net: pkt_filter: Add priority filters to enable quality of service #93341

ClaCodes · 2025-07-18T16:17:54Z

Quality of Service for Ethernet

To enable quality-of-service (QoS) applications, add net_pkt_filters with result NET_CONTINUE, that allow to change the priority of an incoming packet dynamically. Then the processing of a network packet will happen on the respective traffic-class-queue. This enables a gradual degrading of availabilit of a service based on its priority (user application configurable).

Results (Sample in this PR)

With QoS

c (x) := command service for priority x (high means higher priority)

e (x) := echo service for priority x (high means higher priority)

Without QoS

c (x) := command service for priority x (high means higher priority)

e (x) := echo service for priority x (high means higher priority)

Future Work

Implement deferring before net_conn_input
QoS for UDP
QoS for TCP

rlubos

Some initial feedback (on the topmost 2 commits).

subsys/net/ip/net_tc.c

subsys/net/ip/net_core.c

ClaCodes · 2025-07-24T16:49:56Z

@rlubos
I reverted along the lines of discussion.
Note that, if we go with the current approach of the RFC, we would have to add the snippet to multiple places:

+               update_priority_l2(pkt);
+               if (!being_processed_by_correct_thread(pkt)) {
+                       net_queue_rx(net_pkt_iface(pkt), pkt);
+                       return NET_OK;
+               }

For example after extracting the l2-header but before: STRUCT_SECTION_FOREACH(net_l3_register, l3) { such that for example arp packets can be offloaded on to another traffic class queue.
and after ip-header processing etc ...

I believe, that a loop and a if-else-chain or a switch would be more readable. Presumably the added computation cost is negligible?

void process_data(struct net_pkt *pkt)
{
    // switch would only work if raw packet sockets not needed for loop back etc.
    switch (pkt->step) {
    case STEP_1:
        do_step_1();
        return NET_CONTINUE;
    case STEP_2:
        do_step_2();
        return NET_CONTINUE;
    case STEP_3:
        do_step_3();
        return NET_CONTINUE;
        // ...
    }
}

void processing_data(struct net_pkt *pkt)
{
    enum net_verdict verdict = NET_CONTINUE;
    do {
        verdict = process_data(pkt);
        if (verdict != NET_CONTINUE) {
            break;
        }
        update_prio(pkt, pkt->step);
        if (!on_correct_thread(pkt)) {
            net_queue_rx(pkt);
            verdict = NET_OK;
        }
    } while (verdict == NET_CONTINUE);
    // ... unref
}

If we do not do this, it will be very hard to follow at which point the processing can be interrupted and then recontinued later. This may be an easy source of bugs.

ClaCodes · 2025-08-16T09:03:38Z

Rebased

Note: contains all commits from #93246 and adds two commit

sonarqubecloud · 2025-08-16T09:39:28Z

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
1.6% Duplication on New Code

See analysis details on SonarQube Cloud

rlubos

I think packet rescheduling logic turned out pretty straightforward and looks fine, however I think we need to think on how would the priority updating look like in practice.

subsys/net/ip/net_tc.c

subsys/net/ip/net_core.c

tpambor · 2025-09-24T15:41:44Z

@ClaCodes Could you rebase this PR as the prerequisites #93246 and #93050 have been merged

ClaCodes · 2025-10-19T18:46:54Z

rebased
fixed nits and typos
Added to net filter shell command -> led to net: tc-mapping: Fix SKIP_FOR_HIGH_PRIO #97877 (additional commit)
Added priority filters to the existing sample in net/pkt_filter. Running net filter on that sample now gives:

Rule  Type        Verdict   Pkt-Prio  Queue  Thread-Prio  Tests
[ 1]  recv        CONTINUE         1      0            1  1    iface[1]
[ 2]  recv        CONTINUE         7      1         SKIP  2    iface[1],eth type[0x88f7]
[ 3]  recv        CONTINUE         2      0            1  2    iface[1],eth type[0x8100]
[ 4]  recv        CONTINUE         1      0            1  2    iface[2],eth vlan type[0x0806]
[ 5]  recv        CONTINUE         1      0            1  2    iface[3],eth vlan type[0x0806]
[ 6]  recv        OK             N/A    N/A          N/A  3    iface[2],eth vlan type[0x0800],size max[200]
[ 7]  recv        OK             N/A    N/A          N/A  3    iface[3],eth vlan type[0x0800],size min[100]
[ 8]  recv        OK             N/A    N/A          N/A  1    iface[1]
[ 9]  recv        OK             N/A    N/A          N/A  2    iface[2],eth vlan type[0x0806]
[10]  recv        OK             N/A    N/A          N/A  2    iface[3],eth vlan type[0x0806]
[11]  recv        DROP           N/A    N/A          N/A  0

ClaCodes · 2025-10-19T18:52:31Z

inverted boolean logic and fixed a problem introduced in earlier revision

jukkar · 2025-10-20T07:07:21Z

samples/net/pkt_filter/src/main.c

-
+/* Match ethernet packets for the precision time protocol (PTP) */
+static NPF_ETH_TYPE_MATCH(match_ptype_ptp, NET_ETH_PTYPE_PTP);
+/* Match ethernet packets for the virual local area network (VLAN) */


typo s/virual/virtual/

jukkar · 2025-10-20T07:07:46Z

samples/net/pkt_filter/src/main.c

-static NPF_RULE(large_ipv4_pkt, NET_OK, ipv4_packet2, minsize_100,
-		match_iface_vlan2);
+static NPF_SIZE_MIN(match_bigger_100, 100);
+/* Match virutal internet protocol traffic */


typo s/virutal/virtual/

jukkar · 2025-10-20T07:07:54Z

samples/net/pkt_filter/src/main.c

+static NPF_SIZE_MIN(match_bigger_100, 100);
+/* Match virutal internet protocol traffic */
+static NPF_ETH_VLAN_TYPE_MATCH(match_ipv4_vlan, NET_ETH_PTYPE_IP);
+/* Match virutal address resolution protocol (ARP) traffic */


Copilot

Pull Request Overview

Enable QoS by introducing priority-only packet filter rules that adjust packet priority (and thus TC queue) without deciding the final verdict, plus supporting infrastructure and a demonstrative Ethernet sample.

Add NET_CONTINUE handling in packet filters and a new NPF_PRIORITY macro to set packet priority and continue evaluation
Refactor TC thread priority computation with base/spread configs; add helpers and unify immediate-queue checks
Introduce a QoS Ethernet sample, enhance shell filter output to show priority, queue, and thread priority

Reviewed Changes

Copilot reviewed 17 out of 17 changed files in this pull request and generated 12 comments.

Show a summary per file

File	Description
subsys/net/pkt_filter/base.c	Add NET_CONTINUE path to set packet priority and continue rule evaluation
include/zephyr/net/net_pkt_filter.h	Add priority field to npf_rule and a new NPF_PRIORITY macro
subsys/net/lib/shell/filter.c	Extend shell output to display priority-to-TC mapping and thread priorities
subsys/net/ip/net_tc.c	Replace lookup tables with computed thread priorities using base/spread; add exported helpers
subsys/net/ip/net_private.h	Add helper APIs and inline “immediate” checks for RX/TX
subsys/net/ip/net_if.c	Use new net_tc_tx_is_immediate helper
subsys/net/ip/net_core.c	Use new net_tc_rx_is_immediate helper
subsys/net/ip/Kconfig	Add configurable TX/RX thread priority spread options
samples/net/qos/ethernet/*	New QoS Ethernet sample demonstrating priority-based degradation and echo/command services
samples/net/pkt_filter/src/main.c	Add and order priority rules; comments and rule set updates for VLAN sample
samples/net/pkt_filter/prj.conf	Enable RX TCs and skip-for-high-prio in sample
doc/connectivity/networking/api/net_pkt_filter.rst	Document NET_CONTINUE semantics and NPF_PRIORITY usage with example
samples/net/qos/* (docs/yaml/cmake)	Sample metadata, build, and documentation

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

Copilot · 2025-10-20T07:21:45Z

samples/net/qos/ethernet/src/main.c

+static NPF_PRIORITY(priority_rule_129, NET_PRIORITY_IC, npf_129);
+static NPF_PRIORITY(priority_rule_130, NET_PRIORITY_NC, npf_130);
+
+static uint32_t simulated_work_time; /* data race ! */


simulated_work_time is read/written from multiple threads (RX handler and TX send path) without synchronization, which is a data race. Use atomic_t with atomic_set/atomic_get, or protect access with a mutex; for example: replace with static atomic_t simulated_work_time; set with atomic_set(&simulated_work_time, work); and read with posix_cpu_hold((uint32_t)atomic_get(&simulated_work_time)).

Yes, as I noted with the comment. But it is very unlikely, that is written while readers are active, and I would like to avoid them blocking each other.

Copilot · 2025-10-20T07:21:45Z

samples/net/qos/ethernet/src/main.c

+
+static int net_if_fake_send(const struct device *dev, struct net_pkt *pkt)
+{
+	LOG_INF("sending service %d, pkt %p", net_pkt_ll_proto_type(pkt), pkt);


net_pkt_ll_proto_type() returns uint16_t, but the format specifier uses %d (signed int). Use %u or PRIu16 to avoid incorrect logging on larger values: LOG_INF("sending service %u, pkt %p", (unsigned)net_pkt_ll_proto_type(pkt), pkt);

Suggested change

LOG_INF("sending service %d, pkt %p", net_pkt_ll_proto_type(pkt), pkt);

LOG_INF("sending service %u, pkt %p", net_pkt_ll_proto_type(pkt), pkt);

Copilot · 2025-10-20T07:21:46Z

samples/net/pkt_filter/src/main.c

-
+/* Match ethernet packets for the precision time protocol (PTP) */
+static NPF_ETH_TYPE_MATCH(match_ptype_ptp, NET_ETH_PTYPE_PTP);
+/* Match ethernet packets for the virual local area network (VLAN) */


Corrected spelling of 'virual' to 'virtual'.

Copilot · 2025-10-20T07:21:46Z

samples/net/pkt_filter/src/main.c

-static NPF_RULE(large_ipv4_pkt, NET_OK, ipv4_packet2, minsize_100,
-		match_iface_vlan2);
+static NPF_SIZE_MIN(match_bigger_100, 100);
+/* Match virutal internet protocol traffic */


Corrected spelling of 'virutal' to 'virtual'.

Copilot · 2025-10-20T07:21:46Z

samples/net/pkt_filter/src/main.c

+static NPF_SIZE_MIN(match_bigger_100, 100);
+/* Match virutal internet protocol traffic */
+static NPF_ETH_VLAN_TYPE_MATCH(match_ipv4_vlan, NET_ETH_PTYPE_IP);
+/* Match virutal address resolution protocol (ARP) traffic */


Corrected spelling of 'virutal' to 'virtual'.

Copilot · 2025-10-20T07:21:47Z

include/zephyr/net/net_pkt_filter.h

-	struct npf_test *tests[];	/**< pointers to @ref npf_test instances */
+	sys_snode_t node;         /**< Slist rule list node */
+	enum net_verdict result;  /**< result if all tests pass */
+	uint8_t priority;         /**< priority in case of NET_CONTINUE */


[nitpick] Use the strong type enum net_priority for the priority field to improve type safety and self-documentation: change to enum net_priority priority; and adjust NPF_PRIORITY macro accordingly.

Suggested change

uint8_t priority; /**< priority in case of NET_CONTINUE */

enum net_priority priority; /**< priority in case of NET_CONTINUE */

Copilot · 2025-10-20T07:21:48Z

subsys/net/ip/net_tc.c

+int net_tc_tx_thread_priority(int tc)
 {
-	/* Initial implementation just maps the traffic class to certain queue.
-	 * If there are less queues than classes, then map them into
-	 * some specific queue.
-	 *
-	 * Lower value in this table means higher thread priority. The
-	 * value is used as a parameter to K_PRIO_COOP() or K_PRIO_PREEMPT()
-	 * which converts it to actual thread priority.
-	 *
-	 * Higher traffic class value means higher priority queue. This means
-	 * that thread_priorities[7] value should contain the highest priority
-	 * for the TX queue handling thread.
-	 *
-	 * For example, if NET_TC_TX_COUNT = 8, which is the maximum number of
-	 * traffic classes, then this priority array will contain following
-	 * values if preemptive priorities are used:
-	 *      7, 6, 5, 4, 3, 2, 1, 0
-	 * and
-	 *      14, 13, 12, 11, 10, 9, 8, 7
-	 * if cooperative priorities are used.
-	 *
-	 * Then these will be converted to following thread priorities if
-	 * CONFIG_NET_TC_THREAD_COOPERATIVE is enabled:
-	 *      -1, -2, -3, -4, -5, -6, -7, -8
-	 *
-	 * and if CONFIG_NET_TC_THREAD_PREEMPTIVE is enabled, following thread
-	 * priorities are used:
-	 *       7, 6, 5, 4, 3, 2, 1, 0
-	 *
-	 * This means that the lowest traffic class 1, will have the lowest
-	 * cooperative priority -1 for coop priorities and 7 for preemptive
-	 * priority.
-	 */
-	static const uint8_t thread_priorities[] = {
-		LISTIFY(NET_TC_TX_COUNT, PRIO_TX, (,))
-	};
-
-	BUILD_ASSERT(NET_TC_TX_COUNT <= CONFIG_NUM_COOP_PRIORITIES,
-		     "Too many traffic classes");
-
-	NET_ASSERT(tc < ARRAY_SIZE(thread_priorities));
-
-	return thread_priorities[tc];
+	int priority;
+	int thread_priority;
+
+	BUILD_ASSERT(BASE_PRIO_TX >= PRIO_SPREAD_TX * (NET_TC_TX_COUNT - 1));
+	thread_priority = BASE_PRIO_TX - PRIO_SPREAD_TX * tc;
+
+	priority = IS_ENABLED(CONFIG_NET_TC_THREAD_COOPERATIVE) ?
+		K_PRIO_COOP(thread_priority) :
+		K_PRIO_PREEMPT(thread_priority);
+	return priority;
 }


Add a runtime range check for tc to prevent out-of-range access in case of future callers: e.g., NET_ASSERT(tc >= 0 && tc < NET_TC_TX_COUNT); before using tc.

Copilot · 2025-10-20T07:21:48Z

subsys/net/ip/net_tc.c

+int net_tc_rx_thread_priority(int tc)
 {
-	static const uint8_t thread_priorities[] = {
-		LISTIFY(NET_TC_RX_COUNT, PRIO_RX, (,))
-	};
+	int priority;
+	int thread_priority;

-	BUILD_ASSERT(NET_TC_RX_COUNT <= CONFIG_NUM_COOP_PRIORITIES,
-		     "Too many traffic classes");
+	BUILD_ASSERT(BASE_PRIO_RX >= PRIO_SPREAD_RX * (NET_TC_RX_COUNT - 1));
+	thread_priority = BASE_PRIO_RX - PRIO_SPREAD_RX * tc;

-	NET_ASSERT(tc < ARRAY_SIZE(thread_priorities));
-
-	return thread_priorities[tc];
+	priority = IS_ENABLED(CONFIG_NET_TC_THREAD_COOPERATIVE) ?
+		K_PRIO_COOP(thread_priority) :
+		K_PRIO_PREEMPT(thread_priority);
+	return priority;
 }


Add a runtime range check for tc to ensure it is within [0, NET_TC_RX_COUNT): e.g., NET_ASSERT(tc >= 0 && tc < NET_TC_RX_COUNT);

Copilot · 2025-10-20T07:21:48Z

subsys/net/ip/Kconfig

+config NET_TC_TX_THREAD_PRIO_SPREAD
+	int "Transmit traffic class thread priority spread"
+	default 1


Constrain these integers with a non-zero positive range to avoid invalid configurations (e.g., add range 1 255 or an appropriate upper bound): range 1 255.

subsys/net/ip/Kconfig

To enable quality-of-service (QoS) applications, allow filters to modify the priority of an incoming packet. This allows that processing can be done on tc-queue-threads with different priorities. Signed-off-by: Cla Mattia Galliard <[email protected]>

Add documentation for newly added priority rules. Signed-off-by: Cla Mattia Galliard <[email protected]>

Add some priority filters to the sample for demonstration purpose. Signed-off-by: Cla Mattia Galliard <[email protected]>

Allow to spread tc threads by more then one priority level and cleanup the computation of the priority. Signed-off-by: Cla Mattia Galliard <[email protected]>

Adjust table to print newly available information. Signed-off-by: Cla Mattia Galliard <[email protected]>

Add a sample, that shows the power of quality of service in Zephyr. Signed-off-by: Cla Mattia Galliard <[email protected]>

sonarqubecloud · 2025-10-20T20:09:38Z

Quality Gate passed

Issues
4 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

ClaCodes · 2025-10-20T20:11:49Z

rebased
fix nits, suggestions and typos

pdgendt

In the future, try to avoid force pushing a rebase combined with changes. This makes the "compare" function in Github useless.

ClaCodes · 2025-10-21T16:08:12Z

In the future, try to avoid force pushing a rebase combined with changes. This makes the "compare" function in Github useless.

Will do. Sorry for causing overhead.

zephyrbot added the area: Networking label Jul 18, 2025

zephyrbot requested review from jukkar, pdgendt, rlubos and ssharks July 18, 2025 16:18

zephyrbot assigned rlubos and jukkar Jul 18, 2025

ClaCodes force-pushed the feature/delay_processing branch 4 times, most recently from 26d64cc to a4849f5 Compare July 19, 2025 12:37

rlubos reviewed Jul 21, 2025

View reviewed changes

ClaCodes force-pushed the feature/delay_processing branch 5 times, most recently from c5f9c6c to 9bd6ae9 Compare July 24, 2025 16:36

ClaCodes force-pushed the feature/delay_processing branch from 9bd6ae9 to d88b7ac Compare July 25, 2025 05:56

ClaCodes force-pushed the feature/delay_processing branch from d88b7ac to 80f5552 Compare August 16, 2025 09:03

ClaCodes mentioned this pull request Aug 16, 2025

net_pkt: Store meta-information of processing in the net_pkt #93298

Closed

ClaCodes force-pushed the feature/delay_processing branch from 80f5552 to a69163e Compare August 16, 2025 09:35

ClaCodes requested a review from rlubos August 16, 2025 09:37

rlubos reviewed Aug 18, 2025

View reviewed changes

subsys/net/ip/net_tc.c Outdated Show resolved Hide resolved

subsys/net/ip/net_core.c Outdated Show resolved Hide resolved

ClaCodes force-pushed the feature/delay_processing branch from a69163e to 43cf73c Compare September 27, 2025 16:30

zephyrbot added area: Logging area: Samples Samples labels Sep 27, 2025

ClaCodes force-pushed the feature/delay_processing branch from 6cd1580 to 76d1a0d Compare October 19, 2025 18:37

ClaCodes force-pushed the feature/delay_processing branch from 76d1a0d to 3470ead Compare October 19, 2025 18:50

ClaCodes force-pushed the feature/delay_processing branch 2 times, most recently from 497c236 to 2f1e4f1 Compare October 19, 2025 19:17

ClaCodes mentioned this pull request Oct 19, 2025

net: tc-mapping: Fix SKIP_FOR_HIGH_PRIO #97877

Merged

ClaCodes force-pushed the feature/delay_processing branch 2 times, most recently from 8b436fa to 7df34e2 Compare October 20, 2025 04:38

jukkar reviewed Oct 20, 2025

View reviewed changes

pdgendt requested a review from Copilot October 20, 2025 07:17

Copilot AI reviewed Oct 20, 2025

View reviewed changes

ClaCodes added 6 commits October 20, 2025 22:03

doc: net_pkt_filter: Document priority rules

4cda66c

Add documentation for newly added priority rules. Signed-off-by: Cla Mattia Galliard <[email protected]>

samples: net: pkt_filter: Add priority filters

98d4262

Add some priority filters to the sample for demonstration purpose. Signed-off-by: Cla Mattia Galliard <[email protected]>

net: tc: allow to spread threads by more then 1 priority level

18f8ed5

Allow to spread tc threads by more then one priority level and cleanup the computation of the priority. Signed-off-by: Cla Mattia Galliard <[email protected]>

net: shell: filter: Update table

8d0a140

Adjust table to print newly available information. Signed-off-by: Cla Mattia Galliard <[email protected]>

samples: net: QoS: Add quality of service sample

32b6758

Add a sample, that shows the power of quality of service in Zephyr. Signed-off-by: Cla Mattia Galliard <[email protected]>

ClaCodes force-pushed the feature/delay_processing branch from 7df34e2 to 32b6758 Compare October 20, 2025 20:03

zephyrbot requested a review from jukkar October 20, 2025 20:05

ClaCodes requested a review from tpambor October 20, 2025 20:13

jukkar approved these changes Oct 21, 2025

View reviewed changes

pdgendt approved these changes Oct 21, 2025

View reviewed changes

rlubos approved these changes Oct 21, 2025

View reviewed changes

tpambor approved these changes Oct 21, 2025

View reviewed changes

jhedberg merged commit 630a496 into zephyrproject-rtos:main Oct 21, 2025
27 checks passed

ClaCodes deleted the feature/delay_processing branch October 28, 2025 06:06

	LOG_INF("sending service %d, pkt %p", net_pkt_ll_proto_type(pkt), pkt);
	LOG_INF("sending service %u, pkt %p", net_pkt_ll_proto_type(pkt), pkt);

	uint8_t priority; /*< priority in case of NET_CONTINUE /
	enum net_priority priority; /*< priority in case of NET_CONTINUE /

net: pkt_filter: Add priority filters to enable quality of service #93341

net: pkt_filter: Add priority filters to enable quality of service #93341

Uh oh!

Conversation

ClaCodes commented Jul 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Quality of Service for Ethernet

Results (Sample in this PR)

With QoS

Without QoS

Future Work

Uh oh!

rlubos left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

ClaCodes commented Jul 24, 2025

Uh oh!

ClaCodes commented Aug 16, 2025

Uh oh!

sonarqubecloud bot commented Aug 16, 2025

Quality Gate passed

Uh oh!

rlubos left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

tpambor commented Sep 24, 2025

Uh oh!

ClaCodes commented Oct 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ClaCodes commented Oct 19, 2025

Uh oh!

jukkar Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

jukkar Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

jukkar Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Copilot AI Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

ClaCodes Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 20, 2025

Choose a reason for hiding this comment

Uh oh!

ClaCodes commented Jul 18, 2025 •

edited

Loading

ClaCodes commented Oct 19, 2025 •

edited

Loading