diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index 200995c5210e..9d455fd543b9 100644
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -41,6 +41,7 @@ struct perf_guest_info_callbacks {
 #include <linux/mutex.h>
 #include <linux/rculist.h>
 #include <linux/rcupdate.h>
+#include <linux/rcupdate_trace.h>
 #include <linux/spinlock.h>
 #include <linux/hrtimer.h>
 #include <linux/fs.h>
@@ -1269,8 +1270,13 @@ static inline struct perf_guest_info_callbacks *perf_get_guest_cbs(void)
         * pending stores/changes to the callback pointers are visible before a
         * non-NULL perf_guest_cbs is visible to readers, and to prevent a
         * module from unloading callbacks while readers are active.
+        *
+        * BPF locked rcu using rcu_read_lock_trace() in
+        * bpf_prog_test_run_syscall()
         */
-       return rcu_dereference(perf_guest_cbs);
+       return rcu_dereference_check(perf_guest_cbs,
+                       rcu_read_lock_trace_held() ||
+                       rcu_read_lock_held());
 }
 extern int perf_register_guest_info_callbacks(struct perf_guest_info_callbacks *callbacks);
 extern int perf_unregister_guest_info_callbacks(struct perf_guest_info_callbacks *callbacks);
diff --git a/kernel/events/callchain.c b/kernel/events/callchain.c
index 58cbe357fb2b..e47402c97ae0 100644
--- a/kernel/events/callchain.c
+++ b/kernel/events/callchain.c
@@ -11,6 +11,7 @@
 #include <linux/perf_event.h>
 #include <linux/slab.h>
 #include <linux/sched/task_stack.h>
+#include <linux/rcupdate_trace.h>
 
 #include "internal.h"
 
@@ -32,7 +33,7 @@ static inline size_t perf_callchain_entry__sizeof(void)
 static DEFINE_PER_CPU(int, callchain_recursion[PERF_NR_CONTEXTS]);
 static atomic_t nr_callchain_events;
 static DEFINE_MUTEX(callchain_mutex);
-static struct callchain_cpus_entries *callchain_cpus_entries;
+static struct callchain_cpus_entries __rcu *callchain_cpus_entries;
 
 
 __weak void perf_callchain_kernel(struct perf_callchain_entry_ctx *entry,
@@ -157,8 +158,13 @@ struct perf_callchain_entry *get_callchain_entry(int *rctx)
        *rctx = get_recursion_context(this_cpu_ptr(callchain_recursion));
        if (*rctx == -1)
                return NULL;
-
-       entries = rcu_dereference(callchain_cpus_entries);
+       /*
+        * BPF locked rcu using rcu_read_lock_trace() in
+        * bpf_prog_test_run_syscall()
+        */
+       entries = rcu_dereference_check(callchain_cpus_entries,
+                       rcu_read_lock_trace_held() ||
+                       rcu_read_lock_held());
        if (!entries) {
                put_recursion_context(this_cpu_ptr(callchain_recursion), *rctx);
                return NULL;