[PATCH net-next v9] Add support to set NAPI threaded for individual NAPI

[PATCH net-next v9] Add support to set NAPI threaded for individual NAPI

[Samiullah Khawaja]

A net device has a threaded sysctl that can be used to enable threaded
NAPI polling on all of the NAPI contexts under that device. Allow
enabling threaded NAPI polling at individual NAPI level using netlink.

Extend the netlink operation `napi-set` and allow setting the threaded
attribute of a NAPI. This will enable the threaded polling on a NAPI
context.

Add a test in `nl_netdev.py` that verifies various cases of threaded
NAPI being set at NAPI and at device level.

Tested
 ./tools/testing/selftests/net/nl_netdev.py
 TAP version 13
 1..7
 ok 1 nl_netdev.empty_check
 ok 2 nl_netdev.lo_check
 ok 3 nl_netdev.page_pool_check
 ok 4 nl_netdev.napi_list_check
 ok 5 nl_netdev.dev_set_threaded
 ok 6 nl_netdev.napi_set_threaded
 ok 7 nl_netdev.nsim_rxq_reset_down
 # Totals: pass:7 fail:0 xfail:0 xpass:0 skip:0 error:0

Signed-off-by: Samiullah Khawaja <skhawaja@google.com>
---

v9:
 - Fix usage of NAPI lower/upper case.
 - Handle napi_set_threaded error.
 - Move napi_set_threaded kdoc to function definition.
 - In documentation, Use the ynl CLI form that is packaged with distros.
 - Remove backtick usage in netdev.yaml for threaded attribute doc.
 - Use nla_get_uint instead of nla_get_u8.

v8:
 - Fix napi_set_threaded kdoc in net/core/dev.h.

v7:
 - rebase
 - stop kthread when napi threaded is unset at napi level.
 - Update selftest to verify that kthread PID is not set when threaded
   is disabled at napi level.

v6:
 - Set the threaded property at device level even if the currently set
   value is same. This is to override any per napi settings. Update
   selftest to verify this scenario.
 - Use u8 instead of uint in netdev_nl_napi_set_config implementation.
 - Extend the selftest to verify the existing behaviour that the PID
   stays valid once threaded napi is enabled. It stays valid even after
   disabling the threaded napi. Also verify that the same kthread(PID)
   is reused when threaded napi is enabled again. Will keep this
   behaviour as based on the discussion on v5.

v5:
 - This patch was part of:
 https://lore.kernel.org/netdev/Z92e2kCYXQ_RsrJh@LQ3V64L9R2/T/
 It is being sent separately for the first time.
 - Change threaded attribute type to uint
 - Set napi threaded state when set at napi level. When set at device
   level overwrite the state of each napi. This is the `write all`
   symantics that is also followed by other configurations.
 - Add a test to verify `write all` symantics.

 Documentation/netlink/specs/netdev.yaml  | 10 +++
 Documentation/networking/napi.rst        |  9 ++-
 include/linux/netdevice.h                |  1 +
 include/uapi/linux/netdev.h              |  1 +
 net/core/dev.c                           | 37 +++++++++-
 net/core/dev.h                           | 13 ++++
 net/core/netdev-genl-gen.c               |  5 +-
 net/core/netdev-genl.c                   | 14 ++++
 tools/include/uapi/linux/netdev.h        |  1 +
 tools/testing/selftests/net/nl_netdev.py | 91 +++++++++++++++++++++++-
 10 files changed, 175 insertions(+), 7 deletions(-)

diff --git a/Documentation/netlink/specs/netdev.yaml b/Documentation/netlink/specs/netdev.yaml
index ce4cfec82100..85d0ea6ac426 100644
--- a/Documentation/netlink/specs/netdev.yaml
+++ b/Documentation/netlink/specs/netdev.yaml
@@ -283,6 +283,14 @@ attribute-sets:
         doc: The timeout, in nanoseconds, of how long to suspend irq
              processing, if event polling finds events
         type: uint
+      -
+        name: threaded
+        doc: Whether the NAPI is configured to operate in threaded polling
+             mode. If this is set to 1 then the NAPI context operates in
+             threaded polling mode.
+        type: uint
+        checks:
+          max: 1
   -
     name: xsk-info
     attributes: []
@@ -694,6 +702,7 @@ operations:
             - defer-hard-irqs
             - gro-flush-timeout
             - irq-suspend-timeout
+            - threaded
       dump:
         request:
           attributes:
@@ -746,6 +755,7 @@ operations:
             - defer-hard-irqs
             - gro-flush-timeout
             - irq-suspend-timeout
+            - threaded
     -
       name: bind-tx
       doc: Bind dmabuf to netdev for TX
diff --git a/Documentation/networking/napi.rst b/Documentation/networking/napi.rst
index d0e3953cae6a..a15754adb041 100644
--- a/Documentation/networking/napi.rst
+++ b/Documentation/networking/napi.rst
@@ -444,7 +444,14 @@ dependent). The NAPI instance IDs will be assigned in the opposite
 order than the process IDs of the kernel threads.
 
 Threaded NAPI is controlled by writing 0/1 to the ``threaded`` file in
-netdev's sysfs directory.
+netdev's sysfs directory. It can also be enabled for a specific NAPI using
+netlink interface.
+
+For example, using the script:
+
+.. code-block:: bash
+
+  $ ynl --family netdev --do napi-set --json='{"id": 66, "threaded": 1}'
 
 .. rubric:: Footnotes
 
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index adb14db25798..35947d3e5f93 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -369,6 +369,7 @@ struct napi_config {
 	u64 irq_suspend_timeout;
 	u32 defer_hard_irqs;
 	cpumask_t affinity_mask;
+	bool threaded;
 	unsigned int napi_id;
 };
 
diff --git a/include/uapi/linux/netdev.h b/include/uapi/linux/netdev.h
index 7eb9571786b8..1f3719a9a0eb 100644
--- a/include/uapi/linux/netdev.h
+++ b/include/uapi/linux/netdev.h
@@ -134,6 +134,7 @@ enum {
 	NETDEV_A_NAPI_DEFER_HARD_IRQS,
 	NETDEV_A_NAPI_GRO_FLUSH_TIMEOUT,
 	NETDEV_A_NAPI_IRQ_SUSPEND_TIMEOUT,
+	NETDEV_A_NAPI_THREADED,
 
 	__NETDEV_A_NAPI_MAX,
 	NETDEV_A_NAPI_MAX = (__NETDEV_A_NAPI_MAX - 1)
diff --git a/net/core/dev.c b/net/core/dev.c
index 5baa4691074f..1f02663f5f9a 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -6963,6 +6963,38 @@ static void napi_stop_kthread(struct napi_struct *napi)
 	napi->thread = NULL;
 }
 
+/**
+ * napi_set_threaded - set NAPI threaded state
+ * @n: napi_struct to set the threaded state on
+ * @threaded: whether this NAPI does threaded polling
+ *
+ * Return: 0 on success and negative errno on failure.
+ */
+int napi_set_threaded(struct napi_struct *napi, bool threaded)
+{
+	if (threaded) {
+		if (!napi->thread) {
+			int err = napi_kthread_create(napi);
+
+			if (err)
+				return err;
+		}
+	}
+
+	if (napi->config)
+		napi->config->threaded = threaded;
+
+	if (!threaded && napi->thread) {
+		napi_stop_kthread(napi);
+	} else {
+		/* Make sure kthread is created before THREADED bit is set. */
+		smp_mb__before_atomic();
+		assign_bit(NAPI_STATE_THREADED, &napi->state, threaded);
+	}
+
+	return 0;
+}
+
 int dev_set_threaded(struct net_device *dev, bool threaded)
 {
 	struct napi_struct *napi;
@@ -6970,9 +7002,6 @@ int dev_set_threaded(struct net_device *dev, bool threaded)
 
 	netdev_assert_locked_or_invisible(dev);
 
-	if (dev->threaded == threaded)
-		return 0;
-
 	if (threaded) {
 		list_for_each_entry(napi, &dev->napi_list, dev_list) {
 			if (!napi->thread) {
@@ -7223,6 +7252,8 @@ static void napi_restore_config(struct napi_struct *n)
 		napi_hash_add(n);
 		n->config->napi_id = n->napi_id;
 	}
+
+	WARN_ON_ONCE(napi_set_threaded(n, n->config->threaded));
 }
 
 static void napi_save_config(struct napi_struct *n)
diff --git a/net/core/dev.h b/net/core/dev.h
index e93f36b7ddf3..7733f48c3259 100644
--- a/net/core/dev.h
+++ b/net/core/dev.h
@@ -315,6 +315,19 @@ static inline void napi_set_irq_suspend_timeout(struct napi_struct *n,
 	WRITE_ONCE(n->irq_suspend_timeout, timeout);
 }
 
+/**
+ * napi_get_threaded - get the NAPI threaded state
+ * @n: napi_struct to get the threaded state from
+ *
+ * Return: the per-NAPI threaded state.
+ */
+static inline bool napi_get_threaded(struct napi_struct *n)
+{
+	return test_bit(NAPI_STATE_THREADED, &n->state);
+}
+
+int napi_set_threaded(struct napi_struct *n, bool threaded);
+
 int rps_cpumask_housekeeping(struct cpumask *mask);
 
 #if defined(CONFIG_DEBUG_NET) && defined(CONFIG_BPF_SYSCALL)
diff --git a/net/core/netdev-genl-gen.c b/net/core/netdev-genl-gen.c
index 4fc44587f493..0994bd68a7e6 100644
--- a/net/core/netdev-genl-gen.c
+++ b/net/core/netdev-genl-gen.c
@@ -92,11 +92,12 @@ static const struct nla_policy netdev_bind_rx_nl_policy[NETDEV_A_DMABUF_FD + 1]
 };
 
 /* NETDEV_CMD_NAPI_SET - do */
-static const struct nla_policy netdev_napi_set_nl_policy[NETDEV_A_NAPI_IRQ_SUSPEND_TIMEOUT + 1] = {
+static const struct nla_policy netdev_napi_set_nl_policy[NETDEV_A_NAPI_THREADED + 1] = {
 	[NETDEV_A_NAPI_ID] = { .type = NLA_U32, },
 	[NETDEV_A_NAPI_DEFER_HARD_IRQS] = NLA_POLICY_FULL_RANGE(NLA_U32, &netdev_a_napi_defer_hard_irqs_range),
 	[NETDEV_A_NAPI_GRO_FLUSH_TIMEOUT] = { .type = NLA_UINT, },
 	[NETDEV_A_NAPI_IRQ_SUSPEND_TIMEOUT] = { .type = NLA_UINT, },
+	[NETDEV_A_NAPI_THREADED] = NLA_POLICY_MAX(NLA_UINT, 1),
 };
 
 /* NETDEV_CMD_BIND_TX - do */
@@ -193,7 +194,7 @@ static const struct genl_split_ops netdev_nl_ops[] = {
 		.cmd		= NETDEV_CMD_NAPI_SET,
 		.doit		= netdev_nl_napi_set_doit,
 		.policy		= netdev_napi_set_nl_policy,
-		.maxattr	= NETDEV_A_NAPI_IRQ_SUSPEND_TIMEOUT,
+		.maxattr	= NETDEV_A_NAPI_THREADED,
 		.flags		= GENL_ADMIN_PERM | GENL_CMD_CAP_DO,
 	},
 	{
diff --git a/net/core/netdev-genl.c b/net/core/netdev-genl.c
index 2afa7b2141aa..5875df372415 100644
--- a/net/core/netdev-genl.c
+++ b/net/core/netdev-genl.c
@@ -184,6 +184,10 @@ netdev_nl_napi_fill_one(struct sk_buff *rsp, struct napi_struct *napi,
 	if (napi->irq >= 0 && nla_put_u32(rsp, NETDEV_A_NAPI_IRQ, napi->irq))
 		goto nla_put_failure;
 
+	if (nla_put_uint(rsp, NETDEV_A_NAPI_THREADED,
+			 napi_get_threaded(napi)))
+		goto nla_put_failure;
+
 	if (napi->thread) {
 		pid = task_pid_nr(napi->thread);
 		if (nla_put_u32(rsp, NETDEV_A_NAPI_PID, pid))
@@ -322,8 +326,18 @@ netdev_nl_napi_set_config(struct napi_struct *napi, struct genl_info *info)
 {
 	u64 irq_suspend_timeout = 0;
 	u64 gro_flush_timeout = 0;
+	u8 threaded = 0;
 	u32 defer = 0;
 
+	if (info->attrs[NETDEV_A_NAPI_THREADED]) {
+		int ret;
+
+		threaded = nla_get_uint(info->attrs[NETDEV_A_NAPI_THREADED]);
+		ret = napi_set_threaded(napi, !!threaded);
+		if (ret)
+			return ret;
+	}
+
 	if (info->attrs[NETDEV_A_NAPI_DEFER_HARD_IRQS]) {
 		defer = nla_get_u32(info->attrs[NETDEV_A_NAPI_DEFER_HARD_IRQS]);
 		napi_set_defer_hard_irqs(napi, defer);
diff --git a/tools/include/uapi/linux/netdev.h b/tools/include/uapi/linux/netdev.h
index 7eb9571786b8..1f3719a9a0eb 100644
--- a/tools/include/uapi/linux/netdev.h
+++ b/tools/include/uapi/linux/netdev.h
@@ -134,6 +134,7 @@ enum {
 	NETDEV_A_NAPI_DEFER_HARD_IRQS,
 	NETDEV_A_NAPI_GRO_FLUSH_TIMEOUT,
 	NETDEV_A_NAPI_IRQ_SUSPEND_TIMEOUT,
+	NETDEV_A_NAPI_THREADED,
 
 	__NETDEV_A_NAPI_MAX,
 	NETDEV_A_NAPI_MAX = (__NETDEV_A_NAPI_MAX - 1)
diff --git a/tools/testing/selftests/net/nl_netdev.py b/tools/testing/selftests/net/nl_netdev.py
index c9109627a741..c8ffade79a52 100755
--- a/tools/testing/selftests/net/nl_netdev.py
+++ b/tools/testing/selftests/net/nl_netdev.py
@@ -35,6 +35,91 @@ def napi_list_check(nf) -> None:
                 ksft_eq(len(napis), 100,
                         comment=f"queue count after reset queue {q} mode {i}")
 
+def napi_set_threaded(nf) -> None:
+    """
+    Test that verifies various cases of napi threaded
+    set and unset at napi and device level.
+    """
+    with NetdevSimDev(queue_count=2) as nsimdev:
+        nsim = nsimdev.nsims[0]
+
+        ip(f"link set dev {nsim.ifname} up")
+
+        napis = nf.napi_get({'ifindex': nsim.ifindex}, dump=True)
+        ksft_eq(len(napis), 2)
+
+        napi0_id = napis[0]['id']
+        napi1_id = napis[1]['id']
+
+        # set napi threaded and verify
+        nf.napi_set({'id': napi0_id, 'threaded': 1})
+        napi0 = nf.napi_get({'id': napi0_id})
+        ksft_eq(napi0['threaded'], 1)
+        ksft_ne(napi0.get('pid'), None)
+
+        # check it is not set for napi1
+        napi1 = nf.napi_get({'id': napi1_id})
+        ksft_eq(napi1['threaded'], 0)
+        ksft_eq(napi1.get('pid'), None)
+
+        ip(f"link set dev {nsim.ifname} down")
+        ip(f"link set dev {nsim.ifname} up")
+
+        # verify if napi threaded is still set
+        napi0 = nf.napi_get({'id': napi0_id})
+        ksft_eq(napi0['threaded'], 1)
+        ksft_ne(napi0.get('pid'), None)
+
+        # check it is still not set for napi1
+        napi1 = nf.napi_get({'id': napi1_id})
+        ksft_eq(napi1['threaded'], 0)
+        ksft_eq(napi1.get('pid'), None)
+
+        # unset napi threaded and verify
+        nf.napi_set({'id': napi0_id, 'threaded': 0})
+        napi0 = nf.napi_get({'id': napi0_id})
+        ksft_eq(napi0['threaded'], 0)
+        ksft_eq(napi0.get('pid'), None)
+
+        # set threaded at device level
+        system(f"echo 1 > /sys/class/net/{nsim.ifname}/threaded")
+
+        # check napi threaded is set for both napis
+        napi0 = nf.napi_get({'id': napi0_id})
+        ksft_eq(napi0['threaded'], 1)
+        ksft_ne(napi0.get('pid'), None)
+        napi1 = nf.napi_get({'id': napi1_id})
+        ksft_eq(napi1['threaded'], 1)
+        ksft_ne(napi1.get('pid'), None)
+
+        # unset threaded at device level
+        system(f"echo 0 > /sys/class/net/{nsim.ifname}/threaded")
+
+        # check napi threaded is unset for both napis
+        napi0 = nf.napi_get({'id': napi0_id})
+        ksft_eq(napi0['threaded'], 0)
+        ksft_eq(napi0.get('pid'), None)
+        napi1 = nf.napi_get({'id': napi1_id})
+        ksft_eq(napi1['threaded'], 0)
+        ksft_eq(napi1.get('pid'), None)
+
+        # set napi threaded for napi0
+        nf.napi_set({'id': napi0_id, 'threaded': 1})
+        napi0 = nf.napi_get({'id': napi0_id})
+        ksft_eq(napi0['threaded'], 1)
+        ksft_ne(napi0.get('pid'), None)
+
+        # unset threaded at device level
+        system(f"echo 0 > /sys/class/net/{nsim.ifname}/threaded")
+
+        # check napi threaded is unset for both napis
+        napi0 = nf.napi_get({'id': napi0_id})
+        ksft_eq(napi0['threaded'], 0)
+        ksft_eq(napi0.get('pid'), None)
+        napi1 = nf.napi_get({'id': napi1_id})
+        ksft_eq(napi1['threaded'], 0)
+        ksft_eq(napi1.get('pid'), None)
+
 def dev_set_threaded(nf) -> None:
     """
     Test that verifies various cases of napi threaded
@@ -56,8 +141,10 @@ def dev_set_threaded(nf) -> None:
 
         # check napi threaded is set for both napis
         napi0 = nf.napi_get({'id': napi0_id})
+        ksft_eq(napi0['threaded'], 1)
         ksft_ne(napi0.get('pid'), None)
         napi1 = nf.napi_get({'id': napi1_id})
+        ksft_eq(napi1['threaded'], 1)
         ksft_ne(napi1.get('pid'), None)
 
         # unset threaded
@@ -65,8 +152,10 @@ def dev_set_threaded(nf) -> None:
 
         # check napi threaded is unset for both napis
         napi0 = nf.napi_get({'id': napi0_id})
+        ksft_eq(napi0['threaded'], 0)
         ksft_eq(napi0.get('pid'), None)
         napi1 = nf.napi_get({'id': napi1_id})
+        ksft_eq(napi1['threaded'], 0)
         ksft_eq(napi1.get('pid'), None)
 
 def nsim_rxq_reset_down(nf) -> None:
@@ -156,7 +245,7 @@ def page_pool_check(nf) -> None:
 def main() -> None:
     nf = NetdevFamily()
     ksft_run([empty_check, lo_check, page_pool_check, napi_list_check,
-              dev_set_threaded, nsim_rxq_reset_down],
+              dev_set_threaded, napi_set_threaded, nsim_rxq_reset_down],
              args=(nf, ))
     ksft_exit()
 

base-commit: b1b36680107ede3a4ec7fa41d052971606d6b325
-- 
2.50.0.rc2.761.g2dc52ea45b-goog

Re: [PATCH net-next v9] Add support to set NAPI threaded for individual NAPI

[Jakub Kicinski]

[swapping Joe's email for a working one]

On Mon, 23 Jun 2025 17:53:16 +0000 Samiullah Khawaja wrote:
> A net device has a threaded sysctl that can be used to enable threaded
> NAPI polling on all of the NAPI contexts under that device. Allow
> enabling threaded NAPI polling at individual NAPI level using netlink.
> 
> Extend the netlink operation `napi-set` and allow setting the threaded
> attribute of a NAPI. This will enable the threaded polling on a NAPI
> context.
> 
> Add a test in `nl_netdev.py` that verifies various cases of threaded
> NAPI being set at NAPI and at device level.

LGTM, but as we discussed many version ago my subjective preference
would be for the per-queue setting to be tri-state (unset=follow the
device setting, enabled, disabled). Rather than have the device level
act as an override when written. It sounded like Willem and Joe thought
that's too complicated and diverges from the existing behavior.
Which is fine by me. But I'd like to see some review tags from them
now :)

Gentlemen?

Re: [PATCH net-next v9] Add support to set NAPI threaded for individual NAPI

[Jakub Kicinski]

On Mon, 23 Jun 2025 17:53:16 +0000 Samiullah Khawaja wrote:
> +/**
> + * napi_set_threaded - set NAPI threaded state
> + * @n: napi_struct to set the threaded state on
> + * @threaded: whether this NAPI does threaded polling
> + *
> + * Return: 0 on success and negative errno on failure.
> + */
> +int napi_set_threaded(struct napi_struct *napi, bool threaded)

Ah, missed that this kdoc is wrong s/n/napi/

For the record:
https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=e5880f95a97928308845dc97fdd239605e06e501
-- 
pw-bot: cr

Re: [PATCH net-next v9] Add support to set NAPI threaded for individual NAPI

[Willem de Bruijn]

Jakub Kicinski wrote:
> [swapping Joe's email for a working one]
> 
> On Mon, 23 Jun 2025 17:53:16 +0000 Samiullah Khawaja wrote:
> > A net device has a threaded sysctl that can be used to enable threaded
> > NAPI polling on all of the NAPI contexts under that device. Allow
> > enabling threaded NAPI polling at individual NAPI level using netlink.
> > 
> > Extend the netlink operation `napi-set` and allow setting the threaded
> > attribute of a NAPI. This will enable the threaded polling on a NAPI
> > context.
> > 
> > Add a test in `nl_netdev.py` that verifies various cases of threaded
> > NAPI being set at NAPI and at device level.
> 
> LGTM, but as we discussed many version ago my subjective preference
> would be for the per-queue setting to be tri-state (unset=follow the
> device setting, enabled, disabled). Rather than have the device level
> act as an override when written. It sounded like Willem and Joe thought
> that's too complicated and diverges from the existing behavior.
> Which is fine by me. But I'd like to see some review tags from them
> now :)
> 
> Gentlemen?

Missed that earlier, sorry.

Reviewed-by: Willem de Bruijn <willemb@google.com>