bpf: Refactor BPF_PROG_RUN into a function

Turn BPF_PROG_RUN into a proper always inlined function. No functional and
performance changes are intended, but it makes it much easier to understand
what's going on with how BPF programs are actually get executed. It's more
obvious what types and callbacks are expected. Also extra () around input
parameters can be dropped, as well as `__` variable prefixes intended to avoid
naming collisions, which makes the code simpler to read and write.

This refactoring also highlighted one extra issue. BPF_PROG_RUN is both
a macro and an enum value (BPF_PROG_RUN == BPF_PROG_TEST_RUN). Turning
BPF_PROG_RUN into a function causes naming conflict compilation error. So
rename BPF_PROG_RUN into lower-case bpf_prog_run(), similar to
bpf_prog_run_xdp(), bpf_prog_run_pin_on_cpu(), etc. All existing callers of
BPF_PROG_RUN, the macro, are switched to bpf_prog_run() explicitly.

Signed-off-by: Andrii Nakryiko <[e-mail]>
Signed-off-by: Daniel Borkmann <[e-mail]>
Acked-by: Yonghong Song <[e-mail]>
Link: https://lore.kernel.org/bpf/[e-mail]

Documentation/networking/filter.rst

···
 converted by the kernel into the new instruction set representation and run
 in the eBPF interpreter. For in-kernel handlers, this all works transparently
 by using bpf_prog_create() for setting up the filter, resp.
-bpf_prog_destroy() for destroying it. The macro
-BPF_PROG_RUN(filter, ctx) transparently invokes eBPF interpreter or JITed
+bpf_prog_destroy() for destroying it. The function
+bpf_prog_run(filter, ctx) transparently invokes eBPF interpreter or JITed
 code to run the filter. 'filter' is a pointer to struct bpf_prog that we
 got from bpf_prog_create(), and 'ctx' the given context (e.g.
 skb pointer). All constraints and restrictions from bpf_check_classic() apply

drivers/media/rc/bpf-lirc.c

···
 	raw->bpf_sample = sample;
 
 	if (raw->progs)
-		BPF_PROG_RUN_ARRAY(raw->progs, &raw->bpf_sample, BPF_PROG_RUN);
+		BPF_PROG_RUN_ARRAY(raw->progs, &raw->bpf_sample, bpf_prog_run);
 }
 
 /*

drivers/net/ppp/ppp_generic.c

···
 		   a four-byte PPP header on each packet */
 		*(u8 *)skb_push(skb, 2) = 1;
 		if (ppp->pass_filter &&
-		    BPF_PROG_RUN(ppp->pass_filter, skb) == 0) {
+		    bpf_prog_run(ppp->pass_filter, skb) == 0) {
 			if (ppp->debug & 1)
 				netdev_printk(KERN_DEBUG, ppp->dev,
 					      "PPP: outbound frame "
···
 		}
 		/* if this packet passes the active filter, record the time */
 		if (!(ppp->active_filter &&
-		      BPF_PROG_RUN(ppp->active_filter, skb) == 0))
+		      bpf_prog_run(ppp->active_filter, skb) == 0))
 			ppp->last_xmit = jiffies;
 		skb_pull(skb, 2);
 #else
···
 
 			*(u8 *)skb_push(skb, 2) = 0;
 			if (ppp->pass_filter &&
-			    BPF_PROG_RUN(ppp->pass_filter, skb) == 0) {
+			    bpf_prog_run(ppp->pass_filter, skb) == 0) {
 				if (ppp->debug & 1)
 					netdev_printk(KERN_DEBUG, ppp->dev,
 						      "PPP: inbound frame "
···
 				return;
 			}
 			if (!(ppp->active_filter &&
-			      BPF_PROG_RUN(ppp->active_filter, skb) == 0))
+			      bpf_prog_run(ppp->active_filter, skb) == 0))
 				ppp->last_recv = jiffies;
 			__skb_pull(skb, 2);
 		} else

drivers/net/team/team_mode_loadbalance.c

···
 	fp = rcu_dereference_bh(lb_priv->fp);
 	if (unlikely(!fp))
 		return 0;
-	lhash = BPF_PROG_RUN(fp, skb);
+	lhash = bpf_prog_run(fp, skb);
 	c = (char *) &lhash;
 	return c[0] ^ c[1] ^ c[2] ^ c[3];
 }

include/linux/bpf.h

···
 /* an array of programs to be executed under rcu_lock.
  *
  * Typical usage:
- * ret = BPF_PROG_RUN_ARRAY(&bpf_prog_array, ctx, BPF_PROG_RUN);
+ * ret = BPF_PROG_RUN_ARRAY(&bpf_prog_array, ctx, bpf_prog_run);
  *
  * the structure returned by bpf_prog_array_alloc() should be populated
  * with program pointers and the last pointer must be NULL.

include/linux/filter.h

···
 
 DECLARE_STATIC_KEY_FALSE(bpf_stats_enabled_key);
 
-#define __BPF_PROG_RUN(prog, ctx, dfunc)	({			\
-	u32 __ret;							\
-	cant_migrate();							\
-	if (static_branch_unlikely(&bpf_stats_enabled_key)) {		\
-		struct bpf_prog_stats *__stats;				\
-		u64 __start = sched_clock();				\
-		__ret = dfunc(ctx, (prog)->insnsi, (prog)->bpf_func);	\
-		__stats = this_cpu_ptr(prog->stats);			\
-		u64_stats_update_begin(&__stats->syncp);		\
-		__stats->cnt++;						\
-		__stats->nsecs += sched_clock() - __start;		\
-		u64_stats_update_end(&__stats->syncp);			\
-	} else {							\
-		__ret = dfunc(ctx, (prog)->insnsi, (prog)->bpf_func);	\
-	}								\
-	__ret; })
+typedef unsigned int (*bpf_dispatcher_fn)(const void *ctx,
+					  const struct bpf_insn *insnsi,
+					  unsigned int (*bpf_func)(const void *,
+								   const struct bpf_insn *));
 
-#define BPF_PROG_RUN(prog, ctx)						\
-	__BPF_PROG_RUN(prog, ctx, bpf_dispatcher_nop_func)
+static __always_inline u32 __bpf_prog_run(const struct bpf_prog *prog,
+					  const void *ctx,
+					  bpf_dispatcher_fn dfunc)
+{
+	u32 ret;
+
+	cant_migrate();
+	if (static_branch_unlikely(&bpf_stats_enabled_key)) {
+		struct bpf_prog_stats *stats;
+		u64 start = sched_clock();
+
+		ret = dfunc(ctx, prog->insnsi, prog->bpf_func);
+		stats = this_cpu_ptr(prog->stats);
+		u64_stats_update_begin(&stats->syncp);
+		stats->cnt++;
+		stats->nsecs += sched_clock() - start;
+		u64_stats_update_end(&stats->syncp);
+	} else {
+		ret = dfunc(ctx, prog->insnsi, prog->bpf_func);
+	}
+	return ret;
+}
+
+static __always_inline u32 bpf_prog_run(const struct bpf_prog *prog, const void *ctx)
+{
+	return __bpf_prog_run(prog, ctx, bpf_dispatcher_nop_func);
+}
 
 /*
  * Use in preemptible and therefore migratable context to make sure that
···
 	u32 ret;
 
 	migrate_disable();
-	ret = __BPF_PROG_RUN(prog, ctx, bpf_dispatcher_nop_func);
+	ret = bpf_prog_run(prog, ctx);
 	migrate_enable();
 	return ret;
 }
···
 		memset(cb_data, 0, sizeof(cb_saved));
 	}
 
-	res = BPF_PROG_RUN(prog, skb);
+	res = bpf_prog_run(prog, skb);
 
 	if (unlikely(prog->cb_access))
 		memcpy(cb_data, cb_saved, sizeof(cb_saved));
···
 	 * under local_bh_disable(), which provides the needed RCU protection
 	 * for accessing map entries.
 	 */
-	u32 act = __BPF_PROG_RUN(prog, xdp, BPF_DISPATCHER_FUNC(xdp));
+	u32 act = __bpf_prog_run(prog, xdp, BPF_DISPATCHER_FUNC(xdp));
 
 	if (static_branch_unlikely(&bpf_master_redirect_enabled_key)) {
 		if (act == XDP_TX && netif_is_bond_slave(xdp->rxq->dev))
···
 		};
 		u32 act;
 
-		act = BPF_PROG_SK_LOOKUP_RUN_ARRAY(run_array, ctx, BPF_PROG_RUN);
+		act = BPF_PROG_SK_LOOKUP_RUN_ARRAY(run_array, ctx, bpf_prog_run);
 		if (act == SK_PASS) {
 			selected_sk = ctx.selected_sk;
 			no_reuseport = ctx.no_reuseport;
···
 		};
 		u32 act;
 
-		act = BPF_PROG_SK_LOOKUP_RUN_ARRAY(run_array, ctx, BPF_PROG_RUN);
+		act = BPF_PROG_SK_LOOKUP_RUN_ARRAY(run_array, ctx, bpf_prog_run);
 		if (act == SK_PASS) {
 			selected_sk = ctx.selected_sk;
 			no_reuseport = ctx.no_reuseport;

kernel/bpf/bpf_iter.c

···
 
 	rcu_read_lock();
 	migrate_disable();
-	ret = BPF_PROG_RUN(prog, ctx);
+	ret = bpf_prog_run(prog, ctx);
 	migrate_enable();
 	rcu_read_unlock();
 

kernel/bpf/cgroup.c

···
 	struct cgroup *cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
 	int ret;
 
-	ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[type], sk, BPF_PROG_RUN);
+	ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[type], sk, bpf_prog_run);
 	return ret == 1 ? 0 : -EPERM;
 }
 EXPORT_SYMBOL(__cgroup_bpf_run_filter_sk);
···
 
 	cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
 	ret = BPF_PROG_RUN_ARRAY_FLAGS(cgrp->bpf.effective[type], &ctx,
-				       BPF_PROG_RUN, flags);
+				       bpf_prog_run, flags);
 
 	return ret == 1 ? 0 : -EPERM;
 }
···
 	int ret;
 
 	ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[type], sock_ops,
-				 BPF_PROG_RUN);
+				 bpf_prog_run);
 	return ret == 1 ? 0 : -EPERM;
 }
 EXPORT_SYMBOL(__cgroup_bpf_run_filter_sock_ops);
···
 	rcu_read_lock();
 	cgrp = task_dfl_cgroup(current);
 	allow = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[type], &ctx,
-				   BPF_PROG_RUN);
+				   bpf_prog_run);
 	rcu_read_unlock();
 
 	return !allow;
···
 
 	rcu_read_lock();
 	cgrp = task_dfl_cgroup(current);
-	ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[type], &ctx, BPF_PROG_RUN);
+	ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[type], &ctx, bpf_prog_run);
 	rcu_read_unlock();
 
 	kfree(ctx.cur_val);
···
 
 	lock_sock(sk);
 	ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[BPF_CGROUP_SETSOCKOPT],
-				 &ctx, BPF_PROG_RUN);
+				 &ctx, bpf_prog_run);
 	release_sock(sk);
 
 	if (!ret) {
···
 
 	lock_sock(sk);
 	ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[BPF_CGROUP_GETSOCKOPT],
-				 &ctx, BPF_PROG_RUN);
+				 &ctx, bpf_prog_run);
 	release_sock(sk);
 
 	if (!ret) {
···
 	 */
 
 	ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[BPF_CGROUP_GETSOCKOPT],
-				 &ctx, BPF_PROG_RUN);
+				 &ctx, bpf_prog_run);
 	if (!ret)
 		return -EPERM;
 

kernel/bpf/core.c

···
  *	@err: pointer to error variable
  *
  * Try to JIT eBPF program, if JIT is not available, use interpreter.
- * The BPF program will be executed via BPF_PROG_RUN() macro.
+ * The BPF program will be executed via bpf_prog_run() function.
  *
  * Return: the &fp argument along with &err set to 0 for success or
  * a negative errno code on failure

kernel/bpf/trampoline.c

···
 	u64_stats_update_end(&stats->syncp);
 }
 
-/* The logic is similar to BPF_PROG_RUN, but with an explicit
+/* The logic is similar to bpf_prog_run(), but with an explicit
  * rcu_read_lock() and migrate_disable() which are required
  * for the trampoline. The macro is split into
  * call __bpf_prog_enter

kernel/bpf/verifier.c

···
 		subprog_end = env->subprog_info[i + 1].start;
 
 		len = subprog_end - subprog_start;
-		/* BPF_PROG_RUN doesn't call subprogs directly,
+		/* bpf_prog_run() doesn't call subprogs directly,
 		 * hence main prog stats include the runtime of subprogs.
 		 * subprogs don't have IDs and not reachable via prog_get_next_id
 		 * func[i]->stats will never be accessed and stays NULL

kernel/events/core.c

···
 	if (unlikely(__this_cpu_inc_return(bpf_prog_active) != 1))
 		goto out;
 	rcu_read_lock();
-	ret = BPF_PROG_RUN(event->prog, &ctx);
+	ret = bpf_prog_run(event->prog, &ctx);
 	rcu_read_unlock();
 out:
 	__this_cpu_dec(bpf_prog_active);

kernel/trace/bpf_trace.c

···
 	 * out of events when it was updated in between this and the
 	 * rcu_dereference() which is accepted risk.
 	 */
-	ret = BPF_PROG_RUN_ARRAY_CHECK(call->prog_array, ctx, BPF_PROG_RUN);
+	ret = BPF_PROG_RUN_ARRAY_CHECK(call->prog_array, ctx, bpf_prog_run);
 
  out:
 	__this_cpu_dec(bpf_prog_active);
···
 {
 	cant_sleep();
 	rcu_read_lock();
-	(void) BPF_PROG_RUN(prog, args);
+	(void) bpf_prog_run(prog, args);
 	rcu_read_unlock();
 }
 

lib/test_bpf.c

···
 	start = ktime_get_ns();
 
 	for (i = 0; i < runs; i++)
-		ret = BPF_PROG_RUN(fp, data);
+		ret = bpf_prog_run(fp, data);
 
 	finish = ktime_get_ns();
 	migrate_enable();

net/bpf/test_run.c

···
 		if (xdp)
 			*retval = bpf_prog_run_xdp(prog, ctx);
 		else
-			*retval = BPF_PROG_RUN(prog, ctx);
+			*retval = bpf_prog_run(prog, ctx);
 	} while (bpf_test_timer_continue(&t, repeat, &ret, time));
 	bpf_reset_run_ctx(old_ctx);
 	bpf_test_timer_leave(&t);
···
 	struct bpf_raw_tp_test_run_info *info = data;
 
 	rcu_read_lock();
-	info->retval = BPF_PROG_RUN(info->prog, info->ctx);
+	info->retval = bpf_prog_run(info->prog, info->ctx);
 	rcu_read_unlock();
 }
 
···
 	bpf_test_timer_enter(&t);
 	do {
 		ctx.selected_sk = NULL;
-		retval = BPF_PROG_SK_LOOKUP_RUN_ARRAY(progs, ctx, BPF_PROG_RUN);
+		retval = BPF_PROG_SK_LOOKUP_RUN_ARRAY(progs, ctx, bpf_prog_run);
 	} while (bpf_test_timer_continue(&t, repeat, &ret, &duration));
 	bpf_test_timer_leave(&t);
 

net/core/filter.c

···
  * Run the eBPF program and then cut skb->data to correct size returned by
  * the program. If pkt_len is 0 we toss packet. If skb->len is smaller
  * than pkt_len we keep whole skb->data. This is the socket level
- * wrapper to BPF_PROG_RUN. It returns 0 if the packet should
+ * wrapper to bpf_prog_run. It returns 0 if the packet should
  * be accepted or -EPERM if the packet should be tossed.
  *
  */
···
 	enum sk_action action;
 
 	bpf_init_reuseport_kern(&reuse_kern, reuse, sk, skb, migrating_sk, hash);
-	action = BPF_PROG_RUN(prog, &reuse_kern);
+	action = bpf_prog_run(prog, &reuse_kern);
 
 	if (action == SK_PASS)
 		return reuse_kern.selected_sk;

net/core/ptp_classifier.c

···
 
 unsigned int ptp_classify_raw(const struct sk_buff *skb)
 {
-	return BPF_PROG_RUN(ptp_insns, skb);
+	return bpf_prog_run(ptp_insns, skb);
 }
 EXPORT_SYMBOL_GPL(ptp_classify_raw);
 

net/netfilter/xt_bpf.c

···
 {
 	const struct xt_bpf_info *info = par->matchinfo;
 
-	return BPF_PROG_RUN(info->filter, skb);
+	return bpf_prog_run(info->filter, skb);
 }
 
 static bool bpf_mt_v1(const struct sk_buff *skb, struct xt_action_param *par)

net/sched/act_bpf.c

···
 	if (at_ingress) {
 		__skb_push(skb, skb->mac_len);
 		bpf_compute_data_pointers(skb);
-		filter_res = BPF_PROG_RUN(filter, skb);
+		filter_res = bpf_prog_run(filter, skb);
 		__skb_pull(skb, skb->mac_len);
 	} else {
 		bpf_compute_data_pointers(skb);
-		filter_res = BPF_PROG_RUN(filter, skb);
+		filter_res = bpf_prog_run(filter, skb);
 	}
 	if (skb_sk_is_prefetched(skb) && filter_res != TC_ACT_OK)
 		skb_orphan(skb);

net/sched/cls_bpf.c

···
 			/* It is safe to push/pull even if skb_shared() */
 			__skb_push(skb, skb->mac_len);
 			bpf_compute_data_pointers(skb);
-			filter_res = BPF_PROG_RUN(prog->filter, skb);
+			filter_res = bpf_prog_run(prog->filter, skb);
 			__skb_pull(skb, skb->mac_len);
 		} else {
 			bpf_compute_data_pointers(skb);
-			filter_res = BPF_PROG_RUN(prog->filter, skb);
+			filter_res = bpf_prog_run(prog->filter, skb);
 		}
 
 		if (prog->exts_integrated) {