diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 5fb3c610fa96a5dba4f5d82fff0c816f91da3a37..f63afdc43beccf2c16cfbb13ae17bc8972c1bb18 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -169,7 +169,9 @@ u64 bpf_tail_call(u64 ctx, u64 r2, u64 index, u64 r4, u64 r5);
 u64 bpf_get_stackid(u64 r1, u64 r2, u64 r3, u64 r4, u64 r5);
 void bpf_fd_array_map_clear(struct bpf_map *map);
 bool bpf_prog_array_compatible(struct bpf_array *array, const struct bpf_prog *fp);
+
 const struct bpf_func_proto *bpf_get_trace_printk_proto(void);
+const struct bpf_func_proto *bpf_get_event_output_proto(void);
 
 #ifdef CONFIG_BPF_SYSCALL
 DECLARE_PER_CPU(int, bpf_prog_active);
diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
index 70eda5aeb304281cb3462f25445886111726f78a..b7b0fb1292e745394d8c61db2fc2f50e53604f8f 100644
--- a/include/uapi/linux/bpf.h
+++ b/include/uapi/linux/bpf.h
@@ -347,6 +347,10 @@ enum bpf_func_id {
 #define BPF_F_ZERO_CSUM_TX		(1ULL << 1)
 #define BPF_F_DONT_FRAGMENT		(1ULL << 2)
 
+/* BPF_FUNC_perf_event_output flags. */
+#define BPF_F_INDEX_MASK		0xffffffffULL
+#define BPF_F_CURRENT_CPU		BPF_F_INDEX_MASK
+
 /* user accessible mirror of in-kernel sk_buff.
  * new fields can only be added to the end of this structure
  */
diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
index be0abf669ced3b5b228a10d5ad6ac99576fae0f7..e4248fe79513093b00dc33a039ed540aef34a1b9 100644
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -764,14 +764,21 @@ const struct bpf_func_proto bpf_map_delete_elem_proto __weak;
 const struct bpf_func_proto bpf_get_prandom_u32_proto __weak;
 const struct bpf_func_proto bpf_get_smp_processor_id_proto __weak;
 const struct bpf_func_proto bpf_ktime_get_ns_proto __weak;
+
 const struct bpf_func_proto bpf_get_current_pid_tgid_proto __weak;
 const struct bpf_func_proto bpf_get_current_uid_gid_proto __weak;
 const struct bpf_func_proto bpf_get_current_comm_proto __weak;
+
 const struct bpf_func_proto * __weak bpf_get_trace_printk_proto(void)
 {
 	return NULL;
 }
 
+const struct bpf_func_proto * __weak bpf_get_event_output_proto(void)
+{
+	return NULL;
+}
+
 /* Always built-in helper functions. */
 const struct bpf_func_proto bpf_tail_call_proto = {
 	.func		= NULL,
diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index f389629dade7a7a685fc907655f8e8aaca21d5f5..780bcbe1d4de33bbb24d2328b231bb056bc641b5 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c
@@ -225,11 +225,12 @@ static const struct bpf_func_proto bpf_perf_event_read_proto = {
 	.arg2_type	= ARG_ANYTHING,
 };
 
-static u64 bpf_perf_event_output(u64 r1, u64 r2, u64 index, u64 r4, u64 size)
+static u64 bpf_perf_event_output(u64 r1, u64 r2, u64 flags, u64 r4, u64 size)
 {
 	struct pt_regs *regs = (struct pt_regs *) (long) r1;
 	struct bpf_map *map = (struct bpf_map *) (long) r2;
 	struct bpf_array *array = container_of(map, struct bpf_array, map);
+	u64 index = flags & BPF_F_INDEX_MASK;
 	void *data = (void *) (long) r4;
 	struct perf_sample_data sample_data;
 	struct perf_event *event;
@@ -239,6 +240,10 @@ static u64 bpf_perf_event_output(u64 r1, u64 r2, u64 index, u64 r4, u64 size)
 		.data = data,
 	};
 
+	if (unlikely(flags & ~(BPF_F_INDEX_MASK)))
+		return -EINVAL;
+	if (index == BPF_F_CURRENT_CPU)
+		index = raw_smp_processor_id();
 	if (unlikely(index >= array->map.max_entries))
 		return -E2BIG;
 
@@ -272,6 +277,33 @@ static const struct bpf_func_proto bpf_perf_event_output_proto = {
 	.arg5_type	= ARG_CONST_STACK_SIZE,
 };
 
+static DEFINE_PER_CPU(struct pt_regs, bpf_pt_regs);
+
+static u64 bpf_event_output(u64 r1, u64 r2, u64 flags, u64 r4, u64 size)
+{
+	struct pt_regs *regs = this_cpu_ptr(&bpf_pt_regs);
+
+	perf_fetch_caller_regs(regs);
+
+	return bpf_perf_event_output((long)regs, r2, flags, r4, size);
+}
+
+static const struct bpf_func_proto bpf_event_output_proto = {
+	.func		= bpf_event_output,
+	.gpl_only	= true,
+	.ret_type	= RET_INTEGER,
+	.arg1_type	= ARG_PTR_TO_CTX,
+	.arg2_type	= ARG_CONST_MAP_PTR,
+	.arg3_type	= ARG_ANYTHING,
+	.arg4_type	= ARG_PTR_TO_STACK,
+	.arg5_type	= ARG_CONST_STACK_SIZE,
+};
+
+const struct bpf_func_proto *bpf_get_event_output_proto(void)
+{
+	return &bpf_event_output_proto;
+}
+
 static const struct bpf_func_proto *tracing_func_proto(enum bpf_func_id func_id)
 {
 	switch (func_id) {
diff --git a/net/core/filter.c b/net/core/filter.c
index 5d2ac2b9d1c4fcce377ed90a2cd861c04eb7f613..218e5de8c402b9fac28083fbf1400590022be514 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -2039,6 +2039,8 @@ tc_cls_act_func_proto(enum bpf_func_id func_id)
 		return &bpf_redirect_proto;
 	case BPF_FUNC_get_route_realm:
 		return &bpf_get_route_realm_proto;
+	case BPF_FUNC_perf_event_output:
+		return bpf_get_event_output_proto();
 	default:
 		return sk_filter_func_proto(func_id);
 	}