selftests/bpf: Add tests for cgrp_local_storage with local percpu kptr

Add a non-sleepable cgrp_local_storage test with percpu kptr. The
test does allocation of percpu data, assigning values to percpu
data and retrieval of percpu data. The de-allocation of percpu
data is done when the map is freed.
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20230827152827.2001784-1-yonghong.song@linux.devSigned-off-by: Alexei Starovoitov <ast@kernel.org>

tools/testing/selftests/bpf/prog_tests/percpu_alloc.c

 // SPDX-License-Identifier: GPL-2.0
 #include <test_progs.h>
 #include "percpu_alloc_array.skel.h"
+#include "percpu_alloc_cgrp_local_storage.skel.h"
 
 static void test_array(void)
 {
@@ -69,10 +70,49 @@ static void test_array_sleepable(void)
 	percpu_alloc_array__destroy(skel);
 }
 
+static void test_cgrp_local_storage(void)
+{
+	struct percpu_alloc_cgrp_local_storage *skel;
+	int err, cgroup_fd, prog_fd;
+	LIBBPF_OPTS(bpf_test_run_opts, topts);
+
+	cgroup_fd = test__join_cgroup("/percpu_alloc");
+	if (!ASSERT_GE(cgroup_fd, 0, "join_cgroup /percpu_alloc"))
+		return;
+
+	skel = percpu_alloc_cgrp_local_storage__open();
+	if (!ASSERT_OK_PTR(skel, "percpu_alloc_cgrp_local_storage__open"))
+		goto close_fd;
+
+	skel->rodata->nr_cpus = libbpf_num_possible_cpus();
+
+	err = percpu_alloc_cgrp_local_storage__load(skel);
+	if (!ASSERT_OK(err, "percpu_alloc_cgrp_local_storage__load"))
+		goto destroy_skel;
+
+	err = percpu_alloc_cgrp_local_storage__attach(skel);
+	if (!ASSERT_OK(err, "percpu_alloc_cgrp_local_storage__attach"))
+		goto destroy_skel;
+
+	prog_fd = bpf_program__fd(skel->progs.test_cgrp_local_storage_1);
+	err = bpf_prog_test_run_opts(prog_fd, &topts);
+	ASSERT_OK(err, "test_run cgrp_local_storage 1-3");
+	ASSERT_EQ(topts.retval, 0, "test_run cgrp_local_storage 1-3");
+	ASSERT_EQ(skel->bss->cpu0_field_d, 2, "cpu0_field_d");
+	ASSERT_EQ(skel->bss->sum_field_c, 1, "sum_field_c");
+
+destroy_skel:
+	percpu_alloc_cgrp_local_storage__destroy(skel);
+close_fd:
+	close(cgroup_fd);
+}
+
 void test_percpu_alloc(void)
 {
 	if (test__start_subtest("array"))
 		test_array();
 	if (test__start_subtest("array_sleepable"))
 		test_array_sleepable();
+	if (test__start_subtest("cgrp_local_storage"))
+		test_cgrp_local_storage();
 }

tools/testing/selftests/bpf/progs/percpu_alloc_cgrp_local_storage.c

+#include "bpf_experimental.h"
+
+struct val_t {
+	long b, c, d;
+};
+
+struct elem {
+	long sum;
+	struct val_t __percpu_kptr *pc;
+};
+
+struct {
+	__uint(type, BPF_MAP_TYPE_CGRP_STORAGE);
+	__uint(map_flags, BPF_F_NO_PREALLOC);
+	__type(key, int);
+	__type(value, struct elem);
+} cgrp SEC(".maps");
+
+const volatile int nr_cpus;
+
+/* Initialize the percpu object */
+SEC("fentry/bpf_fentry_test1")
+int BPF_PROG(test_cgrp_local_storage_1)
+{
+	struct task_struct *task;
+	struct val_t __percpu_kptr *p;
+	struct elem *e;
+
+	task = bpf_get_current_task_btf();
+	e = bpf_cgrp_storage_get(&cgrp, task->cgroups->dfl_cgrp, 0,
+				 BPF_LOCAL_STORAGE_GET_F_CREATE);
+	if (!e)
+		return 0;
+
+	p = bpf_percpu_obj_new(struct val_t);
+	if (!p)
+		return 0;
+
+	p = bpf_kptr_xchg(&e->pc, p);
+	if (p)
+		bpf_percpu_obj_drop(p);
+
+	return 0;
+}
+
+/* Percpu data collection */
+SEC("fentry/bpf_fentry_test2")
+int BPF_PROG(test_cgrp_local_storage_2)
+{
+	struct task_struct *task;
+	struct val_t __percpu_kptr *p;
+	struct val_t *v;
+	struct elem *e;
+
+	task = bpf_get_current_task_btf();
+	e = bpf_cgrp_storage_get(&cgrp, task->cgroups->dfl_cgrp, 0, 0);
+	if (!e)
+		return 0;
+
+	p = e->pc;
+	if (!p)
+		return 0;
+
+	v = bpf_per_cpu_ptr(p, 0);
+	if (!v)
+		return 0;
+	v->c = 1;
+	v->d = 2;
+	return 0;
+}
+
+int cpu0_field_d, sum_field_c;
+
+/* Summarize percpu data collection */
+SEC("fentry/bpf_fentry_test3")
+int BPF_PROG(test_cgrp_local_storage_3)
+{
+	struct task_struct *task;
+	struct val_t __percpu_kptr *p;
+	struct val_t *v;
+	struct elem *e;
+	int i;
+
+	task = bpf_get_current_task_btf();
+	e = bpf_cgrp_storage_get(&cgrp, task->cgroups->dfl_cgrp, 0, 0);
+	if (!e)
+		return 0;
+
+	p = e->pc;
+	if (!p)
+		return 0;
+
+	bpf_for(i, 0, nr_cpus) {
+		v = bpf_per_cpu_ptr(p, i);
+		if (v) {
+			if (i == 0)
+				cpu0_field_d = v->d;
+			sum_field_c += v->c;
+		}
+	}
+
+	return 0;
+}
+
+char _license[] SEC("license") = "GPL";