ethtool: provide customized dim profile management

The NetDIM library, currently leveraged by an array of NICs, delivers
excellent acceleration benefits. Nevertheless, NICs vary significantly
in their dim profile list prerequisites.

Specifically, virtio-net backends may present diverse sw or hw device
implementation, making a one-size-fits-all parameter list impractical.
On Alibaba Cloud, the virtio DPU's performance under the default DIM
profile falls short of expectations, partly due to a mismatch in
parameter configuration.

I also noticed that ice/idpf/ena and other NICs have customized
profilelist or placed some restrictions on dim capabilities.

Motivated by this, I tried adding new params for "ethtool -C" that provides
a per-device control to modify and access a device's interrupt parameters.

Usage
========
The target NIC is named ethx.

Assume that ethx only declares support for rx profile setting
(with DIM_PROFILE_RX flag set in profile_flags) and supports modification
of usec and pkt fields.

1. Query the currently customized list of the device

$ ethtool -c ethx
...
rx-profile:
{.usec =   1, .pkts = 256, .comps = n/a,},
{.usec =   8, .pkts = 256, .comps = n/a,},
{.usec =  64, .pkts = 256, .comps = n/a,},
{.usec = 128, .pkts = 256, .comps = n/a,},
{.usec = 256, .pkts = 256, .comps = n/a,}
tx-profile:   n/a

2. Tune
$ ethtool -C ethx rx-profile 1,1,n_2,n,n_3,3,n_4,4,n_n,5,n
"n" means do not modify this field.
$ ethtool -c ethx
...
rx-profile:
{.usec =   1, .pkts =   1, .comps = n/a,},
{.usec =   2, .pkts = 256, .comps = n/a,},
{.usec =   3, .pkts =   3, .comps = n/a,},
{.usec =   4, .pkts =   4, .comps = n/a,},
{.usec = 256, .pkts =   5, .comps = n/a,}
tx-profile:   n/a

3. Hint
If the device does not support some type of customized dim profiles,
the corresponding "n/a" will display.

If the "n/a" field is being modified, -EOPNOTSUPP will be reported.

Signed-off-by: Heng Qi <hengqi@linux.alibaba.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20240621101353.107425-4-hengqi@linux.alibaba.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Documentation/netlink/specs/ethtool.yaml

@@ -414,6 +414,26 @@ attribute-sets:
         name: combined-count
         type: u32
 
+  -
+    name: irq-moderation
+    attributes:
+      -
+        name: usec
+        type: u32
+      -
+        name: pkts
+        type: u32
+      -
+        name: comps
+        type: u32
+  -
+    name: profile
+    attributes:
+      -
+        name: irq-moderation
+        type: nest
+        multi-attr: true
+        nested-attributes: irq-moderation
   -
     name: coalesce
     attributes:
@@ -502,6 +522,15 @@ attribute-sets:
       -
         name: tx-aggr-time-usecs
         type: u32
+      -
+        name: rx-profile
+        type: nest
+        nested-attributes: profile
+      -
+        name: tx-profile
+        type: nest
+        nested-attributes: profile
+
   -
     name: pause-stat
     attributes:
@@ -1325,6 +1354,8 @@ operations:
             - tx-aggr-max-bytes
             - tx-aggr-max-frames
             - tx-aggr-time-usecs
+            - rx-profile
+            - tx-profile
       dump: *coalesce-get-op
     -
       name: coalesce-set

Documentation/networking/ethtool-netlink.rst

@@ -1033,6 +1033,8 @@ Kernel response contents:
   ``ETHTOOL_A_COALESCE_TX_AGGR_MAX_BYTES``     u32     max aggr size, Tx
   ``ETHTOOL_A_COALESCE_TX_AGGR_MAX_FRAMES``    u32     max aggr packets, Tx
   ``ETHTOOL_A_COALESCE_TX_AGGR_TIME_USECS``    u32     time (us), aggr, Tx
+  ``ETHTOOL_A_COALESCE_RX_PROFILE``            nested  profile of DIM, Rx
+  ``ETHTOOL_A_COALESCE_TX_PROFILE``            nested  profile of DIM, Tx
   ===========================================  ======  =======================
 
 Attributes are only included in reply if their value is not zero or the
@@ -1062,6 +1064,10 @@ block should be sent.
 This feature is mainly of interest for specific USB devices which does not cope
 well with frequent small-sized URBs transmissions.
 
+``ETHTOOL_A_COALESCE_RX_PROFILE`` and ``ETHTOOL_A_COALESCE_TX_PROFILE`` refer
+to DIM parameters, see `Generic Network Dynamic Interrupt Moderation (Net DIM)
+<https://www.kernel.org/doc/Documentation/networking/net_dim.rst>`_.
+
 COALESCE_SET
 ============
 
@@ -1098,6 +1104,8 @@ Request contents:
   ``ETHTOOL_A_COALESCE_TX_AGGR_MAX_BYTES``     u32     max aggr size, Tx
   ``ETHTOOL_A_COALESCE_TX_AGGR_MAX_FRAMES``    u32     max aggr packets, Tx
   ``ETHTOOL_A_COALESCE_TX_AGGR_TIME_USECS``    u32     time (us), aggr, Tx
+  ``ETHTOOL_A_COALESCE_RX_PROFILE``            nested  profile of DIM, Rx
+  ``ETHTOOL_A_COALESCE_TX_PROFILE``            nested  profile of DIM, Tx
   ===========================================  ======  =======================
 
 Request is rejected if it attributes declared as unsupported by driver (i.e.

Documentation/networking/net_dim.rst

@@ -169,6 +169,48 @@ usage is not complete but it should make the outline of the usage clear.
 	...
   }
 
+
+Tuning DIM
+==========
+
+Net DIM serves a range of network devices and delivers excellent acceleration
+benefits. Yet, it has been observed that some preset configurations of DIM may
+not align seamlessly with the varying specifications of network devices, and
+this discrepancy has been identified as a factor to the suboptimal performance
+outcomes of DIM-enabled network devices, related to a mismatch in profiles.
+
+To address this issue, Net DIM introduces a per-device control to modify and
+access a device's ``rx-profile`` and ``tx-profile`` parameters:
+Assume that the target network device is named ethx, and ethx only declares
+support for RX profile setting and supports modification of ``usec`` field
+and ``pkts`` field (See the data structure:
+:c:type:`struct dim_cq_moder <dim_cq_moder>`).
+
+You can use ethtool to modify the current RX DIM profile where all
+values are 64::
+
+    $ ethtool -C ethx rx-profile 1,1,n_2,2,n_3,n,n_n,4,n_n,n,n
+
+``n`` means do not modify this field, and ``_`` separates structure
+elements of the profile array.
+
+Querying the current profiles using::
+
+    $ ethtool -c ethx
+    ...
+    rx-profile:
+    {.usec =   1, .pkts =   1, .comps = n/a,},
+    {.usec =   2, .pkts =   2, .comps = n/a,},
+    {.usec =   3, .pkts =  64, .comps = n/a,},
+    {.usec =  64, .pkts =   4, .comps = n/a,},
+    {.usec =  64, .pkts =  64, .comps = n/a,}
+    tx-profile:   n/a
+
+If the network device does not support specific fields of DIM profiles,
+the corresponding ``n/a`` will display. If the ``n/a`` field is being
+modified, error messages will be reported.
+
+
 Dynamic Interrupt Moderation (DIM) library API
 ==============================================
 

include/linux/dim.h

@@ -10,6 +10,8 @@
 #include <linux/types.h>
 #include <linux/workqueue.h>
 
+struct net_device;
+
 /* Number of DIM profiles and period mode. */
 #define NET_DIM_PARAMS_NUM_PROFILES 5
 #define NET_DIM_DEFAULT_RX_CQ_PKTS_FROM_EQE 256
@@ -45,12 +47,45 @@
  * @pkts: CQ packet counter suggestion (by DIM)
  * @comps: Completion counter
  * @cq_period_mode: CQ period count mode (from CQE/EQE)
+ * @rcu: for asynchronous kfree_rcu
  */
 struct dim_cq_moder {
 	u16 usec;
 	u16 pkts;
 	u16 comps;
 	u8 cq_period_mode;
+	struct rcu_head rcu;
+};
+
+#define DIM_PROFILE_RX		BIT(0)	/* support rx profile modification */
+#define DIM_PROFILE_TX		BIT(1)	/* support tx profile modification */
+
+#define DIM_COALESCE_USEC	BIT(0)	/* support usec field modification */
+#define DIM_COALESCE_PKTS	BIT(1)	/* support pkts field modification */
+#define DIM_COALESCE_COMPS	BIT(2)	/* support comps field modification */
+
+/**
+ * struct dim_irq_moder - Structure for irq moderation information.
+ * Used to collect irq moderation related information.
+ *
+ * @profile_flags: DIM_PROFILE_*
+ * @coal_flags: DIM_COALESCE_* for Rx and Tx
+ * @dim_rx_mode: Rx DIM period count mode: CQE or EQE
+ * @dim_tx_mode: Tx DIM period count mode: CQE or EQE
+ * @rx_profile: DIM profile list for Rx
+ * @tx_profile: DIM profile list for Tx
+ * @rx_dim_work: Rx DIM worker scheduled by net_dim()
+ * @tx_dim_work: Tx DIM worker scheduled by net_dim()
+ */
+struct dim_irq_moder {
+	u8 profile_flags;
+	u8 coal_flags;
+	u8 dim_rx_mode;
+	u8 dim_tx_mode;
+	struct dim_cq_moder __rcu *rx_profile;
+	struct dim_cq_moder __rcu *tx_profile;
+	void (*rx_dim_work)(struct work_struct *work);
+	void (*tx_dim_work)(struct work_struct *work);
 };
 
 /**
@@ -198,6 +233,29 @@ enum dim_step_result {
 	DIM_ON_EDGE,
 };
 
+/**
+ * net_dim_init_irq_moder - collect information to initialize irq moderation
+ * @dev: target network device
+ * @profile_flags: Rx or Tx profile modification capability
+ * @coal_flags: irq moderation params flags
+ * @rx_mode: CQ period mode for Rx
+ * @tx_mode: CQ period mode for Tx
+ * @rx_dim_work: Rx worker called after dim decision
+ * @tx_dim_work: Tx worker called after dim decision
+ *
+ * Return: 0 on success or a negative error code.
+ */
+int net_dim_init_irq_moder(struct net_device *dev, u8 profile_flags,
+			   u8 coal_flags, u8 rx_mode, u8 tx_mode,
+			   void (*rx_dim_work)(struct work_struct *work),
+			   void (*tx_dim_work)(struct work_struct *work));
+
+/**
+ * net_dim_free_irq_moder - free fields for irq moderation
+ * @dev: target network device
+ */
+void net_dim_free_irq_moder(struct net_device *dev);
+
 /**
  *	dim_on_top - check if current state is a good place to stop (top location)
  *	@dim: DIM context

include/linux/ethtool.h

@@ -284,7 +284,9 @@ bool ethtool_convert_link_mode_to_legacy_u32(u32 *legacy_u32,
 #define ETHTOOL_COALESCE_TX_AGGR_MAX_BYTES	BIT(24)
 #define ETHTOOL_COALESCE_TX_AGGR_MAX_FRAMES	BIT(25)
 #define ETHTOOL_COALESCE_TX_AGGR_TIME_USECS	BIT(26)
-#define ETHTOOL_COALESCE_ALL_PARAMS		GENMASK(26, 0)
+#define ETHTOOL_COALESCE_RX_PROFILE		BIT(27)
+#define ETHTOOL_COALESCE_TX_PROFILE		BIT(28)
+#define ETHTOOL_COALESCE_ALL_PARAMS		GENMASK(28, 0)
 
 #define ETHTOOL_COALESCE_USECS						\
 	(ETHTOOL_COALESCE_RX_USECS | ETHTOOL_COALESCE_TX_USECS)

include/linux/netdevice.h

@@ -2402,6 +2402,9 @@ struct net_device {
 	/** @page_pools: page pools created for this netdevice */
 	struct hlist_head	page_pools;
 #endif
+
+	/** @irq_moder: dim parameters used if IS_ENABLED(CONFIG_DIMLIB). */
+	struct dim_irq_moder	*irq_moder;
 };
 #define to_net_dev(d) container_of(d, struct net_device, dev)
 

include/uapi/linux/ethtool_netlink.h

@@ -415,12 +415,34 @@ enum {
 	ETHTOOL_A_COALESCE_TX_AGGR_MAX_BYTES,		/* u32 */
 	ETHTOOL_A_COALESCE_TX_AGGR_MAX_FRAMES,		/* u32 */
 	ETHTOOL_A_COALESCE_TX_AGGR_TIME_USECS,		/* u32 */
+	/* nest - _A_PROFILE_IRQ_MODERATION */
+	ETHTOOL_A_COALESCE_RX_PROFILE,
+	/* nest - _A_PROFILE_IRQ_MODERATION */
+	ETHTOOL_A_COALESCE_TX_PROFILE,
 
 	/* add new constants above here */
 	__ETHTOOL_A_COALESCE_CNT,
 	ETHTOOL_A_COALESCE_MAX = (__ETHTOOL_A_COALESCE_CNT - 1)
 };
 
+enum {
+	ETHTOOL_A_PROFILE_UNSPEC,
+	/* nest, _A_IRQ_MODERATION_* */
+	ETHTOOL_A_PROFILE_IRQ_MODERATION,
+	__ETHTOOL_A_PROFILE_CNT,
+	ETHTOOL_A_PROFILE_MAX = (__ETHTOOL_A_PROFILE_CNT - 1)
+};
+
+enum {
+	ETHTOOL_A_IRQ_MODERATION_UNSPEC,
+	ETHTOOL_A_IRQ_MODERATION_USEC,			/* u32 */
+	ETHTOOL_A_IRQ_MODERATION_PKTS,			/* u32 */
+	ETHTOOL_A_IRQ_MODERATION_COMPS,			/* u32 */
+
+	__ETHTOOL_A_IRQ_MODERATION_CNT,
+	ETHTOOL_A_IRQ_MODERATION_MAX = (__ETHTOOL_A_IRQ_MODERATION_CNT - 1)
+};
+
 /* PAUSE */
 
 enum {

lib/dim/net_dim.c

@@ -4,6 +4,7 @@
  */
 
 #include <linux/dim.h>
+#include <linux/rtnetlink.h>
 
 /*
  * Net DIM profiles:
@@ -95,6 +96,75 @@ net_dim_get_def_tx_moderation(u8 cq_period_mode)
 }
 EXPORT_SYMBOL(net_dim_get_def_tx_moderation);
 
+int net_dim_init_irq_moder(struct net_device *dev, u8 profile_flags,
+			   u8 coal_flags, u8 rx_mode, u8 tx_mode,
+			   void (*rx_dim_work)(struct work_struct *work),
+			   void (*tx_dim_work)(struct work_struct *work))
+{
+	struct dim_cq_moder *rxp = NULL, *txp;
+	struct dim_irq_moder *moder;
+	int len;
+
+	dev->irq_moder = kzalloc(sizeof(*dev->irq_moder), GFP_KERNEL);
+	if (!dev->irq_moder)
+		return -ENOMEM;
+
+	moder = dev->irq_moder;
+	len = NET_DIM_PARAMS_NUM_PROFILES * sizeof(*moder->rx_profile);
+
+	moder->coal_flags = coal_flags;
+	moder->profile_flags = profile_flags;
+
+	if (profile_flags & DIM_PROFILE_RX) {
+		moder->rx_dim_work = rx_dim_work;
+		moder->dim_rx_mode = rx_mode;
+		rxp = kmemdup(rx_profile[rx_mode], len, GFP_KERNEL);
+		if (!rxp)
+			goto free_moder;
+
+		rcu_assign_pointer(moder->rx_profile, rxp);
+	}
+
+	if (profile_flags & DIM_PROFILE_TX) {
+		moder->tx_dim_work = tx_dim_work;
+		moder->dim_tx_mode = tx_mode;
+		txp = kmemdup(tx_profile[tx_mode], len, GFP_KERNEL);
+		if (!txp)
+			goto free_rxp;
+
+		rcu_assign_pointer(moder->tx_profile, txp);
+	}
+
+	return 0;
+
+free_rxp:
+	kfree(rxp);
+free_moder:
+	kfree(moder);
+	return -ENOMEM;
+}
+EXPORT_SYMBOL(net_dim_init_irq_moder);
+
+/* RTNL lock is held. */
+void net_dim_free_irq_moder(struct net_device *dev)
+{
+	struct dim_cq_moder *rxp, *txp;
+
+	if (!dev->irq_moder)
+		return;
+
+	rxp = rtnl_dereference(dev->irq_moder->rx_profile);
+	txp = rtnl_dereference(dev->irq_moder->tx_profile);
+
+	rcu_assign_pointer(dev->irq_moder->rx_profile, NULL);
+	rcu_assign_pointer(dev->irq_moder->tx_profile, NULL);
+
+	kfree_rcu(rxp, rcu);
+	kfree_rcu(txp, rcu);
+	kfree(dev->irq_moder);
+}
+EXPORT_SYMBOL(net_dim_free_irq_moder);
+
 static int net_dim_step(struct dim *dim)
 {
 	if (dim->tired == (NET_DIM_PARAMS_NUM_PROFILES * 2))

net/Kconfig

@@ -508,6 +508,7 @@ config FAILOVER
 
 config ETHTOOL_NETLINK
 	bool "Netlink interface for ethtool"
+	select DIMLIB
 	default y
 	help
 	  An alternative userspace interface for ethtool based on generic