@@ -343,7 +343,7 @@ Date: Jan 2021
KernelVersion: 5.12
Contact: netdev@vger.kernel.org
Description:
- Boolean value to control the threaded mode per device. User could
+ Integer value to control the threaded mode per device. User could
set this value to enable/disable threaded mode for all napi
belonging to this device, without the need to do device up/down.
@@ -351,4 +351,5 @@ Description:
== ==================================
0 threaded mode disabled for this dev
1 threaded mode enabled for this dev
+ 2 threaded mode enabled, and busy polling enabled.
== ==================================
@@ -78,6 +78,10 @@ definitions:
name: qstats-scope
type: flags
entries: [ queue ]
+ -
+ name: napi-threaded
+ type: enum
+ entries: [ disable, enable, busy-poll-enable ]
attribute-sets:
-
@@ -271,11 +275,11 @@ attribute-sets:
-
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.
+ mode. If this is set to `enable` then the NAPI context operates
+ in threaded polling mode. If this is set to `busy-poll-enable`
+ then the NAPI kthread also does busypolling.
type: u32
- checks:
- max: 1
+ enum: napi-threaded
-
name: queue
attributes:
@@ -232,7 +232,9 @@ are not well known).
Busy polling is enabled by either setting ``SO_BUSY_POLL`` on
selected sockets or using the global ``net.core.busy_poll`` and
``net.core.busy_read`` sysctls. An io_uring API for NAPI busy polling
-also exists.
+also exists. Threaded polling of NAPI also has a mode to busy poll for
+packets (:ref:`threaded busy polling<threaded_busy_poll>`) using the same
+thread that is used for NAPI processing.
epoll-based busy polling
------------------------
@@ -395,6 +397,69 @@ Therefore, setting ``gro_flush_timeout`` and ``napi_defer_hard_irqs`` is
the recommended usage, because otherwise setting ``irq-suspend-timeout``
might not have any discernible effect.
+.. _threaded_busy_poll:
+
+Threaded NAPI busy polling
+--------------------------
+
+Threaded napi allows processing of packets from each NAPI in a kthread in
+kernel. Threaded napi busy polling extends this and adds support to do
+continuous busy polling of this napi. This can be used to enable busy polling
+independent of userspace application or the API (epoll, io_uring, raw sockets)
+being used in userspace to process the packets.
+
+It can be enabled for each NAPI using netlink interface or at device level using
+the threaded NAPI sysctl.
+
+For example, using following script:
+
+.. code-block:: bash
+
+ $ kernel-source/tools/net/ynl/pyynl/cli.py \
+ --spec Documentation/netlink/specs/netdev.yaml \
+ --do napi-set \
+ --json='{"id": 66,
+ "threaded": "busy-poll-enable"}'
+
+
+Enabling it for each NAPI allows finer control to enable busy pollling for
+only a set of NIC queues which will get traffic with low latency requirements.
+
+Depending on application requirement, user might want to set affinity of the
+kthread that is busy polling each NAPI. User might also want to set priority
+and the scheduler of the thread depending on the latency requirements.
+
+For a hard low-latency application, user might want to dedicate the full core
+for the NAPI polling so the NIC queue descriptors are picked up from the queue
+as soon as they appear. For more relaxed low-latency requirement, user might
+want to share the core with other threads.
+
+Once threaded busy polling is enabled for a NAPI, PID of the kthread can be
+fetched using netlink interface so the affinity, priority and scheduler
+configuration can be done.
+
+For example, following script can be used to fetch the pid:
+
+.. code-block:: bash
+
+ $ kernel-source/tools/net/ynl/pyynl/cli.py \
+ --spec Documentation/netlink/specs/netdev.yaml \
+ --do napi-get \
+ --json='{"id": 66}'
+
+This will output something like following, the pid `258` is the PID of the
+kthread that is polling this NAPI.
+
+.. code-block:: bash
+
+ $ {'defer-hard-irqs': 0,
+ 'gro-flush-timeout': 0,
+ 'id': 66,
+ 'ifindex': 2,
+ 'irq-suspend-timeout': 0,
+ 'pid': 258,
+ 'threaded': 'enable'}
+
.. _threaded:
Threaded NAPI
@@ -2688,7 +2688,7 @@ static int atl1c_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
adapter->mii.mdio_write = atl1c_mdio_write;
adapter->mii.phy_id_mask = 0x1f;
adapter->mii.reg_num_mask = MDIO_CTRL_REG_MASK;
- dev_set_threaded(netdev, true);
+ dev_set_threaded(netdev, NETDEV_NAPI_THREADED_ENABLE);
for (i = 0; i < adapter->rx_queue_count; ++i)
netif_napi_add(netdev, &adapter->rrd_ring[i].napi,
atl1c_clean_rx);
@@ -156,7 +156,7 @@ static int mlxsw_pci_napi_devs_init(struct mlxsw_pci *mlxsw_pci)
}
strscpy(mlxsw_pci->napi_dev_rx->name, "mlxsw_rx",
sizeof(mlxsw_pci->napi_dev_rx->name));
- dev_set_threaded(mlxsw_pci->napi_dev_rx, true);
+ dev_set_threaded(mlxsw_pci->napi_dev_rx, NETDEV_NAPI_THREADED_ENABLE);
return 0;
@@ -3075,7 +3075,7 @@ static int ravb_probe(struct platform_device *pdev)
if (info->coalesce_irqs) {
netdev_sw_irq_coalesce_default_on(ndev);
if (num_present_cpus() == 1)
- dev_set_threaded(ndev, true);
+ dev_set_threaded(ndev, NETDEV_NAPI_THREADED_ENABLE);
}
/* Network device register */
@@ -935,7 +935,7 @@ static int ath10k_snoc_hif_start(struct ath10k *ar)
bitmap_clear(ar_snoc->pending_ce_irqs, 0, CE_COUNT_MAX);
- dev_set_threaded(ar->napi_dev, true);
+ dev_set_threaded(ar->napi_dev, NETDEV_NAPI_THREADED_ENABLE);
ath10k_core_napi_enable(ar);
ath10k_snoc_irq_enable(ar);
ath10k_snoc_rx_post(ar);
@@ -352,7 +352,7 @@ struct napi_config {
u64 gro_flush_timeout;
u64 irq_suspend_timeout;
u32 defer_hard_irqs;
- bool threaded;
+ u8 threaded;
unsigned int napi_id;
};
@@ -410,6 +410,8 @@ enum {
NAPI_STATE_PREFER_BUSY_POLL, /* prefer busy-polling over softirq processing*/
NAPI_STATE_THREADED, /* The poll is performed inside its own thread*/
NAPI_STATE_SCHED_THREADED, /* Napi is currently scheduled in threaded mode */
+ NAPI_STATE_THREADED_BUSY_POLL, /* The threaded napi poller will busy poll */
+ NAPI_STATE_SCHED_THREADED_BUSY_POLL, /* The threaded napi poller is busy polling */
};
enum {
@@ -423,8 +425,14 @@ enum {
NAPIF_STATE_PREFER_BUSY_POLL = BIT(NAPI_STATE_PREFER_BUSY_POLL),
NAPIF_STATE_THREADED = BIT(NAPI_STATE_THREADED),
NAPIF_STATE_SCHED_THREADED = BIT(NAPI_STATE_SCHED_THREADED),
+ NAPIF_STATE_THREADED_BUSY_POLL = BIT(NAPI_STATE_THREADED_BUSY_POLL),
+ NAPIF_STATE_SCHED_THREADED_BUSY_POLL
+ = BIT(NAPI_STATE_SCHED_THREADED_BUSY_POLL),
};
+#define NAPIF_STATE_THREADED_BUSY_POLL_MASK \
+ (NAPIF_STATE_THREADED | NAPIF_STATE_THREADED_BUSY_POLL)
+
enum gro_result {
GRO_MERGED,
GRO_MERGED_FREE,
@@ -571,16 +579,18 @@ static inline bool napi_complete(struct napi_struct *n)
return napi_complete_done(n, 0);
}
-int dev_set_threaded(struct net_device *dev, bool threaded);
+int dev_set_threaded(struct net_device *dev,
+ enum netdev_napi_threaded threaded);
/*
* napi_set_threaded - set napi threaded state
* @napi: NAPI context
- * @threaded: whether this napi does threaded polling
+ * @threaded: threading mode
*
* Return 0 on success and negative errno on failure.
*/
-int napi_set_threaded(struct napi_struct *napi, bool threaded);
+int napi_set_threaded(struct napi_struct *napi,
+ enum netdev_napi_threaded threaded);
void napi_disable(struct napi_struct *n);
void napi_disable_locked(struct napi_struct *n);
@@ -2404,7 +2414,7 @@ struct net_device {
struct sfp_bus *sfp_bus;
struct lock_class_key *qdisc_tx_busylock;
bool proto_down;
- bool threaded;
+ u8 threaded;
/* priv_flags_slow, ungrouped to save space */
unsigned long see_all_hwtstamp_requests:1;
@@ -74,6 +74,12 @@ enum netdev_qstats_scope {
NETDEV_QSTATS_SCOPE_QUEUE = 1,
};
+enum netdev_napi_threaded {
+ NETDEV_NAPI_THREADED_DISABLE,
+ NETDEV_NAPI_THREADED_ENABLE,
+ NETDEV_NAPI_THREADED_BUSY_POLL_ENABLE,
+};
+
enum {
NETDEV_A_DEV_IFINDEX = 1,
NETDEV_A_DEV_PAD,
@@ -78,6 +78,7 @@
#include <linux/slab.h>
#include <linux/sched.h>
#include <linux/sched/isolation.h>
+#include <linux/sched/types.h>
#include <linux/sched/mm.h>
#include <linux/smpboot.h>
#include <linux/mutex.h>
@@ -6403,7 +6404,8 @@ bool napi_complete_done(struct napi_struct *n, int work_done)
* the guarantee we will be called later.
*/
if (unlikely(n->state & (NAPIF_STATE_NPSVC |
- NAPIF_STATE_IN_BUSY_POLL)))
+ NAPIF_STATE_IN_BUSY_POLL |
+ NAPIF_STATE_SCHED_THREADED_BUSY_POLL)))
return false;
if (work_done) {
@@ -6792,8 +6794,10 @@ static void init_gro_hash(struct napi_struct *napi)
napi->gro_bitmask = 0;
}
-int napi_set_threaded(struct napi_struct *napi, bool threaded)
+int napi_set_threaded(struct napi_struct *napi,
+ enum netdev_napi_threaded threaded)
{
+ unsigned long val;
if (napi->dev->threaded)
return -EINVAL;
@@ -6811,14 +6815,20 @@ int napi_set_threaded(struct napi_struct *napi, bool threaded)
/* Make sure kthread is created before THREADED bit is set. */
smp_mb__before_atomic();
- assign_bit(NAPI_STATE_THREADED, &napi->state, threaded);
+ val = 0;
+ if (threaded == NETDEV_NAPI_THREADED_BUSY_POLL_ENABLE)
+ val |= NAPIF_STATE_THREADED_BUSY_POLL;
+ if (threaded)
+ val |= NAPIF_STATE_THREADED;
+ set_mask_bits(&napi->state, NAPIF_STATE_THREADED_BUSY_POLL_MASK, val);
return 0;
}
-int dev_set_threaded(struct net_device *dev, bool threaded)
+int dev_set_threaded(struct net_device *dev, enum netdev_napi_threaded threaded)
{
struct napi_struct *napi;
+ unsigned long val;
int err = 0;
netdev_assert_locked_or_invisible(dev);
@@ -6826,17 +6836,22 @@ int dev_set_threaded(struct net_device *dev, bool threaded)
if (dev->threaded == threaded)
return 0;
+ val = 0;
if (threaded) {
/* Check if threaded is set at napi level already */
list_for_each_entry(napi, &dev->napi_list, dev_list)
if (test_bit(NAPI_STATE_THREADED, &napi->state))
return -EINVAL;
+ val |= NAPIF_STATE_THREADED;
+ if (threaded == NETDEV_NAPI_THREADED_BUSY_POLL_ENABLE)
+ val |= NAPIF_STATE_THREADED_BUSY_POLL;
+
list_for_each_entry(napi, &dev->napi_list, dev_list) {
if (!napi->thread) {
err = napi_kthread_create(napi);
if (err) {
- threaded = false;
+ threaded = NETDEV_NAPI_THREADED_DISABLE;
break;
}
}
@@ -6855,9 +6870,13 @@ int dev_set_threaded(struct net_device *dev, bool threaded)
* polled. In this case, the switch between threaded mode and
* softirq mode will happen in the next round of napi_schedule().
* This should not cause hiccups/stalls to the live traffic.
+ *
+ * Switch to busy_poll threaded napi will occur after the threaded
+ * napi is scheduled.
*/
list_for_each_entry(napi, &dev->napi_list, dev_list)
- assign_bit(NAPI_STATE_THREADED, &napi->state, threaded);
+ set_mask_bits(&napi->state,
+ NAPIF_STATE_THREADED_BUSY_POLL_MASK, val);
return err;
}
@@ -7235,7 +7254,7 @@ static int napi_thread_wait(struct napi_struct *napi)
return -1;
}
-static void napi_threaded_poll_loop(struct napi_struct *napi)
+static void napi_threaded_poll_loop(struct napi_struct *napi, bool busy_poll)
{
struct bpf_net_context __bpf_net_ctx, *bpf_net_ctx;
struct softnet_data *sd;
@@ -7264,22 +7283,53 @@ static void napi_threaded_poll_loop(struct napi_struct *napi)
}
skb_defer_free_flush(sd);
bpf_net_ctx_clear(bpf_net_ctx);
+
+ /* Push the skbs up the stack if busy polling. */
+ if (busy_poll)
+ __napi_gro_flush_helper(napi);
local_bh_enable();
- if (!repoll)
+ /* If busy polling then do not break here because we need to
+ * call cond_resched and rcu_softirq_qs_periodic to prevent
+ * watchdog warnings.
+ */
+ if (!repoll && !busy_poll)
break;
rcu_softirq_qs_periodic(last_qs);
cond_resched();
+
+ if (!repoll)
+ break;
}
}
static int napi_threaded_poll(void *data)
{
struct napi_struct *napi = data;
+ bool busy_poll_sched;
+ unsigned long val;
+ bool busy_poll;
+
+ while (!napi_thread_wait(napi)) {
+ /* Once woken up, this means that we are scheduled as threaded
+ * napi and this thread owns the napi context, if busy poll
+ * state is set then we busy poll this napi.
+ */
+ val = READ_ONCE(napi->state);
+ busy_poll = val & NAPIF_STATE_THREADED_BUSY_POLL;
+ busy_poll_sched = val & NAPIF_STATE_SCHED_THREADED_BUSY_POLL;
+
+ /* Do not busy poll if napi is disabled. */
+ if (unlikely(val & NAPIF_STATE_DISABLE))
+ busy_poll = false;
+
+ if (busy_poll != busy_poll_sched)
+ assign_bit(NAPI_STATE_SCHED_THREADED_BUSY_POLL,
+ &napi->state, busy_poll);
- while (!napi_thread_wait(napi))
- napi_threaded_poll_loop(napi);
+ napi_threaded_poll_loop(napi, busy_poll);
+ }
return 0;
}
@@ -12474,7 +12524,7 @@ static void run_backlog_napi(unsigned int cpu)
{
struct softnet_data *sd = per_cpu_ptr(&softnet_data, cpu);
- napi_threaded_poll_loop(&sd->backlog);
+ napi_threaded_poll_loop(&sd->backlog, false);
}
static void backlog_napi_setup(unsigned int cpu)
@@ -657,7 +657,7 @@ static int modify_napi_threaded(struct net_device *dev, unsigned long val)
if (list_empty(&dev->napi_list))
return -EOPNOTSUPP;
- if (val != 0 && val != 1)
+ if (val > NETDEV_NAPI_THREADED_BUSY_POLL_ENABLE)
return -EOPNOTSUPP;
ret = dev_set_threaded(dev, val);
@@ -97,7 +97,7 @@ static const struct nla_policy netdev_napi_set_nl_policy[NETDEV_A_NAPI_THREADED
[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_U32, 1),
+ [NETDEV_A_NAPI_THREADED] = NLA_POLICY_MAX(NLA_U32, 2),
};
/* Ops table for netdev */
@@ -329,7 +329,7 @@ netdev_nl_napi_set_config(struct napi_struct *napi, struct genl_info *info)
if (info->attrs[NETDEV_A_NAPI_THREADED]) {
threaded = nla_get_u32(info->attrs[NETDEV_A_NAPI_THREADED]);
- napi_set_threaded(napi, !!threaded);
+ napi_set_threaded(napi, threaded);
}
if (info->attrs[NETDEV_A_NAPI_DEFER_HARD_IRQS]) {
@@ -74,6 +74,12 @@ enum netdev_qstats_scope {
NETDEV_QSTATS_SCOPE_QUEUE = 1,
};
+enum netdev_napi_threaded {
+ NETDEV_NAPI_THREADED_DISABLE,
+ NETDEV_NAPI_THREADED_ENABLE,
+ NETDEV_NAPI_THREADED_BUSY_POLL_ENABLE,
+};
+
enum {
NETDEV_A_DEV_IFINDEX = 1,
NETDEV_A_DEV_PAD,