diff mbox series

[1/3] ethdev: add strict queue to pre-configuration flow hints

Message ID 20220921145409.511328-2-michaelba@nvidia.com (mailing list archive)
State Superseded, archived
Delegated to: Andrew Rybchenko
Headers
Series ethdev: AGE action preparation |

Checks

Context Check Description
ci/checkpatch success coding style OK

Commit Message

Michael Baum Sept. 21, 2022, 2:54 p.m. UTC 
  The data-path focused flow rule management can manage flow rules in more
optimized way than traditional one by using hints provided by
application in initialization phase.

In addition to the current hints we have in port attr, more hints could
be provided by application about its behaviour.

One example is how the application do with the same flow rule ?
A. create/destroy flow on same queue but query flow on different queue
   or queue-less way (i.e, counter query)
B. All flow operations will be exactly on the same queue, by which PMD
   could be in more optimized way then A because resource could be
   isolated and access based on queue, without lock, for example.

This patch add flag about above situation and could be extended to cover
more situations.

Signed-off-by: Michael Baum <michaelba@nvidia.com>
---
 app/test-pmd/cmdline_flow.c                 | 10 ++++++++++
 doc/guides/testpmd_app_ug/testpmd_funcs.rst |  4 ++--
 lib/ethdev/rte_flow.h                       | 14 ++++++++++++++
 3 files changed, 26 insertions(+), 2 deletions(-)

Comments

Ori Kam Sept. 29, 2022, 12:04 p.m. UTC | #1 
Hi Michael,

> -----Original Message-----
> From: Michael Baum <michaelba@nvidia.com>
> Sent: Wednesday, 21 September 2022 17:54
> 
> The data-path focused flow rule management can manage flow rules in more
> optimized way than traditional one by using hints provided by
> application in initialization phase.
> 
> In addition to the current hints we have in port attr, more hints could
> be provided by application about its behaviour.
> 
> One example is how the application do with the same flow rule ?
> A. create/destroy flow on same queue but query flow on different queue
>    or queue-less way (i.e, counter query)
> B. All flow operations will be exactly on the same queue, by which PMD
>    could be in more optimized way then A because resource could be
>    isolated and access based on queue, without lock, for example.
> 
> This patch add flag about above situation and could be extended to cover
> more situations.
> 
> Signed-off-by: Michael Baum <michaelba@nvidia.com>
> ---

Acked-by: Ori Kam <orika@nvidia.com>
Thanks,
Ori
diff mbox series

Patch

diff --git a/app/test-pmd/cmdline_flow.c b/app/test-pmd/cmdline_flow.c
index 7f50028eb7..a982083d27 100644
--- a/app/test-pmd/cmdline_flow.c
+++ b/app/test-pmd/cmdline_flow.c
@@ -219,6 +219,7 @@  enum index {
 	CONFIG_COUNTERS_NUMBER,
 	CONFIG_AGING_OBJECTS_NUMBER,
 	CONFIG_METERS_NUMBER,
+	CONFIG_FLAGS,
 
 	/* Indirect action arguments */
 	INDIRECT_ACTION_CREATE,
@@ -1081,6 +1082,7 @@  static const enum index next_config_attr[] = {
 	CONFIG_COUNTERS_NUMBER,
 	CONFIG_AGING_OBJECTS_NUMBER,
 	CONFIG_METERS_NUMBER,
+	CONFIG_FLAGS,
 	END,
 	ZERO,
 };
@@ -2667,6 +2669,14 @@  static const struct token token_list[] = {
 		.args = ARGS(ARGS_ENTRY(struct buffer,
 					args.configure.port_attr.nb_meters)),
 	},
+	[CONFIG_FLAGS] = {
+		.name = "flags",
+		.help = "configuration flags",
+		.next = NEXT(next_config_attr,
+			     NEXT_ENTRY(COMMON_UNSIGNED)),
+		.args = ARGS(ARGS_ENTRY(struct buffer,
+					args.configure.port_attr.flags)),
+	},
 	/* Top-level command. */
 	[PATTERN_TEMPLATE] = {
 		.name = "pattern_template",
diff --git a/doc/guides/testpmd_app_ug/testpmd_funcs.rst b/doc/guides/testpmd_app_ug/testpmd_funcs.rst
index 330e34427d..6c12e0286c 100644
--- a/doc/guides/testpmd_app_ug/testpmd_funcs.rst
+++ b/doc/guides/testpmd_app_ug/testpmd_funcs.rst
@@ -3082,7 +3082,7 @@  following sections.
        [queues_number {number}] [queues_size {size}]
        [counters_number {number}]
        [aging_counters_number {number}]
-       [meters_number {number}]
+       [meters_number {number}] [flags {number}]
 
 - Create a pattern template::
    flow pattern_template {port_id} create [pattern_template_id {id}]
@@ -3233,7 +3233,7 @@  for asynchronous flow creation/destruction operations. It is bound to
        [queues_number {number}] [queues_size {size}]
        [counters_number {number}]
        [aging_counters_number {number}]
-       [meters_number {number}]
+       [meters_number {number}] [flags {number}]
 
 If successful, it will show::
 
diff --git a/lib/ethdev/rte_flow.h b/lib/ethdev/rte_flow.h
index a79f1e7ef0..c552771472 100644
--- a/lib/ethdev/rte_flow.h
+++ b/lib/ethdev/rte_flow.h
@@ -4874,6 +4874,12 @@  rte_flow_flex_item_release(uint16_t port_id,
 			   const struct rte_flow_item_flex_handle *handle,
 			   struct rte_flow_error *error);
 
+/**
+ * Indicate all operations for a given flow rule will _strictly_
+ * happen on the same queue (create/destroy/query/update).
+ */
+#define RTE_FLOW_PORT_FLAG_STRICT_QUEUE RTE_BIT32(0)
+
 /**
  * @warning
  * @b EXPERIMENTAL: this API may change without prior notice.
@@ -4902,6 +4908,10 @@  struct rte_flow_port_info {
 	 * @see RTE_FLOW_ACTION_TYPE_METER
 	 */
 	uint32_t max_nb_meters;
+	/**
+	 * Port supported flags (RTE_FLOW_PORT_FLAG_*).
+	 */
+	uint32_t supported_flags;
 };
 
 /**
@@ -4971,6 +4981,10 @@  struct rte_flow_port_attr {
 	 * @see RTE_FLOW_ACTION_TYPE_METER
 	 */
 	uint32_t nb_meters;
+	/**
+	 * Port flags (RTE_FLOW_PORT_FLAG_*).
+	 */
+	uint32_t flags;
 };
 
 /**