Received-SPF: pass (google.com: domain of linux-kernel+bounces-106861-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Date: Mon, 18 Mar 2024 18:30:54 -0600
From: Jose Fernandez <josef@netflix.com>
To: Jiri Olsa <olsajiri@gmail.com>
Cc: Alexei Starovoitov <ast@kernel.org>,
	Daniel Borkmann <daniel@iogearbox.net>,
	Andrii Nakryiko <andrii@kernel.org>,
	Martin KaFai Lau <martin.lau@linux.dev>,
	Eduard Zingerman <eddyz87@gmail.com>, Song Liu <song@kernel.org>,
	Yonghong Song <yonghong.song@linux.dev>,
	John Fastabend <john.fastabend@gmail.com>,
	KP Singh <kpsingh@kernel.org>, Stanislav Fomichev <sdf@google.com>,
	Hao Luo <haoluo@google.com>, bpf@vger.kernel.org,
	linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org,
	Tycho Andersen <tycho@tycho.pizza>
Subject: Re: [PATCH V2 bpf-next 2/2] selftests/bpf: add selftest for
 bpf_task_get_cgroup
Message-ID: <20240319003054.tegzffeydaec45ql@SlimPro9i>
References: <20240316162241.628855-1-josef@netflix.com>
 <20240316162241.628855-2-josef@netflix.com>
 <ZfgQXlU_wKpXScwh@krava>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <ZfgQXlU_wKpXScwh@krava>

On 24/03/18 10:58AM, Jiri Olsa wrote:
> On Sat, Mar 16, 2024 at 10:22:41AM -0600, Jose Fernandez wrote:
> 
> SNIP
> 
> > +void test_task_get_cgroup(void)
> > +{
> > +	struct test_task_get_cgroup *skel;
> > +	int err, fd;
> > +	pid_t pid;
> > +	__u64 cgroup_id, expected_cgroup_id;
> > +	const struct timespec req = {
> > +		.tv_sec = 1,
> > +		.tv_nsec = 0,
> > +	};
> > +
> > +	fd = test__join_cgroup(TEST_CGROUP);
> > +	if (!ASSERT_OK(fd < 0, "test_join_cgroup_TEST_CGROUP"))
> > +		return;
> > +
> > +	skel = test_task_get_cgroup__open();
> > +	if (!ASSERT_OK_PTR(skel, "test_task_get_cgroup__open"))
> > +		goto cleanup;
> > +
> > +	err = test_task_get_cgroup__load(skel);
> > +	if (!ASSERT_OK(err, "test_task_get_cgroup__load"))
> > +		goto cleanup;
> 
> nit, you could call test_task_get_cgroup__open_and_load

I'll rename.

> > +
> > +	err = test_task_get_cgroup__attach(skel);
> > +	if (!ASSERT_OK(err, "test_task_get_cgroup__attach"))
> > +		goto cleanup;
> > +
> > +	pid = getpid();
> > +	expected_cgroup_id = get_cgroup_id(TEST_CGROUP);
> > +	if (!ASSERT_GT(expected_cgroup_id, 0, "get_cgroup_id"))
> > +		goto cleanup;
> > +
> > +	/* Trigger nanosleep to enter the sched_switch tracepoint */
> > +	/* The previous task should be this process */
> > +	syscall(__NR_nanosleep, &req, NULL);
> 
> would smaller sleep do? also we have our own usleep (in test_progs.c)
> that calls nanosleep

Yes a smaller sleep should be fine.
I'll reduce the sleep and use the usleep helper.

> > +
> > +	err = bpf_map_lookup_elem(bpf_map__fd(skel->maps.pid_to_cgid_map), &pid,
> > +				  &cgroup_id);
> > +
> > +	if (!ASSERT_OK(err, "bpf_map_lookup_elem"))
> > +		goto cleanup;
> > +
> > +	ASSERT_EQ(cgroup_id, expected_cgroup_id, "cgroup_id");
> > +
> > +cleanup:
> > +	test_task_get_cgroup__destroy(skel);
> > +	close(fd);
> > +}
> > diff --git a/tools/testing/selftests/bpf/progs/test_task_get_cgroup.c b/tools/testing/selftests/bpf/progs/test_task_get_cgroup.c
> > new file mode 100644
> > index 000000000000..580f8f0657d5
> > --- /dev/null
> > +++ b/tools/testing/selftests/bpf/progs/test_task_get_cgroup.c
> > @@ -0,0 +1,37 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +// Copyright 2024 Netflix, Inc.
> > +
> > +#include "vmlinux.h"
> > +#include <bpf/bpf_helpers.h>
> > +#include <bpf/bpf_tracing.h>
> > +
> > +struct cgroup *bpf_task_get_cgroup(struct task_struct *task) __ksym;
> > +void bpf_cgroup_release(struct cgroup *cgrp) __ksym;
> > +
> > +struct {
> > +	__uint(type, BPF_MAP_TYPE_HASH);
> > +	__uint(max_entries, 4096);
> > +	__type(key, __u32);
> > +	__type(value, __u64);
> > +} pid_to_cgid_map SEC(".maps");
> > +
> > +SEC("tp_btf/sched_switch")
> > +int BPF_PROG(sched_switch, bool preempt, struct task_struct *prev,
> > +	     struct task_struct *next)
> > +{
> > +	struct cgroup *cgrp;
> > +	u64 cgroup_id;
> > +	u32 pid;
> > +
> 
> could you filter for your pid in here like we do in other places,
> (eg in progs/kprobe_multi.c)
> 
> in which case you won't need hash map, but just a single value
> to store the cgroup id to
> 
> jirka

I'll apply this suggestion as well and include it in V3. 
Thanks for the feedback.

> 
> > +	cgrp = bpf_task_get_cgroup(prev);
> > +	if (cgrp == NULL)
> > +		return 0;
> > +	cgroup_id = cgrp->kn->id;
> > +	pid = prev->pid;
> > +	bpf_map_update_elem(&pid_to_cgid_map, &pid, &cgroup_id, BPF_ANY);
> > +
> > +	bpf_cgroup_release(cgrp);
> > +	return 0;
> > +}
> > +
> > +char _license[] SEC("license") = "GPL";
> > -- 
> > 2.40.1
> >