net: dev: Makes sure netif_rx() can be invoked in any context.

Dave suggested a while ago (eleven years by now) "Let's make netif_rx()
work in all contexts and get rid of netif_rx_ni()". Eric agreed and
pointed out that modern devices should use netif_receive_skb() to avoid
the overhead.
In the meantime someone added another variant, netif_rx_any_context(),
which behaves as suggested.

netif_rx() must be invoked with disabled bottom halves to ensure that
pending softirqs, which were raised within the function, are handled.
netif_rx_ni() can be invoked only from process context (bottom halves
must be enabled) because the function handles pending softirqs without
checking if bottom halves were disabled or not.
netif_rx_any_context() invokes on the former functions by checking
in_interrupts().

netif_rx() could be taught to handle both cases (disabled and enabled
bottom halves) by simply disabling bottom halves while invoking
netif_rx_internal(). The local_bh_enable() invocation will then invoke
pending softirqs only if the BH-disable counter drops to zero.

Eric is concerned about the overhead of BH-disable+enable especially in
regard to the loopback driver. As critical as this driver is, it will
receive a shortcut to avoid the additional overhead which is not needed.

Add a local_bh_disable() section in netif_rx() to ensure softirqs are
handled if needed.
Provide __netif_rx() which does not disable BH and has a lockdep assert
to ensure that interrupts are disabled. Use this shortcut in the
loopback driver and in drivers/net/*.c.
Make netif_rx_ni() and netif_rx_any_context() invoke netif_rx() so they
can be removed once they are no more users left.

Link: https://lkml.kernel.org/r/[e-mail]
Signed-off-by: Sebastian Andrzej Siewior <[e-mail]>
Reviewed-by: Eric Dumazet <[e-mail]>
Reviewed-by: Toke Høiland-Jørgensen <[e-mail]>
Signed-off-by: David S. Miller <[e-mail]>

drivers/net/amt.c

···
 	skb->pkt_type = PACKET_MULTICAST;
 	skb->ip_summed = CHECKSUM_NONE;
 	len = skb->len;
-	if (netif_rx(skb) == NET_RX_SUCCESS) {
+	if (__netif_rx(skb) == NET_RX_SUCCESS) {
 		amt_update_gw_status(amt, AMT_STATUS_RECEIVED_QUERY, true);
 		dev_sw_netstats_rx_add(amt->dev, len);
 	} else {
···
 	skb->pkt_type = PACKET_MULTICAST;
 	skb->ip_summed = CHECKSUM_NONE;
 	len = skb->len;
-	if (netif_rx(skb) == NET_RX_SUCCESS) {
+	if (__netif_rx(skb) == NET_RX_SUCCESS) {
 		amt_update_relay_status(tunnel, AMT_STATUS_RECEIVED_UPDATE,
 					true);
 		dev_sw_netstats_rx_add(amt->dev, len);

drivers/net/geneve.c

···
 		}
 
 		skb->protocol = eth_type_trans(skb, geneve->dev);
-		netif_rx(skb);
+		__netif_rx(skb);
 		dst_release(&rt->dst);
 		return -EMSGSIZE;
 	}
···
 		}
 
 		skb->protocol = eth_type_trans(skb, geneve->dev);
-		netif_rx(skb);
+		__netif_rx(skb);
 		dst_release(dst);
 		return -EMSGSIZE;
 	}

drivers/net/gtp.c

···
 
 	dev_sw_netstats_rx_add(pctx->dev, skb->len);
 
-	netif_rx(skb);
+	__netif_rx(skb);
 	return 0;
 
 err:

drivers/net/loopback.c

···
 
 	skb_orphan(skb);
 
-	/* Before queueing this packet to netif_rx(),
+	/* Before queueing this packet to __netif_rx(),
 	 * make sure dst is refcounted.
 	 */
 	skb_dst_force(skb);
···
 	skb->protocol = eth_type_trans(skb, dev);
 
 	len = skb->len;
-	if (likely(netif_rx(skb) == NET_RX_SUCCESS))
+	if (likely(__netif_rx(skb) == NET_RX_SUCCESS))
 		dev_lstats_add(dev, len);
 
 	return NETDEV_TX_OK;

drivers/net/macsec.c

···
 				else
 					nskb->pkt_type = PACKET_MULTICAST;
 
-				netif_rx(nskb);
+				__netif_rx(nskb);
 			}
 			continue;
 		}
···
 
 		nskb->dev = ndev;
 
-		if (netif_rx(nskb) == NET_RX_SUCCESS) {
+		if (__netif_rx(nskb) == NET_RX_SUCCESS) {
 			u64_stats_update_begin(&secy_stats->syncp);
 			secy_stats->stats.InPktsUntagged++;
 			u64_stats_update_end(&secy_stats->syncp);
···
 
 		macsec_reset_skb(nskb, macsec->secy.netdev);
 
-		ret = netif_rx(nskb);
+		ret = __netif_rx(nskb);
 		if (ret == NET_RX_SUCCESS) {
 			u64_stats_update_begin(&secy_stats->syncp);
 			secy_stats->stats.InPktsUnknownSCI++;

drivers/net/macvlan.c

···
 	if (ether_addr_equal_64bits(eth_hdr(skb)->h_dest, dev->dev_addr))
 		nskb->pkt_type = PACKET_HOST;
 
-	ret = netif_rx(nskb);
+	ret = __netif_rx(nskb);
 	macvlan_count_rx(vlan, len, ret == NET_RX_SUCCESS, false);
 }
 
···
 			/* forward to original port. */
 			vlan = src;
 			ret = macvlan_broadcast_one(skb, vlan, eth, 0) ?:
-			      netif_rx(skb);
+			      __netif_rx(skb);
 			handle_res = RX_HANDLER_CONSUMED;
 			goto out;
 		}

drivers/net/mhi_net.c

···
 		u64_stats_inc(&mhi_netdev->stats.rx_packets);
 		u64_stats_add(&mhi_netdev->stats.rx_bytes, skb->len);
 		u64_stats_update_end(&mhi_netdev->stats.rx_syncp);
-		netif_rx(skb);
+		__netif_rx(skb);
 	}
 
 	/* Refill if RX buffers queue becomes low */

drivers/net/ntb_netdev.c

···
 	skb->protocol = eth_type_trans(skb, ndev);
 	skb->ip_summed = CHECKSUM_NONE;
 
-	if (netif_rx(skb) == NET_RX_DROP) {
+	if (__netif_rx(skb) == NET_RX_DROP) {
 		ndev->stats.rx_errors++;
 		ndev->stats.rx_dropped++;
 	} else {

drivers/net/rionet.c

···
 		skb_put(rnet->rx_skb[i], RIO_MAX_MSG_SIZE);
 		rnet->rx_skb[i]->protocol =
 		    eth_type_trans(rnet->rx_skb[i], ndev);
-		error = netif_rx(rnet->rx_skb[i]);
+		error = __netif_rx(rnet->rx_skb[i]);
 
 		if (error == NET_RX_DROP) {
 			ndev->stats.rx_dropped++;

drivers/net/sb1000.c

···
 
 	/* datagram completed: send to upper level */
 	skb_trim(skb, dlen);
-	netif_rx(skb);
+	__netif_rx(skb);
 	stats->rx_bytes+=dlen;
 	stats->rx_packets++;
 	lp->rx_skb[ns] = NULL;

drivers/net/veth.c

···
 {
 	return __dev_forward_skb(dev, skb) ?: xdp ?
 		veth_xdp_rx(rq, skb) :
-		netif_rx(skb);
+		__netif_rx(skb);
 }
 
 /* return true if the specified skb has chances of GRO aggregation

drivers/net/vrf.c

···
 
 	skb->protocol = eth_type_trans(skb, dev);
 
-	if (likely(netif_rx(skb) == NET_RX_SUCCESS))
+	if (likely(__netif_rx(skb) == NET_RX_SUCCESS))
 		vrf_rx_stats(dev, len);
 	else
 		this_cpu_inc(dev->dstats->rx_drps);

drivers/net/vxlan.c

···
 	tx_stats->tx_bytes += len;
 	u64_stats_update_end(&tx_stats->syncp);
 
-	if (netif_rx(skb) == NET_RX_SUCCESS) {
+	if (__netif_rx(skb) == NET_RX_SUCCESS) {
 		u64_stats_update_begin(&rx_stats->syncp);
 		rx_stats->rx_packets++;
 		rx_stats->rx_bytes += len;

include/linux/netdevice.h

···
 void generic_xdp_tx(struct sk_buff *skb, struct bpf_prog *xdp_prog);
 int do_xdp_generic(struct bpf_prog *xdp_prog, struct sk_buff *skb);
 int netif_rx(struct sk_buff *skb);
-int netif_rx_ni(struct sk_buff *skb);
-int netif_rx_any_context(struct sk_buff *skb);
+int __netif_rx(struct sk_buff *skb);
+
+static inline int netif_rx_ni(struct sk_buff *skb)
+{
+	return netif_rx(skb);
+}
+
+static inline int netif_rx_any_context(struct sk_buff *skb)
+{
+	return netif_rx(skb);
+}
+
 int netif_receive_skb(struct sk_buff *skb);
 int netif_receive_skb_core(struct sk_buff *skb);
 void netif_receive_skb_list_internal(struct list_head *head);

include/trace/events/net.h

···
 	TP_ARGS(skb)
 );
 
-DEFINE_EVENT(net_dev_rx_verbose_template, netif_rx_ni_entry,
-
-	TP_PROTO(const struct sk_buff *skb),
-
-	TP_ARGS(skb)
-);
-
 DECLARE_EVENT_CLASS(net_dev_rx_exit_template,
 
 	TP_PROTO(int ret),
···
 );
 
 DEFINE_EVENT(net_dev_rx_exit_template, netif_rx_exit,
-
-	TP_PROTO(int ret),
-
-	TP_ARGS(ret)
-);
-
-DEFINE_EVENT(net_dev_rx_exit_template, netif_rx_ni_exit,
 
 	TP_PROTO(int ret),
 

net/core/dev.c

···
 }
 
 /**
+ *	__netif_rx	-	Slightly optimized version of netif_rx
+ *	@skb: buffer to post
+ *
+ *	This behaves as netif_rx except that it does not disable bottom halves.
+ *	As a result this function may only be invoked from the interrupt context
+ *	(either hard or soft interrupt).
+ */
+int __netif_rx(struct sk_buff *skb)
+{
+	int ret;
+
+	lockdep_assert_once(hardirq_count() | softirq_count());
+
+	trace_netif_rx_entry(skb);
+	ret = netif_rx_internal(skb);
+	trace_netif_rx_exit(ret);
+	return ret;
+}
+EXPORT_SYMBOL(__netif_rx);
+
+/**
  *	netif_rx	-	post buffer to the network code
  *	@skb: buffer to post
  *
  *	This function receives a packet from a device driver and queues it for
- *	the upper (protocol) levels to process.  It always succeeds. The buffer
- *	may be dropped during processing for congestion control or by the
- *	protocol layers.
+ *	the upper (protocol) levels to process via the backlog NAPI device. It
+ *	always succeeds. The buffer may be dropped during processing for
+ *	congestion control or by the protocol layers.
+ *	The network buffer is passed via the backlog NAPI device. Modern NIC
+ *	driver should use NAPI and GRO.
+ *	This function can used from any context.
  *
  *	return values:
  *	NET_RX_SUCCESS	(no congestion)
  *	NET_RX_DROP     (packet was dropped)
  *
  */
-
 int netif_rx(struct sk_buff *skb)
 {
 	int ret;
 
+	local_bh_disable();
 	trace_netif_rx_entry(skb);
-
 	ret = netif_rx_internal(skb);
 	trace_netif_rx_exit(ret);
-
+	local_bh_enable();
 	return ret;
 }
 EXPORT_SYMBOL(netif_rx);
-
-int netif_rx_ni(struct sk_buff *skb)
-{
-	int err;
-
-	trace_netif_rx_ni_entry(skb);
-
-	preempt_disable();
-	err = netif_rx_internal(skb);
-	if (local_softirq_pending())
-		do_softirq();
-	preempt_enable();
-	trace_netif_rx_ni_exit(err);
-
-	return err;
-}
-EXPORT_SYMBOL(netif_rx_ni);
-
-int netif_rx_any_context(struct sk_buff *skb)
-{
-	/*
-	 * If invoked from contexts which do not invoke bottom half
-	 * processing either at return from interrupt or when softrqs are
-	 * reenabled, use netif_rx_ni() which invokes bottomhalf processing
-	 * directly.
-	 */
-	if (in_interrupt())
-		return netif_rx(skb);
-	else
-		return netif_rx_ni(skb);
-}
-EXPORT_SYMBOL(netif_rx_any_context);
 
 static __latent_entropy void net_tx_action(struct softirq_action *h)
 {