linux/kernel/bpf/cpumap.c

// SPDX-License-Identifier: GPL-2.0-only
/* bpf/cpumap.c
 *
 * Copyright (c) 2017 Jesper Dangaard Brouer, Red Hat Inc.
 */

/**
 * DOC: cpu map
 * The 'cpumap' is primarily used as a backend map for XDP BPF helper
 * call bpf_redirect_map() and XDP_REDIRECT action, like 'devmap'.
 *
 * Unlike devmap which redirects XDP frames out to another NIC device,
 * this map type redirects raw XDP frames to another CPU.  The remote
 * CPU will do SKB-allocation and call the normal network stack.
 */
/*
 * This is a scalability and isolation mechanism, that allow
 * separating the early driver network XDP layer, from the rest of the
 * netstack, and assigning dedicated CPUs for this stage.  This
 * basically allows for 10G wirespeed pre-filtering via bpf.
 */
#include <linux/bitops.h>
#include <linux/bpf.h>
#include <linux/filter.h>
#include <linux/ptr_ring.h>
#include <net/xdp.h>
#include <net/hotdata.h>

#include <linux/sched.h>
#include <linux/workqueue.h>
#include <linux/kthread.h>
#include <linux/completion.h>
#include <trace/events/xdp.h>
#include <linux/btf_ids.h>

#include <linux/netdevice.h>
#include <net/gro.h>

/* General idea: XDP packets getting XDP redirected to another CPU,
 * will maximum be stored/queued for one driver ->poll() call.  It is
 * guaranteed that queueing the frame and the flush operation happen on
 * same CPU.  Thus, cpu_map_flush operation can deduct via this_cpu_ptr()
 * which queue in bpf_cpu_map_entry contains packets.
 */

#define CPU_MAP_BULK_SIZE 8  /* 8 == one cacheline on 64-bit archs */
struct bpf_cpu_map_entry;
struct bpf_cpu_map;

struct xdp_bulk_queue {
	void *q[CPU_MAP_BULK_SIZE];
	struct list_head flush_node;
	struct bpf_cpu_map_entry *obj;
	unsigned int count;
};

/* Struct for every remote "destination" CPU in map */
struct bpf_cpu_map_entry {
	u32 cpu;    /* kthread CPU and map index */
	int map_id; /* Back reference to map */

	/* XDP can run multiple RX-ring queues, need __percpu enqueue store */
	struct xdp_bulk_queue __percpu *bulkq;

	/* Queue with potential multi-producers, and single-consumer kthread */
	struct ptr_ring *queue;
	struct task_struct *kthread;

	struct bpf_cpumap_val value;
	struct bpf_prog *prog;
	struct gro_node gro;

	struct completion kthread_running;
	struct rcu_work free_work;
};

struct bpf_cpu_map {
	struct bpf_map map;
	/* Below members specific for map type */
	struct bpf_cpu_map_entry __rcu **cpu_map;
};

static struct bpf_map *cpu_map_alloc(union bpf_attr *attr)
{
	u32 value_size = attr->value_size;
	struct bpf_cpu_map *cmap;

	/* check sanity of attributes */
	if (attr->max_entries == 0 || attr->key_size != 4 ||
	    (value_size != offsetofend(struct bpf_cpumap_val, qsize) &&
	     value_size != offsetofend(struct bpf_cpumap_val, bpf_prog.fd)) ||
	    attr->map_flags & ~BPF_F_NUMA_NODE)
		return ERR_PTR(-EINVAL);

	/* Pre-limit array size based on NR_CPUS, not final CPU check */
	if (attr->max_entries > NR_CPUS)
		return ERR_PTR(-E2BIG);

	cmap = bpf_map_area_alloc(sizeof(*cmap), NUMA_NO_NODE);
	if (!cmap)
		return ERR_PTR(-ENOMEM);

	bpf_map_init_from_attr(&cmap->map, attr);

	/* Alloc array for possible remote "destination" CPUs */
	cmap->cpu_map = bpf_map_area_alloc(cmap->map.max_entries *
					   sizeof(struct bpf_cpu_map_entry *),
					   cmap->map.numa_node);
	if (!cmap->cpu_map) {
		bpf_map_area_free(cmap);
		return ERR_PTR(-ENOMEM);
	}

	return &cmap->map;
}

static void __cpu_map_ring_cleanup(struct ptr_ring *ring)
{
	/* The tear-down procedure should have made sure that queue is
	 * empty.  See __cpu_map_entry_replace() and work-queue
	 * invoked cpu_map_kthread_stop(). Catch any broken behaviour
	 * gracefully and warn once.
	 */
	void *ptr;

	while ((ptr = ptr_ring_consume(ring))) {
		WARN_ON_ONCE(1);
		if (unlikely(__ptr_test_bit(0, &ptr))) {
			__ptr_clear_bit(0, &ptr);
			kfree_skb(ptr);
			continue;
		}
		xdp_return_frame(ptr);
	}
}

static u32 cpu_map_bpf_prog_run_skb(struct bpf_cpu_map_entry *rcpu,
				    void **skbs, u32 skb_n,
				    struct xdp_cpumap_stats *stats)
{
	struct xdp_buff xdp;
	u32 act, pass = 0;
	int err;

	for (u32 i = 0; i < skb_n; i++) {
		struct sk_buff *skb = skbs[i];

		act = bpf_prog_run_generic_xdp(skb, &xdp, rcpu->prog);
		switch (act) {
		case XDP_PASS:
			skbs[pass++] = skb;
			break;
		case XDP_REDIRECT:
			err = xdp_do_generic_redirect(skb->dev, skb, &xdp,
						      rcpu->prog);
			if (unlikely(err)) {
				kfree_skb(skb);
				stats->drop++;
			} else {
				stats->redirect++;
			}
			break;
		default:
			bpf_warn_invalid_xdp_action(NULL, rcpu->prog, act);
			fallthrough;
		case XDP_ABORTED:
			trace_xdp_exception(skb->dev, rcpu->prog, act);
			fallthrough;
		case XDP_DROP:
			napi_consume_skb(skb, true);
			stats->drop++;
			break;
		}
	}

	stats->pass += pass;

	return pass;
}

static int cpu_map_bpf_prog_run_xdp(struct bpf_cpu_map_entry *rcpu,
				    void **frames, int n,
				    struct xdp_cpumap_stats *stats)
{
	struct xdp_rxq_info rxq = {};
	struct xdp_buff xdp;
	int i, nframes = 0;

	xdp_set_return_frame_no_direct();
	xdp.rxq = &rxq;

	for (i = 0; i < n; i++) {
		struct xdp_frame *xdpf = frames[i];
		u32 act;
		int err;

		rxq.dev = xdpf->dev_rx;
		rxq.mem.type = xdpf->mem_type;
		/* TODO: report queue_index to xdp_rxq_info */

		xdp_convert_frame_to_buff(xdpf, &xdp);

		act = bpf_prog_run_xdp(rcpu->prog, &xdp);
		switch (act) {
		case XDP_PASS:
			err = xdp_update_frame_from_buff(&xdp, xdpf);
			if (err < 0) {
				xdp_return_frame(xdpf);
				stats->drop++;
			} else {
				frames[nframes++] = xdpf;
			}
			break;
		case XDP_REDIRECT:
			err = xdp_do_redirect(xdpf->dev_rx, &xdp,
					      rcpu->prog);
			if (unlikely(err)) {
				xdp_return_frame(xdpf);
				stats->drop++;
			} else {
				stats->redirect++;
			}
			break;
		default:
			bpf_warn_invalid_xdp_action(NULL, rcpu->prog, act);
			fallthrough;
		case XDP_DROP:
			xdp_return_frame(xdpf);
			stats->drop++;
			break;
		}
	}

	xdp_clear_return_frame_no_direct();
	stats->pass += nframes;

	return nframes;
}

#define CPUMAP_BATCH 8

struct cpu_map_ret {
	u32 xdp_n;
	u32 skb_n;
};

static void cpu_map_bpf_prog_run(struct bpf_cpu_map_entry *rcpu, void **frames,
				 void **skbs, struct cpu_map_ret *ret,
				 struct xdp_cpumap_stats *stats)
{
	struct bpf_net_context __bpf_net_ctx, *bpf_net_ctx;

	if (!rcpu->prog)
		goto out;

	rcu_read_lock();
	bpf_net_ctx = bpf_net_ctx_set(&__bpf_net_ctx);

	ret->xdp_n = cpu_map_bpf_prog_run_xdp(rcpu, frames, ret->xdp_n, stats);
	if (unlikely(ret->skb_n))
		ret->skb_n = cpu_map_bpf_prog_run_skb(rcpu, skbs, ret->skb_n,
						      stats);

	if (stats->redirect)
		xdp_do_flush();

	bpf_net_ctx_clear(bpf_net_ctx);
	rcu_read_unlock();

out:
	if (unlikely(ret->skb_n) && ret->xdp_n)
		memmove(&skbs[ret->xdp_n], skbs, ret->skb_n * sizeof(*skbs));
}

static void cpu_map_gro_flush(struct bpf_cpu_map_entry *rcpu, bool empty)
{
	/*
	 * If the ring is not empty, there'll be a new iteration soon, and we
	 * only need to do a full flush if a tick is long (> 1 ms).
	 * If the ring is empty, to not hold GRO packets in the stack for too
	 * long, do a full flush.
	 * This is equivalent to how NAPI decides whether to perform a full
	 * flush.
	 */
	gro_flush(&rcpu->gro, !empty && HZ >= 1000);
	gro_normal_list(&rcpu->gro);
}

static int cpu_map_kthread_run(void *data)
{
	struct bpf_cpu_map_entry *rcpu = data;
	unsigned long last_qs = jiffies;
	u32 packets = 0;

	complete(&rcpu->kthread_running);
	set_current_state(TASK_INTERRUPTIBLE);

	/* When kthread gives stop order, then rcpu have been disconnected
	 * from map, thus no new packets can enter. Remaining in-flight
	 * per CPU stored packets are flushed to this queue.  Wait honoring
	 * kthread_stop signal until queue is empty.
	 */
	while (!kthread_should_stop() || !__ptr_ring_empty(rcpu->queue)) {
		struct xdp_cpumap_stats stats = {}; /* zero stats */
		unsigned int kmem_alloc_drops = 0, sched = 0;
		struct cpu_map_ret ret = { };
		void *frames[CPUMAP_BATCH];
		void *skbs[CPUMAP_BATCH];
		u32 i, n, m;
		bool empty;

		/* Release CPU reschedule checks */
		if (__ptr_ring_empty(rcpu->queue)) {
			set_current_state(TASK_INTERRUPTIBLE);
			/* Recheck to avoid lost wake-up */
			if (__ptr_ring_empty(rcpu->queue)) {
				schedule();
				sched = 1;
				last_qs = jiffies;
			} else {
				__set_current_state(TASK_RUNNING);
			}
		} else {
			rcu_softirq_qs_periodic(last_qs);
			sched = cond_resched();
		}

		/*
		 * The bpf_cpu_map_entry is single consumer, with this
		 * kthread CPU pinned. Lockless access to ptr_ring
		 * consume side valid as no-resize allowed of queue.
		 */
		n = __ptr_ring_consume_batched(rcpu->queue, frames,
					       CPUMAP_BATCH);
		for (i = 0; i < n; i++) {
			void *f = frames[i];
			struct page *page;

			if (unlikely(__ptr_test_bit(0, &f))) {
				struct sk_buff *skb = f;

				__ptr_clear_bit(0, &skb);
				skbs[ret.skb_n++] = skb;
				continue;
			}

			frames[ret.xdp_n++] = f;
			page = virt_to_page(f);

			/* Bring struct page memory area to curr CPU. Read by
			 * build_skb_around via page_is_pfmemalloc(), and when
			 * freed written by page_frag_free call.
			 */
			prefetchw(page);
		}

		local_bh_disable();

		/* Support running another XDP prog on this CPU */
		cpu_map_bpf_prog_run(rcpu, frames, skbs, &ret, &stats);
		if (!ret.xdp_n)
			goto stats;

		m = napi_skb_cache_get_bulk(skbs, ret.xdp_n);
		if (unlikely(m < ret.xdp_n)) {
			for (i = m; i < ret.xdp_n; i++)
				xdp_return_frame(frames[i]);

			if (ret.skb_n)
				memmove(&skbs[m], &skbs[ret.xdp_n],
					ret.skb_n * sizeof(*skbs));

			kmem_alloc_drops += ret.xdp_n - m;
			ret.xdp_n = m;
		}

		for (i = 0; i < ret.xdp_n; i++) {
			struct xdp_frame *xdpf = frames[i];

			/* Can fail only when !skb -- already handled above */
			__xdp_build_skb_from_frame(xdpf, skbs[i], xdpf->dev_rx);
		}

stats:
		/* Feedback loop via tracepoint.
		 * NB: keep before recv to allow measuring enqueue/dequeue latency.
		 */
		trace_xdp_cpumap_kthread(rcpu->map_id, n, kmem_alloc_drops,
					 sched, &stats);

		for (i = 0; i < ret.xdp_n + ret.skb_n; i++)
			gro_receive_skb(&rcpu->gro, skbs[i]);

		/* Flush either every 64 packets or in case of empty ring */
		packets += n;
		empty = __ptr_ring_empty(rcpu->queue);
		if (packets >= NAPI_POLL_WEIGHT || empty) {
			cpu_map_gro_flush(rcpu, empty);
			packets = 0;
		}

		local_bh_enable(); /* resched point, may call do_softirq() */
	}
	__set_current_state(TASK_RUNNING);

	return 0;
}

static int __cpu_map_load_bpf_program(struct bpf_cpu_map_entry *rcpu,
				      struct bpf_map *map, int fd)
{
	struct bpf_prog *prog;

	prog = bpf_prog_get_type(fd, BPF_PROG_TYPE_XDP);
	if (IS_ERR(prog))
		return PTR_ERR(prog);

	if (prog->expected_attach_type != BPF_XDP_CPUMAP ||
	    !bpf_prog_map_compatible(map, prog)) {
		bpf_prog_put(prog);
		return -EINVAL;
	}

	rcpu->value.bpf_prog.id = prog->aux->id;
	rcpu->prog = prog;

	return 0;
}

static struct bpf_cpu_map_entry *
__cpu_map_entry_alloc(struct bpf_map *map, struct bpf_cpumap_val *value,
		      u32 cpu)
{
	int numa, err, i, fd = value->bpf_prog.fd;
	gfp_t gfp = GFP_KERNEL | __GFP_NOWARN;
	struct bpf_cpu_map_entry *rcpu;
	struct xdp_bulk_queue *bq;

	/* Have map->numa_node, but choose node of redirect target CPU */
	numa = cpu_to_node(cpu);

	rcpu = bpf_map_kmalloc_node(map, sizeof(*rcpu), gfp | __GFP_ZERO, numa);
	if (!rcpu)
		return NULL;

	/* Alloc percpu bulkq */
	rcpu->bulkq = bpf_map_alloc_percpu(map, sizeof(*rcpu->bulkq),
					   sizeof(void *), gfp);
	if (!rcpu->bulkq)
		goto free_rcu;

	for_each_possible_cpu(i) {
		bq = per_cpu_ptr(rcpu->bulkq, i);
		bq->obj = rcpu;
	}

	/* Alloc queue */
	rcpu->queue = bpf_map_kmalloc_node(map, sizeof(*rcpu->queue), gfp,
					   numa);
	if (!rcpu->queue)
		goto free_bulkq;

	err = ptr_ring_init(rcpu->queue, value->qsize, gfp);
	if (err)
		goto free_queue;

	rcpu->cpu    = cpu;
	rcpu->map_id = map->id;
	rcpu->value.qsize  = value->qsize;
	gro_init(&rcpu->gro);

	if (fd > 0 && __cpu_map_load_bpf_program(rcpu, map, fd))
		goto free_ptr_ring;

	/* Setup kthread */
	init_completion(&rcpu->kthread_running);
	rcpu->kthread = kthread_create_on_node(cpu_map_kthread_run, rcpu, numa,
					       "cpumap/%d/map:%d", cpu,
					       map->id);
	if (IS_ERR(rcpu->kthread))
		goto free_prog;

	/* Make sure kthread runs on a single CPU */
	kthread_bind(rcpu->kthread, cpu);
	wake_up_process(rcpu->kthread);

	/* Make sure kthread has been running, so kthread_stop() will not
	 * stop the kthread prematurely and all pending frames or skbs
	 * will be handled by the kthread before kthread_stop() returns.
	 */
	wait_for_completion(&rcpu->kthread_running);

	return rcpu;

free_prog:
	if (rcpu->prog)
		bpf_prog_put(rcpu->prog);
free_ptr_ring:
	gro_cleanup(&rcpu->gro);
	ptr_ring_cleanup(rcpu->queue, NULL);
free_queue:
	kfree(rcpu->queue);
free_bulkq:
	free_percpu(rcpu->bulkq);
free_rcu:
	kfree(rcpu);
	return NULL;
}

static void __cpu_map_entry_free(struct work_struct *work)
{
	struct bpf_cpu_map_entry *rcpu;

	/* This cpu_map_entry have been disconnected from map and one
	 * RCU grace-period have elapsed. Thus, XDP cannot queue any
	 * new packets and cannot change/set flush_needed that can
	 * find this entry.
	 */
	rcpu = container_of(to_rcu_work(work), struct bpf_cpu_map_entry, free_work);

	/* kthread_stop will wake_up_process and wait for it to complete.
	 * cpu_map_kthread_run() makes sure the pointer ring is empty
	 * before exiting.
	 */
	kthread_stop(rcpu->kthread);

	if (rcpu->prog)
		bpf_prog_put(rcpu->prog);
	gro_cleanup(&rcpu->gro);
	/* The queue should be empty at this point */
	__cpu_map_ring_cleanup(rcpu->queue);
	ptr_ring_cleanup(rcpu->queue, NULL);
	kfree(rcpu->queue);
	free_percpu(rcpu->bulkq);
	kfree(rcpu);
}

/* After the xchg of the bpf_cpu_map_entry pointer, we need to make sure the old
 * entry is no longer in use before freeing. We use queue_rcu_work() to call
 * __cpu_map_entry_free() in a separate workqueue after waiting for an RCU grace
 * period. This means that (a) all pending enqueue and flush operations have
 * completed (because of the RCU callback), and (b) we are in a workqueue
 * context where we can stop the kthread and wait for it to exit before freeing
 * everything.
 */
static void __cpu_map_entry_replace(struct bpf_cpu_map *cmap,
				    u32 key_cpu, struct bpf_cpu_map_entry *rcpu)
{
	struct bpf_cpu_map_entry *old_rcpu;

	old_rcpu = unrcu_pointer(xchg(&cmap->cpu_map[key_cpu], RCU_INITIALIZER(rcpu)));
	if (old_rcpu) {
		INIT_RCU_WORK(&old_rcpu->free_work, __cpu_map_entry_free);
		queue_rcu_work(system_wq, &old_rcpu->free_work);
	}
}

static long cpu_map_delete_elem(struct bpf_map *map, void *key)
{
	struct bpf_cpu_map *cmap = container_of(map, struct bpf_cpu_map, map);
	u32 key_cpu = *(u32 *)key;

	if (key_cpu >= map->max_entries)
		return -EINVAL;

	/* notice caller map_delete_elem() uses rcu_read_lock() */
	__cpu_map_entry_replace(cmap, key_cpu, NULL);
	return 0;
}

static long cpu_map_update_elem(struct bpf_map *map, void *key, void *value,
				u64 map_flags)
{
	struct bpf_cpu_map *cmap = container_of(map, struct bpf_cpu_map, map);
	struct bpf_cpumap_val cpumap_value = {};
	struct bpf_cpu_map_entry *rcpu;
	/* Array index key correspond to CPU number */
	u32 key_cpu = *(u32 *)key;

	memcpy(&cpumap_value, value, map->value_size);

	if (unlikely(map_flags > BPF_EXIST))
		return -EINVAL;
	if (unlikely(key_cpu >= cmap->map.max_entries))
		return -E2BIG;
	if (unlikely(map_flags == BPF_NOEXIST))
		return -EEXIST;
	if (unlikely(cpumap_value.qsize > 16384)) /* sanity limit on qsize */
		return -EOVERFLOW;

	/* Make sure CPU is a valid possible cpu */
	if (key_cpu >= nr_cpumask_bits || !cpu_possible(key_cpu))
		return -ENODEV;

	if (cpumap_value.qsize == 0) {
		rcpu = NULL; /* Same as deleting */
	} else {
		/* Updating qsize cause re-allocation of bpf_cpu_map_entry */
		rcpu = __cpu_map_entry_alloc(map, &cpumap_value, key_cpu);
		if (!rcpu)
			return -ENOMEM;
	}
	rcu_read_lock();
	__cpu_map_entry_replace(cmap, key_cpu, rcpu);
	rcu_read_unlock();
	return 0;
}

static void cpu_map_free(struct bpf_map *map)
{
	struct bpf_cpu_map *cmap = container_of(map, struct bpf_cpu_map, map);
	u32 i;

	/* At this point bpf_prog->aux->refcnt == 0 and this map->refcnt == 0,
	 * so the bpf programs (can be more than one that used this map) were
	 * disconnected from events. Wait for outstanding critical sections in
	 * these programs to complete. synchronize_rcu() below not only
	 * guarantees no further "XDP/bpf-side" reads against
	 * bpf_cpu_map->cpu_map, but also ensure pending flush operations
	 * (if any) are completed.
	 */
	synchronize_rcu();

	/* The only possible user of bpf_cpu_map_entry is
	 * cpu_map_kthread_run().
	 */
	for (i = 0; i < cmap->map.max_entries; i++) {
		struct bpf_cpu_map_entry *rcpu;

		rcpu = rcu_dereference_raw(cmap->cpu_map[i]);
		if (!rcpu)
			continue;

		/* Stop kthread and cleanup entry directly */
		__cpu_map_entry_free(&rcpu->free_work.work);
	}
	bpf_map_area_free(cmap->cpu_map);
	bpf_map_area_free(cmap);
}

/* Elements are kept alive by RCU; either by rcu_read_lock() (from syscall) or
 * by local_bh_disable() (from XDP calls inside NAPI). The
 * rcu_read_lock_bh_held() below makes lockdep accept both.
 */
static void *__cpu_map_lookup_elem(struct bpf_map *map, u32 key)
{
	struct bpf_cpu_map *cmap = container_of(map, struct bpf_cpu_map, map);
	struct bpf_cpu_map_entry *rcpu;

	if (key >= map->max_entries)
		return NULL;

	rcpu = rcu_dereference_check(cmap->cpu_map[key],
				     rcu_read_lock_bh_held());
	return rcpu;
}

static void *cpu_map_lookup_elem(struct bpf_map *map, void *key)
{
	struct bpf_cpu_map_entry *rcpu =
		__cpu_map_lookup_elem(map, *(u32 *)key);

	return rcpu ? &rcpu->value : NULL;
}

static int cpu_map_get_next_key(struct bpf_map *map, void *key, void *next_key)
{
	struct bpf_cpu_map *cmap = container_of(map, struct bpf_cpu_map, map);
	u32 index = key ? *(u32 *)key : U32_MAX;
	u32 *next = next_key;

	if (index >= cmap->map.max_entries) {
		*next = 0;
		return 0;
	}

	if (index == cmap->map.max_entries - 1)
		return -ENOENT;
	*next = index + 1;
	return 0;
}

static long cpu_map_redirect(struct bpf_map *map, u64 index, u64 flags)
{
	return __bpf_xdp_redirect_map(map, index, flags, 0,
				      __cpu_map_lookup_elem);
}

static u64 cpu_map_mem_usage(const struct bpf_map *map)
{
	u64 usage = sizeof(struct bpf_cpu_map);

	/* Currently the dynamically allocated elements are not counted */
	usage += (u64)map->max_entries * sizeof(struct bpf_cpu_map_entry *);
	return usage;
}

BTF_ID_LIST_SINGLE(cpu_map_btf_ids, struct, bpf_cpu_map)
const struct bpf_map_ops cpu_map_ops = {
	.map_meta_equal		= bpf_map_meta_equal,
	.map_alloc		= cpu_map_alloc,
	.map_free		= cpu_map_free,
	.map_delete_elem	= cpu_map_delete_elem,
	.map_update_elem	= cpu_map_update_elem,
	.map_lookup_elem	= cpu_map_lookup_elem,
	.map_get_next_key	= cpu_map_get_next_key,
	.map_check_btf		= map_check_no_btf,
	.map_mem_usage		= cpu_map_mem_usage,
	.map_btf_id		= &cpu_map_btf_ids[0],
	.map_redirect		= cpu_map_redirect,
};

static void bq_flush_to_queue(struct xdp_bulk_queue *bq)
{
	struct bpf_cpu_map_entry *rcpu = bq->obj;
	unsigned int processed = 0, drops = 0;
	const int to_cpu = rcpu->cpu;
	struct ptr_ring *q;
	int i;

	if (unlikely(!bq->count))
		return;

	q = rcpu->queue;
	spin_lock(&q->producer_lock);

	for (i = 0; i < bq->count; i++) {
		struct xdp_frame *xdpf = bq->q[i];
		int err;

		err = __ptr_ring_produce(q, xdpf);
		if (err) {
			drops++;
			xdp_return_frame_rx_napi(xdpf);
		}
		processed++;
	}
	bq->count = 0;
	spin_unlock(&q->producer_lock);

	__list_del_clearprev(&bq->flush_node);

	/* Feedback loop via tracepoints */
	trace_xdp_cpumap_enqueue(rcpu->map_id, processed, drops, to_cpu);
}

/* Runs under RCU-read-side, plus in softirq under NAPI protection.
 * Thus, safe percpu variable access.
 */
static void bq_enqueue(struct bpf_cpu_map_entry *rcpu, struct xdp_frame *xdpf)
{
	struct xdp_bulk_queue *bq = this_cpu_ptr(rcpu->bulkq);

	if (unlikely(bq->count == CPU_MAP_BULK_SIZE))
		bq_flush_to_queue(bq);

	/* Notice, xdp_buff/page MUST be queued here, long enough for
	 * driver to code invoking us to finished, due to driver
	 * (e.g. ixgbe) recycle tricks based on page-refcnt.
	 *
	 * Thus, incoming xdp_frame is always queued here (else we race
	 * with another CPU on page-refcnt and remaining driver code).
	 * Queue time is very short, as driver will invoke flush
	 * operation, when completing napi->poll call.
	 */
	bq->q[bq->count++] = xdpf;

	if (!bq->flush_node.prev) {
		struct list_head *flush_list = bpf_net_ctx_get_cpu_map_flush_list();

		list_add(&bq->flush_node, flush_list);
	}
}

int cpu_map_enqueue(struct bpf_cpu_map_entry *rcpu, struct xdp_frame *xdpf,
		    struct net_device *dev_rx)
{
	/* Info needed when constructing SKB on remote CPU */
	xdpf->dev_rx = dev_rx;

	bq_enqueue(rcpu, xdpf);
	return 0;
}

int cpu_map_generic_redirect(struct bpf_cpu_map_entry *rcpu,
			     struct sk_buff *skb)
{
	int ret;

	__skb_pull(skb, skb->mac_len);
	skb_set_redirected(skb, false);
	__ptr_set_bit(0, &skb);

	ret = ptr_ring_produce(rcpu->queue, skb);
	if (ret < 0)
		goto trace;

	wake_up_process(rcpu->kthread);
trace:
	trace_xdp_cpumap_enqueue(rcpu->map_id, !ret, !!ret, rcpu->cpu);
	return ret;
}

void __cpu_map_flush(struct list_head *flush_list)
{
	struct xdp_bulk_queue *bq, *tmp;

	list_for_each_entry_safe(bq, tmp, flush_list, flush_node) {
		bq_flush_to_queue(bq);

		/* If already running, costs spin_lock_irqsave + smb_mb */
		wake_up_process(bq->obj->kthread);
	}
}