* [PATCH iwl-next v5 1/2] i40e: move ATR sample rate from ring to PF level
2026-10-08 13:18 [PATCH iwl-next v5 0/2] i40e: add runtime-tunable ATR sample rate mheib
@ 2026-10-08 13:18 ` mheib
2026-10-08 13:18 ` [PATCH iwl-next v5 2/2] i40e: add devlink parameter for Flow Director ATR sample rate mheib
1 sibling, 0 replies; 3+ messages in thread
From: mheib @ 2026-10-08 13:18 UTC (permalink / raw)
To: intel-wired-lan
Cc: netdev, jiri, davem, edumazet, kuba, pabeni, anthony.l.nguyen,
przemyslaw.kitszel, andrew+netdev, horms, Mohammad Heib
From: Mohammad Heib <mheib@redhat.com>
The ATR sample rate is currently stored per-ring and initialized when each
TX ring is configured. Since the sample rate is a global policy that
applies uniformly across all rings, it makes more sense to store it at
the PF level.
Move atr_sample_rate from struct i40e_ring to struct i40e_pf and initialize
it once during i40e_sw_init(). Update i40e_atr() to reference the PF-level
field. Change atr_count from u8 to u32 to match the sample rate type.
This also changes a user-visible behavior: previously, re-enabling ATR via
"ethtool --set-priv-flags ethX flow-director-atr on" did not resume
sampling until an ifdown/ifup or PF reset, because the per-ring sample
rate was latched at ring configuration time. With the PF-level rate always
initialized, the flag takes effect immediately.
Trigger a PF reset when re-enabling ATR so that fd_ena is re-evaluated
in the TX queue HW context.
Signed-off-by: Mohammad Heib <mheib@redhat.com>
---
drivers/net/ethernet/intel/i40e/i40e.h | 1 +
drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 12 +++++++++++-
drivers/net/ethernet/intel/i40e/i40e_main.c | 9 +++------
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 6 ++++--
drivers/net/ethernet/intel/i40e/i40e_txrx.h | 3 +--
5 files changed, 20 insertions(+), 11 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index a9d79bbd3ea7..528ef329a647 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -486,6 +486,7 @@ struct i40e_pf {
u16 rss_size_max; /* HW defined max RSS queues */
u16 fdir_pf_filter_count; /* num of guaranteed filters for this PF */
u16 num_alloc_vsi; /* num VSIs this driver supports */
+ u32 atr_sample_rate;
bool wol_en;
struct hlist_head fdir_filter_list;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
index 0ae50d7f3465..9b1ecb2c4edc 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
@@ -5300,6 +5300,16 @@ static int i40e_set_priv_flags(struct net_device *dev, u32 flags)
test_bit(I40E_FLAG_SOURCE_PRUNING_DIS, changed_flags))
reset_needed = BIT(__I40E_PF_RESET_REQUESTED);
+ /* Re-enabling ATR requires a reset to reprogram fd_ena in the
+ * TX queue HW context. Skip when FD_SB (ntuple) is already
+ * enabled, since i40e_configure_tx_ring() sets fd_ena when
+ * either FD flag is active.
+ */
+ if (test_bit(I40E_FLAG_FD_ATR_ENA, changed_flags) &&
+ test_bit(I40E_FLAG_FD_ATR_ENA, new_flags) &&
+ !test_bit(I40E_FLAG_FD_SB_ENA, new_flags))
+ reset_needed = BIT(__I40E_PF_RESET_REQUESTED);
+
/* Before we finalize any flag changes, we need to perform some
* checks to ensure that the changes are supported and safe.
*/
@@ -5431,7 +5441,7 @@ static int i40e_set_priv_flags(struct net_device *dev, u32 flags)
case LIBIE_AQ_RC_EEXIST:
dev_warn(&pf->pdev->dev,
"FW LLDP agent is already running\n");
- reset_needed = 0;
+ reset_needed &= ~I40E_PF_RESET_AND_REBUILD_FLAG;
break;
case LIBIE_AQ_RC_EPERM:
dev_warn(&pf->pdev->dev,
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 109ba6b20f8e..d83116596eda 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -3459,12 +3459,7 @@ static int i40e_configure_tx_ring(struct i40e_ring *ring)
ring->xsk_pool = i40e_xsk_pool(ring);
/* some ATR related tx ring init */
- if (test_bit(I40E_FLAG_FD_ATR_ENA, vsi->back->flags)) {
- ring->atr_sample_rate = I40E_DEFAULT_ATR_SAMPLE_RATE;
- ring->atr_count = 0;
- } else {
- ring->atr_sample_rate = 0;
- }
+ ring->atr_count = 0;
/* configure XPS */
i40e_config_xps_tx_ring(ring);
@@ -12754,6 +12749,8 @@ static int i40e_sw_init(struct i40e_pf *pf)
}
}
+ pf->atr_sample_rate = I40E_DEFAULT_ATR_SAMPLE_RATE;
+
if ((pf->hw.func_caps.fd_filters_guaranteed > 0) ||
(pf->hw.func_caps.fd_filters_best_effort > 0)) {
set_bit(I40E_FLAG_FD_ATR_ENA, pf->flags);
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index 2032465c9b7c..acb47a9fd86f 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -2887,6 +2887,7 @@ static void i40e_atr(struct i40e_ring *tx_ring, struct sk_buff *skb,
struct tcphdr *th;
unsigned int hlen;
u32 flex_ptype, dtype_cmd;
+ u32 sample_rate;
int l4_proto;
u16 i;
@@ -2898,7 +2899,8 @@ static void i40e_atr(struct i40e_ring *tx_ring, struct sk_buff *skb,
return;
/* if sampling is disabled do nothing */
- if (!tx_ring->atr_sample_rate)
+ sample_rate = READ_ONCE(pf->atr_sample_rate);
+ if (!sample_rate)
return;
/* Currently only IPv4/IPv6 with TCP is supported */
@@ -2950,7 +2952,7 @@ static void i40e_atr(struct i40e_ring *tx_ring, struct sk_buff *skb,
if (!th->fin &&
!th->syn &&
!th->rst &&
- (tx_ring->atr_count < tx_ring->atr_sample_rate))
+ (tx_ring->atr_count < sample_rate))
return;
tx_ring->atr_count = 0;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.h b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
index bb741ff3e5f2..be587f804e7a 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
@@ -372,8 +372,7 @@ struct i40e_ring {
u16 next_to_clean;
u16 xdp_tx_active;
- u8 atr_sample_rate;
- u8 atr_count;
+ u32 atr_count;
bool ring_active; /* is ring online or not */
bool arm_wb; /* do something to arm write back */
--
2.56.0
^ permalink raw reply related [flat|nested] 3+ messages in thread* [PATCH iwl-next v5 2/2] i40e: add devlink parameter for Flow Director ATR sample rate
2026-10-08 13:18 [PATCH iwl-next v5 0/2] i40e: add runtime-tunable ATR sample rate mheib
2026-10-08 13:18 ` [PATCH iwl-next v5 1/2] i40e: move ATR sample rate from ring to PF level mheib
@ 2026-10-08 13:18 ` mheib
1 sibling, 0 replies; 3+ messages in thread
From: mheib @ 2026-10-08 13:18 UTC (permalink / raw)
To: intel-wired-lan
Cc: netdev, jiri, davem, edumazet, kuba, pabeni, anthony.l.nguyen,
przemyslaw.kitszel, andrew+netdev, horms, Mohammad Heib
From: Mohammad Heib <mheib@redhat.com>
The i40e driver uses Flow Director ATR to periodically update flow
steering information for active TCP flows. The update frequency is
currently controlled by I40E_DEFAULT_ATR_SAMPLE_RATE and is fixed at
driver build time.
On systems with a large number of queues and high-rate TCP workloads,
the default sampling interval can result in frequent Flow Director
reprogramming for long-lived flows.
The amount of TCP packet reordering observed on some systems is
sensitive to the ATR sampling interval. Increasing the interval reduces
Flow Director programming activity and can significantly reduce the
associated reordering.
Since the optimal sampling interval depends on the workload and system
configuration, a single fixed value is not suitable for all deployments.
Add a devlink parameter to allow administrators to tune the ATR sample
rate at runtime without rebuilding the driver or disabling ATR
functionality entirely.
Signed-off-by: Mohammad Heib <mheib@redhat.com>
---
Documentation/networking/devlink/i40e.rst | 21 +++++++++
.../net/ethernet/intel/i40e/i40e_devlink.c | 46 +++++++++++++++++++
2 files changed, 67 insertions(+)
diff --git a/Documentation/networking/devlink/i40e.rst b/Documentation/networking/devlink/i40e.rst
index 51c887f0dc83..d280c17fe987 100644
--- a/Documentation/networking/devlink/i40e.rst
+++ b/Documentation/networking/devlink/i40e.rst
@@ -40,6 +40,27 @@ Parameters
The default value is ``0`` (internal calculation is used).
+.. list-table:: Driver specific parameters implemented
+ :widths: 5 5 90
+
+ * - Name
+ - Mode
+ - Notes
+ * - ``atr_sample_rate``
+ - runtime
+ - Controls how frequently Flow Director ATR updates flow steering
+ information for active TCP flows.
+
+ ATR programs Flow Director entries based on sampled TCP packets
+ per Tx queue. The sampling interval is the number of TCP packets
+ transmitted on each queue between ATR updates.
+
+ Lower values increase Flow Director programming activity, while
+ higher values reduce the update frequency.
+
+ Setting to ``0`` disables ATR sampling (no filters will be programmed).
+
+ The default value is ``20``.
Info versions
=============
diff --git a/drivers/net/ethernet/intel/i40e/i40e_devlink.c b/drivers/net/ethernet/intel/i40e/i40e_devlink.c
index 229179ccc131..8af2487776f2 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_devlink.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_devlink.c
@@ -33,12 +33,58 @@ static int i40e_max_mac_per_vf_get(struct devlink *devlink,
return 0;
}
+static int i40e_atr_sample_rate_set(struct devlink *devlink,
+ u32 id,
+ struct devlink_param_gset_ctx *ctx,
+ struct netlink_ext_ack *extack)
+{
+ struct i40e_pf *pf = devlink_priv(devlink);
+ u32 sample_rate = ctx->val.vu32;
+
+ WRITE_ONCE(pf->atr_sample_rate, sample_rate);
+
+ /* Setting rate to 0 disables sampling but does not remove
+ * existing ATR entries. Flush them so stale filters do not
+ * persist on devices without HW ATR eviction.
+ */
+ if (!sample_rate) {
+ set_bit(__I40E_FD_ATR_AUTO_DISABLED, pf->state);
+ set_bit(__I40E_FD_FLUSH_REQUESTED, pf->state);
+ }
+
+ return 0;
+}
+
+static int i40e_atr_sample_rate_get(struct devlink *devlink,
+ u32 id,
+ struct devlink_param_gset_ctx *ctx,
+ struct netlink_ext_ack *extack)
+{
+ struct i40e_pf *pf = devlink_priv(devlink);
+
+ ctx->val.vu32 = READ_ONCE(pf->atr_sample_rate);
+
+ return 0;
+}
+
+enum i40e_dl_param_id {
+ I40E_DEVLINK_PARAM_ID_BASE = DEVLINK_PARAM_GENERIC_ID_MAX,
+ I40E_DEVLINK_PARAM_ID_ATR_SAMPLE_RATE,
+};
+
static const struct devlink_param i40e_dl_params[] = {
DEVLINK_PARAM_GENERIC(MAX_MAC_PER_VF,
BIT(DEVLINK_PARAM_CMODE_RUNTIME),
i40e_max_mac_per_vf_get,
i40e_max_mac_per_vf_set,
NULL),
+ DEVLINK_PARAM_DRIVER(I40E_DEVLINK_PARAM_ID_ATR_SAMPLE_RATE,
+ "atr_sample_rate",
+ DEVLINK_PARAM_TYPE_U32,
+ BIT(DEVLINK_PARAM_CMODE_RUNTIME),
+ i40e_atr_sample_rate_get,
+ i40e_atr_sample_rate_set,
+ NULL),
};
static void i40e_info_get_dsn(struct i40e_pf *pf, char *buf, size_t len)
--
2.56.0
^ permalink raw reply related [flat|nested] 3+ messages in thread