Loading...
Note: File does not exist in v3.5.6.
1// SPDX-License-Identifier: GPL-2.0-only
2/* bpf/cpumap.c
3 *
4 * Copyright (c) 2017 Jesper Dangaard Brouer, Red Hat Inc.
5 */
6
7/**
8 * DOC: cpu map
9 * The 'cpumap' is primarily used as a backend map for XDP BPF helper
10 * call bpf_redirect_map() and XDP_REDIRECT action, like 'devmap'.
11 *
12 * Unlike devmap which redirects XDP frames out to another NIC device,
13 * this map type redirects raw XDP frames to another CPU. The remote
14 * CPU will do SKB-allocation and call the normal network stack.
15 */
16/*
17 * This is a scalability and isolation mechanism, that allow
18 * separating the early driver network XDP layer, from the rest of the
19 * netstack, and assigning dedicated CPUs for this stage. This
20 * basically allows for 10G wirespeed pre-filtering via bpf.
21 */
22#include <linux/bitops.h>
23#include <linux/bpf.h>
24#include <linux/filter.h>
25#include <linux/ptr_ring.h>
26#include <net/xdp.h>
27#include <net/hotdata.h>
28
29#include <linux/sched.h>
30#include <linux/workqueue.h>
31#include <linux/kthread.h>
32#include <linux/completion.h>
33#include <trace/events/xdp.h>
34#include <linux/btf_ids.h>
35
36#include <linux/netdevice.h> /* netif_receive_skb_list */
37#include <linux/etherdevice.h> /* eth_type_trans */
38
39/* General idea: XDP packets getting XDP redirected to another CPU,
40 * will maximum be stored/queued for one driver ->poll() call. It is
41 * guaranteed that queueing the frame and the flush operation happen on
42 * same CPU. Thus, cpu_map_flush operation can deduct via this_cpu_ptr()
43 * which queue in bpf_cpu_map_entry contains packets.
44 */
45
46#define CPU_MAP_BULK_SIZE 8 /* 8 == one cacheline on 64-bit archs */
47struct bpf_cpu_map_entry;
48struct bpf_cpu_map;
49
50struct xdp_bulk_queue {
51 void *q[CPU_MAP_BULK_SIZE];
52 struct list_head flush_node;
53 struct bpf_cpu_map_entry *obj;
54 unsigned int count;
55};
56
57/* Struct for every remote "destination" CPU in map */
58struct bpf_cpu_map_entry {
59 u32 cpu; /* kthread CPU and map index */
60 int map_id; /* Back reference to map */
61
62 /* XDP can run multiple RX-ring queues, need __percpu enqueue store */
63 struct xdp_bulk_queue __percpu *bulkq;
64
65 /* Queue with potential multi-producers, and single-consumer kthread */
66 struct ptr_ring *queue;
67 struct task_struct *kthread;
68
69 struct bpf_cpumap_val value;
70 struct bpf_prog *prog;
71
72 struct completion kthread_running;
73 struct rcu_work free_work;
74};
75
76struct bpf_cpu_map {
77 struct bpf_map map;
78 /* Below members specific for map type */
79 struct bpf_cpu_map_entry __rcu **cpu_map;
80};
81
82static struct bpf_map *cpu_map_alloc(union bpf_attr *attr)
83{
84 u32 value_size = attr->value_size;
85 struct bpf_cpu_map *cmap;
86
87 /* check sanity of attributes */
88 if (attr->max_entries == 0 || attr->key_size != 4 ||
89 (value_size != offsetofend(struct bpf_cpumap_val, qsize) &&
90 value_size != offsetofend(struct bpf_cpumap_val, bpf_prog.fd)) ||
91 attr->map_flags & ~BPF_F_NUMA_NODE)
92 return ERR_PTR(-EINVAL);
93
94 /* Pre-limit array size based on NR_CPUS, not final CPU check */
95 if (attr->max_entries > NR_CPUS)
96 return ERR_PTR(-E2BIG);
97
98 cmap = bpf_map_area_alloc(sizeof(*cmap), NUMA_NO_NODE);
99 if (!cmap)
100 return ERR_PTR(-ENOMEM);
101
102 bpf_map_init_from_attr(&cmap->map, attr);
103
104 /* Alloc array for possible remote "destination" CPUs */
105 cmap->cpu_map = bpf_map_area_alloc(cmap->map.max_entries *
106 sizeof(struct bpf_cpu_map_entry *),
107 cmap->map.numa_node);
108 if (!cmap->cpu_map) {
109 bpf_map_area_free(cmap);
110 return ERR_PTR(-ENOMEM);
111 }
112
113 return &cmap->map;
114}
115
116static void __cpu_map_ring_cleanup(struct ptr_ring *ring)
117{
118 /* The tear-down procedure should have made sure that queue is
119 * empty. See __cpu_map_entry_replace() and work-queue
120 * invoked cpu_map_kthread_stop(). Catch any broken behaviour
121 * gracefully and warn once.
122 */
123 void *ptr;
124
125 while ((ptr = ptr_ring_consume(ring))) {
126 WARN_ON_ONCE(1);
127 if (unlikely(__ptr_test_bit(0, &ptr))) {
128 __ptr_clear_bit(0, &ptr);
129 kfree_skb(ptr);
130 continue;
131 }
132 xdp_return_frame(ptr);
133 }
134}
135
136static void cpu_map_bpf_prog_run_skb(struct bpf_cpu_map_entry *rcpu,
137 struct list_head *listp,
138 struct xdp_cpumap_stats *stats)
139{
140 struct sk_buff *skb, *tmp;
141 struct xdp_buff xdp;
142 u32 act;
143 int err;
144
145 list_for_each_entry_safe(skb, tmp, listp, list) {
146 act = bpf_prog_run_generic_xdp(skb, &xdp, rcpu->prog);
147 switch (act) {
148 case XDP_PASS:
149 break;
150 case XDP_REDIRECT:
151 skb_list_del_init(skb);
152 err = xdp_do_generic_redirect(skb->dev, skb, &xdp,
153 rcpu->prog);
154 if (unlikely(err)) {
155 kfree_skb(skb);
156 stats->drop++;
157 } else {
158 stats->redirect++;
159 }
160 return;
161 default:
162 bpf_warn_invalid_xdp_action(NULL, rcpu->prog, act);
163 fallthrough;
164 case XDP_ABORTED:
165 trace_xdp_exception(skb->dev, rcpu->prog, act);
166 fallthrough;
167 case XDP_DROP:
168 skb_list_del_init(skb);
169 kfree_skb(skb);
170 stats->drop++;
171 return;
172 }
173 }
174}
175
176static int cpu_map_bpf_prog_run_xdp(struct bpf_cpu_map_entry *rcpu,
177 void **frames, int n,
178 struct xdp_cpumap_stats *stats)
179{
180 struct xdp_rxq_info rxq = {};
181 struct xdp_buff xdp;
182 int i, nframes = 0;
183
184 xdp_set_return_frame_no_direct();
185 xdp.rxq = &rxq;
186
187 for (i = 0; i < n; i++) {
188 struct xdp_frame *xdpf = frames[i];
189 u32 act;
190 int err;
191
192 rxq.dev = xdpf->dev_rx;
193 rxq.mem = xdpf->mem;
194 /* TODO: report queue_index to xdp_rxq_info */
195
196 xdp_convert_frame_to_buff(xdpf, &xdp);
197
198 act = bpf_prog_run_xdp(rcpu->prog, &xdp);
199 switch (act) {
200 case XDP_PASS:
201 err = xdp_update_frame_from_buff(&xdp, xdpf);
202 if (err < 0) {
203 xdp_return_frame(xdpf);
204 stats->drop++;
205 } else {
206 frames[nframes++] = xdpf;
207 stats->pass++;
208 }
209 break;
210 case XDP_REDIRECT:
211 err = xdp_do_redirect(xdpf->dev_rx, &xdp,
212 rcpu->prog);
213 if (unlikely(err)) {
214 xdp_return_frame(xdpf);
215 stats->drop++;
216 } else {
217 stats->redirect++;
218 }
219 break;
220 default:
221 bpf_warn_invalid_xdp_action(NULL, rcpu->prog, act);
222 fallthrough;
223 case XDP_DROP:
224 xdp_return_frame(xdpf);
225 stats->drop++;
226 break;
227 }
228 }
229
230 xdp_clear_return_frame_no_direct();
231
232 return nframes;
233}
234
235#define CPUMAP_BATCH 8
236
237static int cpu_map_bpf_prog_run(struct bpf_cpu_map_entry *rcpu, void **frames,
238 int xdp_n, struct xdp_cpumap_stats *stats,
239 struct list_head *list)
240{
241 struct bpf_net_context __bpf_net_ctx, *bpf_net_ctx;
242 int nframes;
243
244 if (!rcpu->prog)
245 return xdp_n;
246
247 rcu_read_lock_bh();
248 bpf_net_ctx = bpf_net_ctx_set(&__bpf_net_ctx);
249
250 nframes = cpu_map_bpf_prog_run_xdp(rcpu, frames, xdp_n, stats);
251
252 if (stats->redirect)
253 xdp_do_flush();
254
255 if (unlikely(!list_empty(list)))
256 cpu_map_bpf_prog_run_skb(rcpu, list, stats);
257
258 bpf_net_ctx_clear(bpf_net_ctx);
259 rcu_read_unlock_bh(); /* resched point, may call do_softirq() */
260
261 return nframes;
262}
263
264static int cpu_map_kthread_run(void *data)
265{
266 struct bpf_cpu_map_entry *rcpu = data;
267 unsigned long last_qs = jiffies;
268
269 complete(&rcpu->kthread_running);
270 set_current_state(TASK_INTERRUPTIBLE);
271
272 /* When kthread gives stop order, then rcpu have been disconnected
273 * from map, thus no new packets can enter. Remaining in-flight
274 * per CPU stored packets are flushed to this queue. Wait honoring
275 * kthread_stop signal until queue is empty.
276 */
277 while (!kthread_should_stop() || !__ptr_ring_empty(rcpu->queue)) {
278 struct xdp_cpumap_stats stats = {}; /* zero stats */
279 unsigned int kmem_alloc_drops = 0, sched = 0;
280 gfp_t gfp = __GFP_ZERO | GFP_ATOMIC;
281 int i, n, m, nframes, xdp_n;
282 void *frames[CPUMAP_BATCH];
283 void *skbs[CPUMAP_BATCH];
284 LIST_HEAD(list);
285
286 /* Release CPU reschedule checks */
287 if (__ptr_ring_empty(rcpu->queue)) {
288 set_current_state(TASK_INTERRUPTIBLE);
289 /* Recheck to avoid lost wake-up */
290 if (__ptr_ring_empty(rcpu->queue)) {
291 schedule();
292 sched = 1;
293 last_qs = jiffies;
294 } else {
295 __set_current_state(TASK_RUNNING);
296 }
297 } else {
298 rcu_softirq_qs_periodic(last_qs);
299 sched = cond_resched();
300 }
301
302 /*
303 * The bpf_cpu_map_entry is single consumer, with this
304 * kthread CPU pinned. Lockless access to ptr_ring
305 * consume side valid as no-resize allowed of queue.
306 */
307 n = __ptr_ring_consume_batched(rcpu->queue, frames,
308 CPUMAP_BATCH);
309 for (i = 0, xdp_n = 0; i < n; i++) {
310 void *f = frames[i];
311 struct page *page;
312
313 if (unlikely(__ptr_test_bit(0, &f))) {
314 struct sk_buff *skb = f;
315
316 __ptr_clear_bit(0, &skb);
317 list_add_tail(&skb->list, &list);
318 continue;
319 }
320
321 frames[xdp_n++] = f;
322 page = virt_to_page(f);
323
324 /* Bring struct page memory area to curr CPU. Read by
325 * build_skb_around via page_is_pfmemalloc(), and when
326 * freed written by page_frag_free call.
327 */
328 prefetchw(page);
329 }
330
331 /* Support running another XDP prog on this CPU */
332 nframes = cpu_map_bpf_prog_run(rcpu, frames, xdp_n, &stats, &list);
333 if (nframes) {
334 m = kmem_cache_alloc_bulk(net_hotdata.skbuff_cache,
335 gfp, nframes, skbs);
336 if (unlikely(m == 0)) {
337 for (i = 0; i < nframes; i++)
338 skbs[i] = NULL; /* effect: xdp_return_frame */
339 kmem_alloc_drops += nframes;
340 }
341 }
342
343 local_bh_disable();
344 for (i = 0; i < nframes; i++) {
345 struct xdp_frame *xdpf = frames[i];
346 struct sk_buff *skb = skbs[i];
347
348 skb = __xdp_build_skb_from_frame(xdpf, skb,
349 xdpf->dev_rx);
350 if (!skb) {
351 xdp_return_frame(xdpf);
352 continue;
353 }
354
355 list_add_tail(&skb->list, &list);
356 }
357
358 /* Feedback loop via tracepoint.
359 * NB: keep before recv to allow measuring enqueue/dequeue latency.
360 */
361 trace_xdp_cpumap_kthread(rcpu->map_id, n, kmem_alloc_drops,
362 sched, &stats);
363
364 netif_receive_skb_list(&list);
365 local_bh_enable(); /* resched point, may call do_softirq() */
366 }
367 __set_current_state(TASK_RUNNING);
368
369 return 0;
370}
371
372static int __cpu_map_load_bpf_program(struct bpf_cpu_map_entry *rcpu,
373 struct bpf_map *map, int fd)
374{
375 struct bpf_prog *prog;
376
377 prog = bpf_prog_get_type(fd, BPF_PROG_TYPE_XDP);
378 if (IS_ERR(prog))
379 return PTR_ERR(prog);
380
381 if (prog->expected_attach_type != BPF_XDP_CPUMAP ||
382 !bpf_prog_map_compatible(map, prog)) {
383 bpf_prog_put(prog);
384 return -EINVAL;
385 }
386
387 rcpu->value.bpf_prog.id = prog->aux->id;
388 rcpu->prog = prog;
389
390 return 0;
391}
392
393static struct bpf_cpu_map_entry *
394__cpu_map_entry_alloc(struct bpf_map *map, struct bpf_cpumap_val *value,
395 u32 cpu)
396{
397 int numa, err, i, fd = value->bpf_prog.fd;
398 gfp_t gfp = GFP_KERNEL | __GFP_NOWARN;
399 struct bpf_cpu_map_entry *rcpu;
400 struct xdp_bulk_queue *bq;
401
402 /* Have map->numa_node, but choose node of redirect target CPU */
403 numa = cpu_to_node(cpu);
404
405 rcpu = bpf_map_kmalloc_node(map, sizeof(*rcpu), gfp | __GFP_ZERO, numa);
406 if (!rcpu)
407 return NULL;
408
409 /* Alloc percpu bulkq */
410 rcpu->bulkq = bpf_map_alloc_percpu(map, sizeof(*rcpu->bulkq),
411 sizeof(void *), gfp);
412 if (!rcpu->bulkq)
413 goto free_rcu;
414
415 for_each_possible_cpu(i) {
416 bq = per_cpu_ptr(rcpu->bulkq, i);
417 bq->obj = rcpu;
418 }
419
420 /* Alloc queue */
421 rcpu->queue = bpf_map_kmalloc_node(map, sizeof(*rcpu->queue), gfp,
422 numa);
423 if (!rcpu->queue)
424 goto free_bulkq;
425
426 err = ptr_ring_init(rcpu->queue, value->qsize, gfp);
427 if (err)
428 goto free_queue;
429
430 rcpu->cpu = cpu;
431 rcpu->map_id = map->id;
432 rcpu->value.qsize = value->qsize;
433
434 if (fd > 0 && __cpu_map_load_bpf_program(rcpu, map, fd))
435 goto free_ptr_ring;
436
437 /* Setup kthread */
438 init_completion(&rcpu->kthread_running);
439 rcpu->kthread = kthread_create_on_node(cpu_map_kthread_run, rcpu, numa,
440 "cpumap/%d/map:%d", cpu,
441 map->id);
442 if (IS_ERR(rcpu->kthread))
443 goto free_prog;
444
445 /* Make sure kthread runs on a single CPU */
446 kthread_bind(rcpu->kthread, cpu);
447 wake_up_process(rcpu->kthread);
448
449 /* Make sure kthread has been running, so kthread_stop() will not
450 * stop the kthread prematurely and all pending frames or skbs
451 * will be handled by the kthread before kthread_stop() returns.
452 */
453 wait_for_completion(&rcpu->kthread_running);
454
455 return rcpu;
456
457free_prog:
458 if (rcpu->prog)
459 bpf_prog_put(rcpu->prog);
460free_ptr_ring:
461 ptr_ring_cleanup(rcpu->queue, NULL);
462free_queue:
463 kfree(rcpu->queue);
464free_bulkq:
465 free_percpu(rcpu->bulkq);
466free_rcu:
467 kfree(rcpu);
468 return NULL;
469}
470
471static void __cpu_map_entry_free(struct work_struct *work)
472{
473 struct bpf_cpu_map_entry *rcpu;
474
475 /* This cpu_map_entry have been disconnected from map and one
476 * RCU grace-period have elapsed. Thus, XDP cannot queue any
477 * new packets and cannot change/set flush_needed that can
478 * find this entry.
479 */
480 rcpu = container_of(to_rcu_work(work), struct bpf_cpu_map_entry, free_work);
481
482 /* kthread_stop will wake_up_process and wait for it to complete.
483 * cpu_map_kthread_run() makes sure the pointer ring is empty
484 * before exiting.
485 */
486 kthread_stop(rcpu->kthread);
487
488 if (rcpu->prog)
489 bpf_prog_put(rcpu->prog);
490 /* The queue should be empty at this point */
491 __cpu_map_ring_cleanup(rcpu->queue);
492 ptr_ring_cleanup(rcpu->queue, NULL);
493 kfree(rcpu->queue);
494 free_percpu(rcpu->bulkq);
495 kfree(rcpu);
496}
497
498/* After the xchg of the bpf_cpu_map_entry pointer, we need to make sure the old
499 * entry is no longer in use before freeing. We use queue_rcu_work() to call
500 * __cpu_map_entry_free() in a separate workqueue after waiting for an RCU grace
501 * period. This means that (a) all pending enqueue and flush operations have
502 * completed (because of the RCU callback), and (b) we are in a workqueue
503 * context where we can stop the kthread and wait for it to exit before freeing
504 * everything.
505 */
506static void __cpu_map_entry_replace(struct bpf_cpu_map *cmap,
507 u32 key_cpu, struct bpf_cpu_map_entry *rcpu)
508{
509 struct bpf_cpu_map_entry *old_rcpu;
510
511 old_rcpu = unrcu_pointer(xchg(&cmap->cpu_map[key_cpu], RCU_INITIALIZER(rcpu)));
512 if (old_rcpu) {
513 INIT_RCU_WORK(&old_rcpu->free_work, __cpu_map_entry_free);
514 queue_rcu_work(system_wq, &old_rcpu->free_work);
515 }
516}
517
518static long cpu_map_delete_elem(struct bpf_map *map, void *key)
519{
520 struct bpf_cpu_map *cmap = container_of(map, struct bpf_cpu_map, map);
521 u32 key_cpu = *(u32 *)key;
522
523 if (key_cpu >= map->max_entries)
524 return -EINVAL;
525
526 /* notice caller map_delete_elem() uses rcu_read_lock() */
527 __cpu_map_entry_replace(cmap, key_cpu, NULL);
528 return 0;
529}
530
531static long cpu_map_update_elem(struct bpf_map *map, void *key, void *value,
532 u64 map_flags)
533{
534 struct bpf_cpu_map *cmap = container_of(map, struct bpf_cpu_map, map);
535 struct bpf_cpumap_val cpumap_value = {};
536 struct bpf_cpu_map_entry *rcpu;
537 /* Array index key correspond to CPU number */
538 u32 key_cpu = *(u32 *)key;
539
540 memcpy(&cpumap_value, value, map->value_size);
541
542 if (unlikely(map_flags > BPF_EXIST))
543 return -EINVAL;
544 if (unlikely(key_cpu >= cmap->map.max_entries))
545 return -E2BIG;
546 if (unlikely(map_flags == BPF_NOEXIST))
547 return -EEXIST;
548 if (unlikely(cpumap_value.qsize > 16384)) /* sanity limit on qsize */
549 return -EOVERFLOW;
550
551 /* Make sure CPU is a valid possible cpu */
552 if (key_cpu >= nr_cpumask_bits || !cpu_possible(key_cpu))
553 return -ENODEV;
554
555 if (cpumap_value.qsize == 0) {
556 rcpu = NULL; /* Same as deleting */
557 } else {
558 /* Updating qsize cause re-allocation of bpf_cpu_map_entry */
559 rcpu = __cpu_map_entry_alloc(map, &cpumap_value, key_cpu);
560 if (!rcpu)
561 return -ENOMEM;
562 }
563 rcu_read_lock();
564 __cpu_map_entry_replace(cmap, key_cpu, rcpu);
565 rcu_read_unlock();
566 return 0;
567}
568
569static void cpu_map_free(struct bpf_map *map)
570{
571 struct bpf_cpu_map *cmap = container_of(map, struct bpf_cpu_map, map);
572 u32 i;
573
574 /* At this point bpf_prog->aux->refcnt == 0 and this map->refcnt == 0,
575 * so the bpf programs (can be more than one that used this map) were
576 * disconnected from events. Wait for outstanding critical sections in
577 * these programs to complete. synchronize_rcu() below not only
578 * guarantees no further "XDP/bpf-side" reads against
579 * bpf_cpu_map->cpu_map, but also ensure pending flush operations
580 * (if any) are completed.
581 */
582 synchronize_rcu();
583
584 /* The only possible user of bpf_cpu_map_entry is
585 * cpu_map_kthread_run().
586 */
587 for (i = 0; i < cmap->map.max_entries; i++) {
588 struct bpf_cpu_map_entry *rcpu;
589
590 rcpu = rcu_dereference_raw(cmap->cpu_map[i]);
591 if (!rcpu)
592 continue;
593
594 /* Stop kthread and cleanup entry directly */
595 __cpu_map_entry_free(&rcpu->free_work.work);
596 }
597 bpf_map_area_free(cmap->cpu_map);
598 bpf_map_area_free(cmap);
599}
600
601/* Elements are kept alive by RCU; either by rcu_read_lock() (from syscall) or
602 * by local_bh_disable() (from XDP calls inside NAPI). The
603 * rcu_read_lock_bh_held() below makes lockdep accept both.
604 */
605static void *__cpu_map_lookup_elem(struct bpf_map *map, u32 key)
606{
607 struct bpf_cpu_map *cmap = container_of(map, struct bpf_cpu_map, map);
608 struct bpf_cpu_map_entry *rcpu;
609
610 if (key >= map->max_entries)
611 return NULL;
612
613 rcpu = rcu_dereference_check(cmap->cpu_map[key],
614 rcu_read_lock_bh_held());
615 return rcpu;
616}
617
618static void *cpu_map_lookup_elem(struct bpf_map *map, void *key)
619{
620 struct bpf_cpu_map_entry *rcpu =
621 __cpu_map_lookup_elem(map, *(u32 *)key);
622
623 return rcpu ? &rcpu->value : NULL;
624}
625
626static int cpu_map_get_next_key(struct bpf_map *map, void *key, void *next_key)
627{
628 struct bpf_cpu_map *cmap = container_of(map, struct bpf_cpu_map, map);
629 u32 index = key ? *(u32 *)key : U32_MAX;
630 u32 *next = next_key;
631
632 if (index >= cmap->map.max_entries) {
633 *next = 0;
634 return 0;
635 }
636
637 if (index == cmap->map.max_entries - 1)
638 return -ENOENT;
639 *next = index + 1;
640 return 0;
641}
642
643static long cpu_map_redirect(struct bpf_map *map, u64 index, u64 flags)
644{
645 return __bpf_xdp_redirect_map(map, index, flags, 0,
646 __cpu_map_lookup_elem);
647}
648
649static u64 cpu_map_mem_usage(const struct bpf_map *map)
650{
651 u64 usage = sizeof(struct bpf_cpu_map);
652
653 /* Currently the dynamically allocated elements are not counted */
654 usage += (u64)map->max_entries * sizeof(struct bpf_cpu_map_entry *);
655 return usage;
656}
657
658BTF_ID_LIST_SINGLE(cpu_map_btf_ids, struct, bpf_cpu_map)
659const struct bpf_map_ops cpu_map_ops = {
660 .map_meta_equal = bpf_map_meta_equal,
661 .map_alloc = cpu_map_alloc,
662 .map_free = cpu_map_free,
663 .map_delete_elem = cpu_map_delete_elem,
664 .map_update_elem = cpu_map_update_elem,
665 .map_lookup_elem = cpu_map_lookup_elem,
666 .map_get_next_key = cpu_map_get_next_key,
667 .map_check_btf = map_check_no_btf,
668 .map_mem_usage = cpu_map_mem_usage,
669 .map_btf_id = &cpu_map_btf_ids[0],
670 .map_redirect = cpu_map_redirect,
671};
672
673static void bq_flush_to_queue(struct xdp_bulk_queue *bq)
674{
675 struct bpf_cpu_map_entry *rcpu = bq->obj;
676 unsigned int processed = 0, drops = 0;
677 const int to_cpu = rcpu->cpu;
678 struct ptr_ring *q;
679 int i;
680
681 if (unlikely(!bq->count))
682 return;
683
684 q = rcpu->queue;
685 spin_lock(&q->producer_lock);
686
687 for (i = 0; i < bq->count; i++) {
688 struct xdp_frame *xdpf = bq->q[i];
689 int err;
690
691 err = __ptr_ring_produce(q, xdpf);
692 if (err) {
693 drops++;
694 xdp_return_frame_rx_napi(xdpf);
695 }
696 processed++;
697 }
698 bq->count = 0;
699 spin_unlock(&q->producer_lock);
700
701 __list_del_clearprev(&bq->flush_node);
702
703 /* Feedback loop via tracepoints */
704 trace_xdp_cpumap_enqueue(rcpu->map_id, processed, drops, to_cpu);
705}
706
707/* Runs under RCU-read-side, plus in softirq under NAPI protection.
708 * Thus, safe percpu variable access.
709 */
710static void bq_enqueue(struct bpf_cpu_map_entry *rcpu, struct xdp_frame *xdpf)
711{
712 struct xdp_bulk_queue *bq = this_cpu_ptr(rcpu->bulkq);
713
714 if (unlikely(bq->count == CPU_MAP_BULK_SIZE))
715 bq_flush_to_queue(bq);
716
717 /* Notice, xdp_buff/page MUST be queued here, long enough for
718 * driver to code invoking us to finished, due to driver
719 * (e.g. ixgbe) recycle tricks based on page-refcnt.
720 *
721 * Thus, incoming xdp_frame is always queued here (else we race
722 * with another CPU on page-refcnt and remaining driver code).
723 * Queue time is very short, as driver will invoke flush
724 * operation, when completing napi->poll call.
725 */
726 bq->q[bq->count++] = xdpf;
727
728 if (!bq->flush_node.prev) {
729 struct list_head *flush_list = bpf_net_ctx_get_cpu_map_flush_list();
730
731 list_add(&bq->flush_node, flush_list);
732 }
733}
734
735int cpu_map_enqueue(struct bpf_cpu_map_entry *rcpu, struct xdp_frame *xdpf,
736 struct net_device *dev_rx)
737{
738 /* Info needed when constructing SKB on remote CPU */
739 xdpf->dev_rx = dev_rx;
740
741 bq_enqueue(rcpu, xdpf);
742 return 0;
743}
744
745int cpu_map_generic_redirect(struct bpf_cpu_map_entry *rcpu,
746 struct sk_buff *skb)
747{
748 int ret;
749
750 __skb_pull(skb, skb->mac_len);
751 skb_set_redirected(skb, false);
752 __ptr_set_bit(0, &skb);
753
754 ret = ptr_ring_produce(rcpu->queue, skb);
755 if (ret < 0)
756 goto trace;
757
758 wake_up_process(rcpu->kthread);
759trace:
760 trace_xdp_cpumap_enqueue(rcpu->map_id, !ret, !!ret, rcpu->cpu);
761 return ret;
762}
763
764void __cpu_map_flush(struct list_head *flush_list)
765{
766 struct xdp_bulk_queue *bq, *tmp;
767
768 list_for_each_entry_safe(bq, tmp, flush_list, flush_node) {
769 bq_flush_to_queue(bq);
770
771 /* If already running, costs spin_lock_irqsave + smb_mb */
772 wake_up_process(bq->obj->kthread);
773 }
774}