Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Song Liu <songliubraving@fb.com>
To:     Andrii Nakryiko <andrii.nakryiko@gmail.com>
CC:     bpf <bpf@vger.kernel.org>, Networking <netdev@vger.kernel.org>,
        open list <linux-kernel@vger.kernel.org>,
        Ingo Molnar <mingo@redhat.com>,
        "Peter Ziljstra" <peterz@infradead.org>,
        Alexei Starovoitov <ast@kernel.org>,
        "Daniel Borkmann" <daniel@iogearbox.net>,
        Andrii Nakryiko <andrii@kernel.org>,
        "john fastabend" <john.fastabend@gmail.com>,
        KP Singh <kpsingh@chromium.org>,
        "Kernel Team" <Kernel-team@fb.com>, Hao Luo <haoluo@google.com>
Subject: Re: [PATCH v2 bpf-next 2/4] selftests/bpf: add non-BPF_LSM test for
 task local storage
Thread-Topic: [PATCH v2 bpf-next 2/4] selftests/bpf: add non-BPF_LSM test for
 task local storage
Thread-Index: AQHW88ShH5IOjKYrS029recyIYC356o7/TsAgAAGEYA=
Date:   Wed, 27 Jan 2021 21:43:16 +0000
Message-ID: <4A77E6CE-82FE-4578-BC13-05583E2EA17D@fb.com>
References: <20210126085923.469759-1-songliubraving@fb.com>
 <20210126085923.469759-3-songliubraving@fb.com>
 <CAEf4BzZLJc9=JgZBmvRazHsZg+VLihaRi-3Pt8wrsT9am-eBGg@mail.gmail.com>
In-Reply-To: <CAEf4BzZLJc9=JgZBmvRazHsZg+VLihaRi-3Pt8wrsT9am-eBGg@mail.gmail.com>
Accept-Language: en-US
Content-Language: en-US
Content-Type: text/plain; charset="us-ascii"
Content-ID: <9952AD7C094D6F42A83ADA1D482C67CD@namprd15.prod.outlook.com>
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-MS-Exchange-CrossTenant-Network-Message-Id: f1afb5e5-a9c5-4e08-f030-08d8c30c8df5
X-MS-Exchange-CrossTenant-originalarrivaltime: 27 Jan 2021 21:43:16.0764
 (UTC)
X-MS-Exchange-CrossTenant-fromentityheader: Hosted
X-MS-Exchange-CrossTenant-id: 8ae927fe-1255-47a7-a2af-5f3a069daaa2
X-MS-Exchange-CrossTenant-mailboxtype: HOSTED
X-MS-Exchange-CrossTenant-userprincipalname: Ru8nb0G+6r1iqAG5d8bQjVhtzpf5zWbpaezGI9KehHGoAk0f2sCTEjldNgxz1dB42XPNCu9JIsTnXDgl73UeTw==
X-MS-Exchange-Transport-CrossTenantHeadersStamped: BYAPR15MB3095
X-OriginatorOrg: fb.com
X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10434:6.0.343,18.0.737
 definitions=2021-01-27_09:2021-01-27,2021-01-27 signatures=0
X-Proofpoint-Spam-Details: rule=fb_default_notspam policy=fb_default score=0 malwarescore=0
 phishscore=0 bulkscore=0 mlxlogscore=999 priorityscore=1501 spamscore=0
 impostorscore=0 lowpriorityscore=0 mlxscore=0 adultscore=0 suspectscore=0
 clxscore=1011 classifier=spam adjust=0 reason=mlx scancount=1
 engine=8.12.0-2009150000 definitions=main-2101270109
X-FB-Internal: deliver
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org


> On Jan 27, 2021, at 1:21 PM, Andrii Nakryiko <andrii.nakryiko@gmail.com> =
wrote:
>=20
> On Tue, Jan 26, 2021 at 1:21 AM Song Liu <songliubraving@fb.com> wrote:
>>=20
>> Task local storage is enabled for tracing programs. Add two tests for
>> task local storage without CONFIG_BPF_LSM.
>>=20
>> The first test measures the duration of a syscall by storing sys_enter
>> time in task local storage.
>>=20
>> The second test checks whether the kernel allows allocating task local
>> storage in exit_creds() (which it should not).
>>=20
>> Signed-off-by: Song Liu <songliubraving@fb.com>
>> ---
>> .../bpf/prog_tests/task_local_storage.c       | 85 +++++++++++++++++++
>> .../selftests/bpf/progs/task_local_storage.c  | 56 ++++++++++++
>> .../bpf/progs/task_local_storage_exit_creds.c | 32 +++++++
>> 3 files changed, 173 insertions(+)
>> create mode 100644 tools/testing/selftests/bpf/prog_tests/task_local_sto=
rage.c
>> create mode 100644 tools/testing/selftests/bpf/progs/task_local_storage.=
c
>> create mode 100644 tools/testing/selftests/bpf/progs/task_local_storage_=
exit_creds.c
>>=20
>> diff --git a/tools/testing/selftests/bpf/prog_tests/task_local_storage.c=
 b/tools/testing/selftests/bpf/prog_tests/task_local_storage.c
>> new file mode 100644
>> index 0000000000000..a8e2d3a476145
>> --- /dev/null
>> +++ b/tools/testing/selftests/bpf/prog_tests/task_local_storage.c
>> @@ -0,0 +1,85 @@
>> +// SPDX-License-Identifier: GPL-2.0
>> +/* Copyright (c) 2021 Facebook */
>> +
>> +#include <sys/types.h>
>> +#include <unistd.h>
>> +#include <test_progs.h>
>> +#include "task_local_storage.skel.h"
>> +#include "task_local_storage_exit_creds.skel.h"
>> +
>> +static unsigned int duration;
>> +
>> +static void check_usleep_duration(struct task_local_storage *skel,
>> +                                 __u64 time_us)
>> +{
>> +       __u64 syscall_duration;
>> +
>> +       usleep(time_us);
>> +
>> +       /* save syscall_duration measure in usleep() */
>> +       syscall_duration =3D skel->bss->syscall_duration;
>> +
>> +       /* time measured by the BPF program (in nanoseconds) should be
>> +        * within +/- 20% of time_us * 1000.
>> +        */
>> +       CHECK(syscall_duration < 800 * time_us, "syscall_duration",
>> +             "syscall_duration was too small\n");
>> +       CHECK(syscall_duration > 1200 * time_us, "syscall_duration",
>> +             "syscall_duration was too big\n");
>=20
> this is going to be very flaky, especially in Travis CI. Can you
> please use something more stable that doesn't rely on time?

Let me try.=20

>=20
>> +}
>> +
>> +static void test_syscall_duration(void)
>> +{
>> +       struct task_local_storage *skel;
>> +       int err;
>> +
>> +       skel =3D task_local_storage__open_and_load();
>> +       if (!ASSERT_OK_PTR(skel, "skel_open_and_load"))
>> +               return;
>> +
>> +       skel->bss->target_pid =3D getpid();
>=20
> you are getting process ID, but comparing it with thread ID in BPF
> code. It will stop working properly if/when tests will be run in
> separate threads, so please use gettid() instead.

Will fix.=20

>=20
>> +
>> +       err =3D task_local_storage__attach(skel);
>> +       if (!ASSERT_OK(err, "skel_attach"))
>> +               goto out;
>> +
>> +       check_usleep_duration(skel, 2000);
>> +       check_usleep_duration(skel, 3000);
>> +       check_usleep_duration(skel, 4000);
>> +
>> +out:
>> +       task_local_storage__destroy(skel);
>> +}
>> +
>> +static void test_exit_creds(void)
>> +{
>> +       struct task_local_storage_exit_creds *skel;
>> +       int err;
>> +
>> +       skel =3D task_local_storage_exit_creds__open_and_load();
>> +       if (!ASSERT_OK_PTR(skel, "skel_open_and_load"))
>> +               return;
>> +
>> +       err =3D task_local_storage_exit_creds__attach(skel);
>> +       if (!ASSERT_OK(err, "skel_attach"))
>> +               goto out;
>> +
>> +       /* trigger at least one exit_creds() */
>> +       if (CHECK_FAIL(system("ls > /dev/null")))
>> +               goto out;
>> +
>> +       /* sync rcu, so the following reads could get latest values */
>> +       kern_sync_rcu();
>=20
> what are we waiting for here? you don't detach anything... system() is
> definitely going to complete by now, so whatever counter was or was
> not updated will be reflected here. Seems like kern_sync_rcu() is not
> needed?

IIUC, without sync_ruc(), even system() is finished, the kernel may not=20
have called exit_creds() for the "ls" task yet. Then the following check
for null_ptr_count !=3D 0 would fail.=20

>=20
>> +       ASSERT_EQ(skel->bss->valid_ptr_count, 0, "valid_ptr_count");
>> +       ASSERT_NEQ(skel->bss->null_ptr_count, 0, "null_ptr_count");
>> +out:
>> +       task_local_storage_exit_creds__destroy(skel);
>> +}
>> +
>> +void test_task_local_storage(void)
>> +{
>> +       if (test__start_subtest("syscall_duration"))
>> +               test_syscall_duration();
>> +       if (test__start_subtest("exit_creds"))
>> +               test_exit_creds();
>> +}
>=20
> [...]
>=20
>> +int valid_ptr_count =3D 0;
>> +int null_ptr_count =3D 0;
>> +
>> +SEC("fentry/exit_creds")
>> +int BPF_PROG(trace_exit_creds, struct task_struct *task)
>> +{
>> +       __u64 *ptr;
>> +
>> +       ptr =3D bpf_task_storage_get(&task_storage, task, 0,
>> +                                  BPF_LOCAL_STORAGE_GET_F_CREATE);
>> +       if (ptr)
>> +               valid_ptr_count++;
>> +       else
>> +               null_ptr_count++;
>=20
>=20
> use atomic increments?

Do you mean __sync_fetch_and_add()?=20

>=20
>> +       return 0;
>> +}
>> --
>> 2.24.1