From patchwork Wed Jul 18 17:33:58 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Claudio X-Patchwork-Id: 10533003 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id 6416C600D0 for ; Wed, 18 Jul 2018 17:34:31 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 8B73920182 for ; Wed, 18 Jul 2018 17:34:29 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 7F1A426222; Wed, 18 Jul 2018 17:34:29 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-7.8 required=2.0 tests=BAYES_00, DKIM_ADSP_CUSTOM_MED, DKIM_SIGNED, FREEMAIL_FROM, MAILING_LIST_MULTI, RCVD_IN_DNSWL_HI, T_DKIM_INVALID autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 53D6B20182 for ; Wed, 18 Jul 2018 17:34:28 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1731498AbeGRSNX (ORCPT ); Wed, 18 Jul 2018 14:13:23 -0400 Received: from mail-wm0-f66.google.com ([74.125.82.66]:38421 "EHLO mail-wm0-f66.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1731455AbeGRSNX (ORCPT ); Wed, 18 Jul 2018 14:13:23 -0400 Received: by mail-wm0-f66.google.com with SMTP id 69-v6so3679588wmf.3; Wed, 18 Jul 2018 10:34:25 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id; bh=ZHZyvZWaYFlZy4t9Me4Hp2dW+wIxR1+rJH+sanP6thA=; b=IDXGPNgOw5OlBwmj9rKhQOUlL64pFA6kO5IZH5iVODOYlMifpA7bvzC23IXAN938X4 DvrglwxC9u+26tSh2Q3j/OS1JNjAUoHKihOqWgGd9TjoEBj0c4tObIkrFrpuijJ4JdRi ovNC/oz5crcJOGpXdjWPeVaqb2WuOCTauEtazRDU8aTzvQHEwM//aKVVFNUQUkiEWwhz bP97n/BB+bcobow6jNAc8jPlj9Vmlk7XGE0odxx6qmXEmIYK000LWaw5rp1iniZVXLJ9 b3zJY0vEpGFNR19rk5i9cJUTTmAkKDzGzFWSEdTxSbzZUmADUxfTUCnepLRZJY54hoFB FWJQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id; bh=ZHZyvZWaYFlZy4t9Me4Hp2dW+wIxR1+rJH+sanP6thA=; b=EQbmmwgmrtDcOtaESLbq3mFOAvNaXmPTDSMpdnaF9wq1pM158Dy/PP1ZdtpM9R4roL VrKPYNDgJbdM5Q8wp7BIP/IUof48U0JRI8tiIe0qtSVxoDg3R2gfVJjhY3wvDSxm8bBR EY9C2Jhxe4AwtLz2oRws1mVwxmmf6kwijpad637VICKfc8lXvTHnFXzvf7+8rA5gM2Iy qMAXL3LOzo8eDk6MyxOq9WK1bdfdFB8e8n6CHrc9YPXmcV3xJ8zzROPv6oBpaHOxLs0z 0Jnerbkpg8dczGkCLrInkQ0tvc+IRA7uwo3yFtRrfwmBb/Cujvffg+JhfROxw6/8ptS8 viHw== X-Gm-Message-State: AOUpUlFuPvLXnO1s6p6jRwjyS5gEybAEbfqUcp72CZ13K6JmprzPfZda YcsJ6AIuK2SIWf6oFSgmghW00989DgQ= X-Google-Smtp-Source: AAOMgpfy8aSB+9ZoBpX84zXfziYxxXabzBXMFFzyxCO/0ubFhIVYQYyP6jtMhJTou/TqBgG+Y3GpGA== X-Received: by 2002:a1c:2807:: with SMTP id o7-v6mr2043601wmo.11.1531935264309; Wed, 18 Jul 2018 10:34:24 -0700 (PDT) Received: from localhost.localdomain (host241-253-dynamic.31-79-r.retail.telecomitalia.it. [79.31.253.241]) by smtp.gmail.com with ESMTPSA id f8-v6sm6197773wrj.9.2018.07.18.10.34.22 (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Wed, 18 Jul 2018 10:34:23 -0700 (PDT) From: Claudio To: linux-kselftest@vger.kernel.org Cc: linux-kernel@vger.kernel.org, cgroups@vger.kernel.org, Claudio , Claudio Zumbo , Shuah Khan , Roman Gushchin , Tejun Heo , kernel-team@fb.com Subject: [PATCH kselftest-next] Add cgroup core selftests Date: Wed, 18 Jul 2018 19:33:58 +0200 Message-Id: <20180718173358.14007-1-claudiozumbo@gmail.com> X-Mailer: git-send-email 2.11.0 Sender: linux-kselftest-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kselftest@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP This commit adds tests for some of the core functionalities of cgroups v2. The commit adds tests for some core principles of croup V2 API: - test_cgcore_internal_process_constraint Tests internal process constraint. You can't add a pid to a domain parent if a controller is enabled. - test_cgcore_top_down_constraint_enable Tests that you can't enable a controller on a child if it's not enabled on the parent. - test_cgcore_top_down_constraint_disable Tests that you can't disable a controller on a parent if it's enabled in a child. - test_cgcore_no_internal_process_constraint_on_threads Tests that there's no internal process constrain on threaded cgroups. You can add threads/processes on a parent with a controller enabled. - test_cgcore_parent_becomes_threaded Tests that when a child becomes threaded the parent type becomes domain threaded. - test_cgcore_invalid_domain In a situation like: A (domain threaded) - B (threaded) - C (domain) it tests that C can't be used until it is turned into a threaded cgroup. The "cgroup.type" file will report "domain (invalid)" in these cases. Operations which fail due to invalid topology use EOPNOTSUPP as the errno. - test_cgcore_populated In a situation like: A(0) - B(0) - C(1) \ D(0) It tests that A, B and C's "populated" fields would be 1 while D's 0. It tests that after the one process in C is moved to root, A,B and C's "populated" fields would flip to "0" and file modified events will be generated on the "cgroup.events" files of both cgroups. Signed-off-by: Claudio Zumbo Cc: Shuah Khan Cc: Roman Gushchin Cc: Tejun Heo Cc: kernel-team@fb.com Acked-by: Tejun Heo Reviewed-by: Roman Gushchin --- tools/testing/selftests/cgroup/Makefile | 2 + tools/testing/selftests/cgroup/cgroup_util.c | 8 + tools/testing/selftests/cgroup/cgroup_util.h | 1 + tools/testing/selftests/cgroup/test_core.c | 395 +++++++++++++++++++++++++++ 4 files changed, 406 insertions(+) create mode 100644 tools/testing/selftests/cgroup/test_core.c diff --git a/tools/testing/selftests/cgroup/Makefile b/tools/testing/selftests/cgroup/Makefile index f7a31392eb2f..23fbaa4a9630 100644 --- a/tools/testing/selftests/cgroup/Makefile +++ b/tools/testing/selftests/cgroup/Makefile @@ -4,7 +4,9 @@ CFLAGS += -Wall all: TEST_GEN_PROGS = test_memcontrol +TEST_GEN_PROGS += test_core include ../lib.mk $(OUTPUT)/test_memcontrol: cgroup_util.c +$(OUTPUT)/test_core: cgroup_util.c diff --git a/tools/testing/selftests/cgroup/cgroup_util.c b/tools/testing/selftests/cgroup/cgroup_util.c index 1e9e3c470561..1c5d2b2a583b 100644 --- a/tools/testing/selftests/cgroup/cgroup_util.c +++ b/tools/testing/selftests/cgroup/cgroup_util.c @@ -229,6 +229,14 @@ int cg_destroy(const char *cgroup) return ret; } +int cg_enter_current(const char *cgroup) +{ + char pidbuf[64]; + + snprintf(pidbuf, sizeof(pidbuf), "%d", getpid()); + return cg_write(cgroup, "cgroup.procs", pidbuf); +} + int cg_run(const char *cgroup, int (*fn)(const char *cgroup, void *arg), void *arg) diff --git a/tools/testing/selftests/cgroup/cgroup_util.h b/tools/testing/selftests/cgroup/cgroup_util.h index fe82a297d4e0..1ff6f9f1abdc 100644 --- a/tools/testing/selftests/cgroup/cgroup_util.h +++ b/tools/testing/selftests/cgroup/cgroup_util.h @@ -32,6 +32,7 @@ extern int cg_write(const char *cgroup, const char *control, char *buf); extern int cg_run(const char *cgroup, int (*fn)(const char *cgroup, void *arg), void *arg); +extern int cg_enter_current(const char *cgroup); extern int cg_run_nowait(const char *cgroup, int (*fn)(const char *cgroup, void *arg), void *arg); diff --git a/tools/testing/selftests/cgroup/test_core.c b/tools/testing/selftests/cgroup/test_core.c new file mode 100644 index 000000000000..be59f9c34ea2 --- /dev/null +++ b/tools/testing/selftests/cgroup/test_core.c @@ -0,0 +1,395 @@ +/* SPDX-License-Identifier: GPL-2.0 */ + +#include +#include +#include +#include +#include + +#include "../kselftest.h" +#include "cgroup_util.h" + +/* + * A(0) - B(0) - C(1) + * \ D(0) + * + * A, B and C's "populated" fields would be 1 while D's 0. + * test that after the one process in C is moved to root, + * A,B and C's "populated" fields would flip to "0" and file + * modified events will be generated on the + * "cgroup.events" files of both cgroups. + */ +static int test_cgcore_populated(const char *root) +{ + int ret = KSFT_FAIL; + char *cg_test_a = NULL, *cg_test_b = NULL; + char *cg_test_c = NULL, *cg_test_d = NULL; + + cg_test_a = cg_name(root, "cg_test_a"); + cg_test_b = cg_name(root, "cg_test_a/cg_test_b"); + cg_test_c = cg_name(root, "cg_test_a/cg_test_b/cg_test_c"); + cg_test_d = cg_name(root, "cg_test_a/cg_test_b/cg_test_d"); + + if (!cg_test_a || !cg_test_b || !cg_test_c || !cg_test_d) + goto cleanup; + + if (cg_create(cg_test_a)) + goto cleanup; + + if (cg_create(cg_test_b)) + goto cleanup; + + if (cg_create(cg_test_c)) + goto cleanup; + + if (cg_create(cg_test_d)) + goto cleanup; + + if (cg_enter_current(cg_test_c)) + goto cleanup; + + if (cg_read_strcmp(cg_test_a, "cgroup.events", "populated 1\n")) + goto cleanup; + + if (cg_read_strcmp(cg_test_b, "cgroup.events", "populated 1\n")) + goto cleanup; + + if (cg_read_strcmp(cg_test_c, "cgroup.events", "populated 1\n")) + goto cleanup; + + if (cg_read_strcmp(cg_test_d, "cgroup.events", "populated 0\n")) + goto cleanup; + + if (cg_enter_current(root)) + goto cleanup; + + if (cg_read_strcmp(cg_test_a, "cgroup.events", "populated 0\n")) + goto cleanup; + + if (cg_read_strcmp(cg_test_b, "cgroup.events", "populated 0\n")) + goto cleanup; + + if (cg_read_strcmp(cg_test_c, "cgroup.events", "populated 0\n")) + goto cleanup; + + if (cg_read_strcmp(cg_test_d, "cgroup.events", "populated 0\n")) + goto cleanup; + + ret = KSFT_PASS; + +cleanup: + if (cg_test_d) + cg_destroy(cg_test_d); + if (cg_test_c) + cg_destroy(cg_test_c); + if (cg_test_b) + cg_destroy(cg_test_b); + if (cg_test_a) + cg_destroy(cg_test_a); + free(cg_test_d); + free(cg_test_c); + free(cg_test_b); + free(cg_test_a); + return ret; +} + +/* + * A (domain threaded) - B (threaded) - C (domain) + * + * test that C can't be used until it is turned into a + * threaded cgroup. "cgroup.type" file will report "domain (invalid)" in + * these cases. Operations which fail due to invalid topology use + * EOPNOTSUPP as the errno. + */ +static int test_cgcore_invalid_domain(const char *root) +{ + int ret = KSFT_FAIL; + char *grandparent = NULL, *parent = NULL, *child = NULL; + + grandparent = cg_name(root, "cg_test_grandparent"); + parent = cg_name(root, "cg_test_grandparent/cg_test_parent"); + child = cg_name(root, "cg_test_grandparent/cg_test_parent/cg_test_child"); + if (!parent || !child || !grandparent) + goto cleanup; + + if (cg_create(grandparent)) + goto cleanup; + + if (cg_create(parent)) + goto cleanup; + + if (cg_create(child)) + goto cleanup; + + if (cg_write(parent, "cgroup.type", "threaded")) + goto cleanup; + + if (cg_read_strcmp(child, "cgroup.type", "domain invalid\n")) + goto cleanup; + + if (!cg_enter_current(child)) + goto cleanup; + + if (errno != EOPNOTSUPP) + goto cleanup; + + ret = KSFT_PASS; + +cleanup: + cg_enter_current(root); + if (child) + cg_destroy(child); + if (parent) + cg_destroy(parent); + if (grandparent) + cg_destroy(grandparent); + free(child); + free(parent); + free(grandparent); + return ret; +} + +/* + * Test that when a child becomes threaded + * the parent type becomes domain threaded. + */ +static int test_cgcore_parent_becomes_threaded(const char *root) +{ + int ret = KSFT_FAIL; + char *parent = NULL, *child = NULL; + + parent = cg_name(root, "cg_test_parent"); + child = cg_name(root, "cg_test_parent/cg_test_child"); + if (!parent || !child) + goto cleanup; + + if (cg_create(parent)) + goto cleanup; + + if (cg_create(child)) + goto cleanup; + + if (cg_write(child, "cgroup.type", "threaded")) + goto cleanup; + + if (cg_read_strcmp(parent, "cgroup.type", "domain threaded\n")) + goto cleanup; + + ret = KSFT_PASS; + +cleanup: + if (child) + cg_destroy(child); + if (parent) + cg_destroy(parent); + free(child); + free(parent); + return ret; + +} + +/* + * Test that there's no internal process constrain on threaded cgroups. + * You can add threads/processes on a parent with a controller enabled. + */ +static int test_cgcore_no_internal_process_constraint_on_threads(const char *root) +{ + int ret = KSFT_FAIL; + char *parent = NULL, *child = NULL; + + if (cg_read_strstr(root, "cgroup.controllers", "cpu") || + cg_read_strstr(root, "cgroup.subtree_control", "cpu")) { + ret = KSFT_SKIP; + goto cleanup; + } + + parent = cg_name(root, "cg_test_parent"); + child = cg_name(root, "cg_test_parent/cg_test_child"); + if (!parent || !child) + goto cleanup; + + if (cg_create(parent)) + goto cleanup; + + if (cg_create(child)) + goto cleanup; + + if (cg_write(parent, "cgroup.type", "threaded")) + goto cleanup; + + if (cg_write(child, "cgroup.type", "threaded")) + goto cleanup; + + if (cg_write(parent, "cgroup.subtree_control", "+cpu")) + goto cleanup; + + if (cg_enter_current(parent)) + goto cleanup; + + ret = KSFT_PASS; + +cleanup: + cg_enter_current(root); + cg_enter_current(root); + if (child) + cg_destroy(child); + if (parent) + cg_destroy(parent); + free(child); + free(parent); + return ret; +} + +/* + * Test that you can't enable a controller on a child if it's not enabled + * on the parent. + */ +static int test_cgcore_top_down_constraint_enable(const char *root) +{ + int ret = KSFT_FAIL; + char *parent = NULL, *child = NULL; + + parent = cg_name(root, "cg_test_parent"); + child = cg_name(root, "cg_test_parent/cg_test_child"); + if (!parent || !child) + goto cleanup; + + if (cg_create(parent)) + goto cleanup; + + if (cg_create(child)) + goto cleanup; + + if (!cg_write(child, "cgroup.subtree_control", "+memory")) + goto cleanup; + + ret = KSFT_PASS; + +cleanup: + if (child) + cg_destroy(child); + if (parent) + cg_destroy(parent); + free(child); + free(parent); + return ret; +} + +/* + * Test that you can't disable a controller on a parent + * if it's enabled in a child. + */ +static int test_cgcore_top_down_constraint_disable(const char *root) +{ + int ret = KSFT_FAIL; + char *parent = NULL, *child = NULL; + + parent = cg_name(root, "cg_test_parent"); + child = cg_name(root, "cg_test_parent/cg_test_child"); + if (!parent || !child) + goto cleanup; + + if (cg_create(parent)) + goto cleanup; + + if (cg_create(child)) + goto cleanup; + + if (cg_write(parent, "cgroup.subtree_control", "+memory")) + goto cleanup; + + if (cg_write(child, "cgroup.subtree_control", "+memory")) + goto cleanup; + + if (!cg_write(parent, "cgroup.subtree_control", "-memory")) + goto cleanup; + + ret = KSFT_PASS; + +cleanup: + if (child) + cg_destroy(child); + if (parent) + cg_destroy(parent); + free(child); + free(parent); + return ret; +} + +/* + * Test internal process constraint. + * You can't add a pid to a domain parent if a controller is enabled. + */ +static int test_cgcore_internal_process_constraint(const char *root) +{ + int ret = KSFT_FAIL; + char *parent = NULL, *child = NULL; + + parent = cg_name(root, "cg_test_parent"); + child = cg_name(root, "cg_test_parent/cg_test_child"); + if (!parent || !child) + goto cleanup; + + if (cg_create(parent)) + goto cleanup; + + if (cg_create(child)) + goto cleanup; + + if (cg_write(parent, "cgroup.subtree_control", "+memory")) + goto cleanup; + + if (!cg_enter_current(parent)) + goto cleanup; + + ret = KSFT_PASS; + +cleanup: + if (child) + cg_destroy(child); + if (parent) + cg_destroy(parent); + free(child); + free(parent); + return ret; +} + +#define T(x) { x, #x } +struct corecg_test { + int (*fn)(const char *root); + const char *name; +} tests[] = { + T(test_cgcore_internal_process_constraint), + T(test_cgcore_top_down_constraint_enable), + T(test_cgcore_top_down_constraint_disable), + T(test_cgcore_no_internal_process_constraint_on_threads), + T(test_cgcore_parent_becomes_threaded), + T(test_cgcore_invalid_domain), + T(test_cgcore_populated), +}; +#undef T + +int main(int argc, char *argv[]) +{ + char root[PATH_MAX]; + int i, ret = EXIT_SUCCESS; + + if (cg_find_unified_root(root, sizeof(root))) + ksft_exit_skip("cgroup v2 isn't mounted\n"); + for (i = 0; i < ARRAY_SIZE(tests); i++) { + switch (tests[i].fn(root)) { + case KSFT_PASS: + ksft_test_result_pass("%s\n", tests[i].name); + break; + case KSFT_SKIP: + ksft_test_result_skip("%s\n", tests[i].name); + break; + default: + ret = EXIT_FAILURE; + ksft_test_result_fail("%s\n", tests[i].name); + break; + } + } + + return ret; +}