returned pointer to reuse for their own purpose
> + *
> + * @param m
> + *   Mbuf
> + * @param offset
> + *   Dynamic offset returned by @ref rte_node_mbuf_dynfield_register()

Missing return field.

> + */
> +__rte_experimental
> +static __rte_always_inline rte_node_mbuf_overload_fields_t *
> +rte_node_mbuf_overload_fields_get(struct rte_mbuf *m, const int offset)
> +{
> +	struct rte_node_mbuf_dynfield *f = NULL;

rte_node_mbuf_dynfield_t

> +
> +	f = RTE_MBUF_DYNFIELD(m, offset, struct rte_node_mbuf_dynfield
> *);

rte_node_mbuf_dynfield_t

> +
> +	return &(f->overloadable_data);
> +}
> +
> +/**
> + *  Register rte_node specific common mbuf dynamic field region. Can be
> called
> + *  in rte_node_register()->init() function to save offset in node->ctx
> + *
> + *  In process() function, node->ctx can be passed to
> + *  - rte_node_mbuf_dynfield_get(mbuf, offset)
> + *  - rte_node_mbuf_overload_fields_get(mbuf, offset)
> + *
> + *  Can be called multiple times by any number of nodes in init() function.
> + *  - Very first call registers dynamic field and returns offset
> + *  - Subsequent calls return same offset
> + *
> + *  @return
> + *   -1 on error

< 0 on error.
Please return and document proper error codes. 

> + *   dynamic field offset on success
> + */
> +__rte_experimental
> +int rte_node_mbuf_dynfield_register(void);
> +
> +#ifdef __cplusplus
> +}
> +#endif
> +
> +#endif
> --
> 2.43.0


  reply	other threads:[~2025-04-18 19:03 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-04-01  4:20 [PATCH 0/2] node: add mbuf dynamic field for nodes Nitin Saxena
2025-04-01  4:20 ` [PATCH 1/2] node: add global node mbuf dynfield Nitin Saxena
2025-04-01 14:15   ` Stephen Hemminger
2025-04-03 10:27     ` Nitin Saxena
2025-04-04  8:11       ` [EXTERNAL] " Pavan Nikhilesh Bhagavatula
2025-04-04 15:21         ` Stephen Hemminger
2025-04-07  6:01           ` Nitin Saxena
2025-04-01  4:20 ` [PATCH 2/2] node: use node mbuf dynfield in ip4 nodes Nitin Saxena
2025-04-03 10:17 ` [PATCH v2 0/2] node: add mbuf dynamic field for nodes Nitin Saxena
2025-04-03 10:17   ` [PATCH v2 1/2] node: add global node mbuf dynfield Nitin Saxena
2025-04-03 10:17   ` [PATCH v2 2/2] node: use node mbuf dynfield in ip4 nodes Nitin Saxena
2025-04-04  7:12 ` [PATCH v3 0/2] node: add mbuf dynamic field for nodes Nitin Saxena
2025-04-04  7:12   ` [PATCH v3 1/2] node: add global node mbuf dynfield Nitin Saxena
2025-04-04 21:35     ` Stephen Hemminger
2025-04-04  7:12   ` [PATCH v3 2/2] node: use node mbuf dynfield in ip4 nodes Nitin Saxena
2025-04-07  7:47 ` [PATCH v4 0/2] node: add mbuf dynamic field for nodes Nitin Saxena
2025-04-07  7:47   ` [PATCH v4 1/2] node: add global node mbuf dynfield Nitin Saxena
2025-04-07  7:47   ` [PATCH v4 2/2] node: use node mbuf dynfield in ip4 nodes Nitin Saxena
2025-04-08  7:35   ` [PATCH v4 0/2] node: add mbuf dynamic field for nodes David Marchand
2025-04-08  8:42     ` Nitin Saxena
2025-04-09 13:55 ` [PATCH v5 " Nitin Saxena
2025-04-09 13:55   ` [PATCH v5 1/2] node: add global node mbuf dynfield Nitin Saxena
2025-04-18 19:03     ` Pavan Nikhilesh Bhagavatula [this message]
2025-04-09 13:55   ` [PATCH v5 2/2] node: use node mbuf dynfield in ip4 nodes Nitin Saxena

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=CO6PR18MB4084AB7880F9D37F0862C636DEBF2@CO6PR18MB4084.namprd18.prod.outlook.com \
    --to=pbhagavatula@marvell.com \
    --cc=cfontain@redhat.com \
    --cc=dev@dpdk.org \
    --cc=jerinj@marvell.com \
    --cc=ndabilpuram@marvell.com \
    --cc=nsaxena16@gmail.com \
    --cc=nsaxena@marvell.com \
    --cc=rjarry@redhat.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).