net: add IFLA_MAX_PACING_OFFLOAD_HORIZON device attribute

Some network devices have the ability to offload EDT (Earliest
Departure Time) which is the model used for TCP pacing and FQ
packet scheduler.

Some of them implement the timing wheel mechanism described in
https://saeed.github.io/files/carousel-sigcomm17.pdf
with an associated 'timing wheel horizon'.

This patch adds dev->max_pacing_offload_horizon expressing
this timing wheel horizon in nsec units.

This is a read-only attribute.

Unless a driver sets it, dev->max_pacing_offload_horizon
is zero.

v2: addressed Jakub feedback ( https://lore.kernel.org/netdev/20240930152304.472767-2-edumazet@google.com/T/#mf6294d714c41cc459962154cc2580ce3c9693663 )
v3: added yaml doc (also per Jakub feedback)

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://patch.msgid.link/20241003121219.2396589-2-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Author: edumazet

Documentation/netlink/specs/rt_link.yaml

@@ -1137,6 +1137,10 @@ attribute-sets:
         name: dpll-pin
         type: nest
         nested-attributes: link-dpll-pin-attrs
+      -
+        name: max-pacing-offload-horizon
+        type: uint
+        doc: EDT offload horizon supported by the device (in nsec).
   -
     name: af-spec-attrs
     attributes:

Documentation/networking/net_cachelines/net_device.rst

@@ -183,3 +183,4 @@ struct_devlink_port*                devlink_port
 struct_dpll_pin*                    dpll_pin                                                        
 struct hlist_head                   page_pools
 struct dim_irq_moder*               irq_moder
+u64                                 max_pacing_offload_horizon

include/linux/netdevice.h

@@ -2009,6 +2009,8 @@ enum netdev_reg_state {
  *	@dpll_pin: Pointer to the SyncE source pin of a DPLL subsystem,
  *		   where the clock is recovered.
  *
+ *	@max_pacing_offload_horizon: max EDT offload horizon in nsec.
+ *
  *	FIXME: cleanup struct net_device such that network protocol info
  *	moves out.
  */
@@ -2399,6 +2401,8 @@ struct net_device {
 	/** @irq_moder: dim parameters used if IS_ENABLED(CONFIG_DIMLIB). */
 	struct dim_irq_moder	*irq_moder;
 
+	u64			max_pacing_offload_horizon;
+
 	u8			priv[] ____cacheline_aligned
 				       __counted_by(priv_len);
 } ____cacheline_aligned;

include/uapi/linux/if_link.h

@@ -377,6 +377,7 @@ enum {
 	IFLA_GSO_IPV4_MAX_SIZE,
 	IFLA_GRO_IPV4_MAX_SIZE,
 	IFLA_DPLL_PIN,
+	IFLA_MAX_PACING_OFFLOAD_HORIZON,
 	__IFLA_MAX
 };
 

net/core/rtnetlink.c

@@ -1118,6 +1118,7 @@ static noinline size_t if_nlmsg_size(const struct net_device *dev,
 	       + nla_total_size(MAX_ADDR_LEN) /* IFLA_PERM_ADDRESS */
 	       + rtnl_devlink_port_size(dev)
 	       + rtnl_dpll_pin_size(dev)
+	       + nla_total_size(8)  /* IFLA_MAX_PACING_OFFLOAD_HORIZON */
 	       + 0;
 }
 
@@ -1867,6 +1868,8 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb,
 			READ_ONCE(dev->tso_max_size)) ||
 	    nla_put_u32(skb, IFLA_TSO_MAX_SEGS,
 			READ_ONCE(dev->tso_max_segs)) ||
+	    nla_put_uint(skb, IFLA_MAX_PACING_OFFLOAD_HORIZON,
+			 READ_ONCE(dev->max_pacing_offload_horizon)) ||
 #ifdef CONFIG_RPS
 	    nla_put_u32(skb, IFLA_NUM_RX_QUEUES,
 			READ_ONCE(dev->num_rx_queues)) ||
@@ -1975,6 +1978,7 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb,
 }
 
 static const struct nla_policy ifla_policy[IFLA_MAX+1] = {
+	[IFLA_UNSPEC]		= { .strict_start_type = IFLA_DPLL_PIN },
 	[IFLA_IFNAME]		= { .type = NLA_STRING, .len = IFNAMSIZ-1 },
 	[IFLA_ADDRESS]		= { .type = NLA_BINARY, .len = MAX_ADDR_LEN },
 	[IFLA_BROADCAST]	= { .type = NLA_BINARY, .len = MAX_ADDR_LEN },

tools/include/uapi/linux/if_link.h

@@ -377,6 +377,7 @@ enum {
 	IFLA_GSO_IPV4_MAX_SIZE,
 	IFLA_GRO_IPV4_MAX_SIZE,
 	IFLA_DPLL_PIN,
+	IFLA_MAX_PACING_OFFLOAD_HORIZON,
 	__IFLA_MAX
 };