From patchwork Fri Aug 4 09:38:03 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Chuyi Zhou X-Patchwork-Id: 13341584 Received: from lindbergh.monkeyblade.net (lindbergh.monkeyblade.net [23.128.96.19]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 905D1A94C for ; Fri, 4 Aug 2023 09:38:24 +0000 (UTC) Received: from mail-pf1-x436.google.com (mail-pf1-x436.google.com [IPv6:2607:f8b0:4864:20::436]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 1E40F30EB for ; Fri, 4 Aug 2023 02:38:23 -0700 (PDT) Received: by mail-pf1-x436.google.com with SMTP id d2e1a72fcca58-686bea20652so1785883b3a.1 for ; Fri, 04 Aug 2023 02:38:23 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=bytedance.com; s=google; t=1691141902; x=1691746702; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=ZnVxinOLSDbVFQcoUFkv5fif4kt7MH5yiiDW0k4Isz4=; b=SnYq7pN1xKt+8a/z1C+XFiuu2Ak1e5y+1Mhw4KDi+FAv5ci79AUGSFoPX+kdGaVg+F ljIZlgKY99++RUbA5xeIKLBsHHuwqQptGQ6yHf0MZ+5awXqsZx8LSY7v2Jeb702P0gGy aSNqUTc0vxHv3ZjhTdhYeX04dsH34jOpByFF8IwXYFZXrLOqi/Y1A0nxMr48P2Mked+W lRhz4HZu79RK7uZdHXLNBgFvOnMSLWkjTr8zn1WtUyr6KDzyoSKWOQMC6Rzh/DpUgiOJ HwIl5c3AZ3RSsOFH/IRN44JBRMBmMiVnj4F2MzK87P7LwMhl5vt40rFY7T6HqGGit5ae fJJQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1691141902; x=1691746702; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=ZnVxinOLSDbVFQcoUFkv5fif4kt7MH5yiiDW0k4Isz4=; b=kVII9/xiYgVXOacGTCXOKylw7XEJV8p0Ub8sF+dLpN1OElNzc+uQTw477gtjY+StT6 kz/2LMlPIyiabOpnEJ7DPFKhO50hv8b0MYlKFYsyH6+Z6/Sk/wAx58Ps6tLb5t4T4qZP NrPYtp+LQ7ROvPLtg3UCcDOD9plh5xF7E5rERMtR0CJ+ykRVrf3Q2ncFJkn47RJOc0bU gIb+0PdLtTqxHMl/J865HAsEAKLQ4iHdR4ILDVF8vQUCsJkuyYVy8V/qFKOO/2SLK/c8 GmuvhqmZ90cVlbtx5uzGBWpGy/0vF4vWp2lSMkTkB+qS2ZEr4TbAKX6oifyi8kerJ5Qo XIHQ== X-Gm-Message-State: AOJu0Yyygf7+fhKOKy/oQwB2uwRVMlIK6Csjj2lsRCrnIXzpVpHJx+Dh cgJ6EvUZKtFdv1lI/b0e/oH05SteOHR7EQOuKsU= X-Google-Smtp-Source: AGHT+IGEaGrIQSMdjLRn7q18dfEfbuGihr94UXKQ14pGWU3sdMqgmFT3kieIlgr3NWLWA4FqGTqF2Q== X-Received: by 2002:a05:6a21:789a:b0:133:7276:324b with SMTP id bf26-20020a056a21789a00b001337276324bmr1346719pzc.23.1691141902614; Fri, 04 Aug 2023 02:38:22 -0700 (PDT) Received: from n37-019-243.byted.org ([180.184.51.142]) by smtp.gmail.com with ESMTPSA id c23-20020aa78817000000b00687933946ddsm1214837pfo.23.2023.08.04.02.38.19 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 04 Aug 2023 02:38:22 -0700 (PDT) From: Chuyi Zhou To: hannes@cmpxchg.org, mhocko@kernel.org, roman.gushchin@linux.dev, ast@kernel.org, daniel@iogearbox.net, andrii@kernel.org, muchun.song@linux.dev Cc: bpf@vger.kernel.org, linux-kernel@vger.kernel.org, wuyun.abel@bytedance.com, robin.lu@bytedance.com, Chuyi Zhou Subject: [RFC PATCH 1/2] mm, oom: Introduce bpf_select_task Date: Fri, 4 Aug 2023 17:38:03 +0800 Message-Id: <20230804093804.47039-2-zhouchuyi@bytedance.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20230804093804.47039-1-zhouchuyi@bytedance.com> References: <20230804093804.47039-1-zhouchuyi@bytedance.com> Precedence: bulk X-Mailing-List: bpf@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-Spam-Status: No, score=-2.1 required=5.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,DKIM_VALID_EF,RCVD_IN_DNSWL_NONE, SPF_HELO_NONE,SPF_NONE,URIBL_BLOCKED autolearn=unavailable autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on lindbergh.monkeyblade.net X-Patchwork-State: RFC This patch adds a new hook bpf_select_task in oom_evaluate_task. It takes oc and current iterating task as parameters and returns a result indicating which one is selected by bpf program. Although bpf_select_task is used to bypass the default method, there are some existing rules should be obeyed. Specifically, we skip these "unkillable" tasks(e.g., kthread, MMF_OOM_SKIP, in_vfork()).So we do not consider tasks with lowest score returned by oom_badness except it was caused by OOM_SCORE_ADJ_MIN. If we attach a prog to the hook, the interface is enabled only when we have successfully chosen at least one valid candidate in previous iteraion. This is to avoid that we find nothing if bpf program rejects all tasks. Signed-off-by: Chuyi Zhou --- mm/oom_kill.c | 57 ++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 50 insertions(+), 7 deletions(-) diff --git a/mm/oom_kill.c b/mm/oom_kill.c index 612b5597d3af..aec4c55ed49a 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -18,6 +18,7 @@ * kernel subsystems and hints as to where to find out what things do. */ +#include #include #include #include @@ -210,6 +211,16 @@ long oom_badness(struct task_struct *p, unsigned long totalpages) if (!p) return LONG_MIN; + /* + * If task is allocating a lot of memory and has been marked to be + * killed first if it triggers an oom, then set points to LONG_MAX. + * It will be selected unless we keep oc->chosen through bpf interface. + */ + if (oom_task_origin(p)) { + task_unlock(p); + return LONG_MAX; + } + /* * Do not even consider tasks which are explicitly marked oom * unkillable or have been already oom reaped or the are in @@ -305,8 +316,30 @@ static enum oom_constraint constrained_alloc(struct oom_control *oc) return CONSTRAINT_NONE; } +enum bpf_select_ret { + BPF_SELECT_DISABLE, + BPF_SELECT_TASK, + BPF_SELECT_CHOSEN, +}; + +__weak noinline int bpf_select_task(struct oom_control *oc, + struct task_struct *task, long badness_points) +{ + return BPF_SELECT_DISABLE; +} + +BTF_SET8_START(oom_bpf_fmodret_ids) +BTF_ID_FLAGS(func, bpf_select_task) +BTF_SET8_END(oom_bpf_fmodret_ids) + +static const struct btf_kfunc_id_set oom_bpf_fmodret_set = { + .owner = THIS_MODULE, + .set = &oom_bpf_fmodret_ids, +}; + static int oom_evaluate_task(struct task_struct *task, void *arg) { + enum bpf_select_ret bpf_ret = BPF_SELECT_DISABLE; struct oom_control *oc = arg; long points; @@ -329,17 +362,23 @@ static int oom_evaluate_task(struct task_struct *task, void *arg) goto abort; } + points = oom_badness(task, oc->totalpages); + /* - * If task is allocating a lot of memory and has been marked to be - * killed first if it triggers an oom, then select it. + * Do not consider tasks with lowest score value except it was caused + * by OOM_SCORE_ADJ_MIN. Give these tasks a chance to be selected by + * bpf interface. */ - if (oom_task_origin(task)) { - points = LONG_MAX; + if (points == LONG_MIN && task->signal->oom_score_adj != OOM_SCORE_ADJ_MIN) + goto next; + + if (oc->chosen) + bpf_ret = bpf_select_task(oc, task, points); + + if (bpf_ret == BPF_SELECT_TASK) goto select; - } - points = oom_badness(task, oc->totalpages); - if (points == LONG_MIN || points < oc->chosen_points) + if (bpf_ret == BPF_SELECT_CHOSEN || points == LONG_MIN || points < oc->chosen_points) goto next; select: @@ -732,10 +771,14 @@ static struct ctl_table vm_oom_kill_table[] = { static int __init oom_init(void) { + int err; oom_reaper_th = kthread_run(oom_reaper, NULL, "oom_reaper"); #ifdef CONFIG_SYSCTL register_sysctl_init("vm", vm_oom_kill_table); #endif + err = register_btf_fmodret_id_set(&oom_bpf_fmodret_set); + if (err) + pr_warn("error while registering oom fmodret entrypoints: %d", err); return 0; } subsys_initcall(oom_init) From patchwork Fri Aug 4 09:38:04 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Chuyi Zhou X-Patchwork-Id: 13341585 X-Patchwork-Delegate: bpf@iogearbox.net Received: from lindbergh.monkeyblade.net (lindbergh.monkeyblade.net [23.128.96.19]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id EFEDAAD29 for ; Fri, 4 Aug 2023 09:38:29 +0000 (UTC) Received: from mail-pf1-x436.google.com (mail-pf1-x436.google.com [IPv6:2607:f8b0:4864:20::436]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 4A2DE30EB for ; Fri, 4 Aug 2023 02:38:28 -0700 (PDT) Received: by mail-pf1-x436.google.com with SMTP id d2e1a72fcca58-686ba29ccb1so1306452b3a.1 for ; Fri, 04 Aug 2023 02:38:28 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=bytedance.com; s=google; t=1691141908; x=1691746708; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=01G/dGmCoWPYOyR3VsPv/eL5+i89PsgHn5Vivri25As=; b=if8fIwb8lSueExAsbM1/AY31VsruqCA6SyziSadQfaBWq7gobGORNAvSo/zEC9G/B/ kx7iEKa2pPfVyH5IjW7i2e+sbal6F4+sIDqsBQ+zlCKBbhEJ9uh377bznGNXWviF0+tH 2kGckdyM1I5UV/+SApN/JG7mlMsv3FAcAid7Zt0pWw0ANaqIZDleN+JBQoyYev9okSe+ B68RzPb+ONsCtSqdCmgGDR//GVLTDlcw9xWlSLtTBW3hpgCXHXg+UvcPgAmhWfuA+Lmz X3+OukpfFFHT+4dWzM5c7Cc0Fm8HVIrE+M3PBAl6+wgMKQgTinPSh0Egt+VYrxtTQAxi ZxLQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1691141908; x=1691746708; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=01G/dGmCoWPYOyR3VsPv/eL5+i89PsgHn5Vivri25As=; b=DfbyevAxuW8euX8rT3MJd02oUPAzHmdzzCAUxvVE7BsdObAIkIleXQN1much/kPMae Ys5+H4tWl9dG8CKqTQ/WE1CVaDQhnUsQeCj+SlFqDQVx3zUStG3B80jLryQwEc5APSDa pjj4ctDWWhZotyqWH5urlMJFvStM96RLS4qbtqXUpvpwZpHtaJxefu+96wWfkluJW9C8 cvXMwn7U/IZjwPkkk3G5Ww39g5PH81/3XVycSE3gT39dpzU2xBMqCAkRzqhRXM8kJgyB 1WBUUfn95P4r61kLXyYSjlFoxMfbkwUdd+ayca+702rSW9wykCHWSGNBkCIc+2s/efdR SYeA== X-Gm-Message-State: AOJu0YxhODq0+/GtjEZcx8rGVIAYfH47ItcJt0xKio6BMQcuH0vtQDVH io9UnPte1IaveqYNPuyhtYimsA== X-Google-Smtp-Source: AGHT+IFjeFRo03zgYga5XtByDq/AZ59O6vv/ZtAYjs3lW/GmRpRrS/F+pY5M7Plb4ZjvmaErR7eQ8A== X-Received: by 2002:a05:6a20:3d84:b0:137:c971:6a0c with SMTP id s4-20020a056a203d8400b00137c9716a0cmr1405040pzi.31.1691141907829; Fri, 04 Aug 2023 02:38:27 -0700 (PDT) Received: from n37-019-243.byted.org ([180.184.51.142]) by smtp.gmail.com with ESMTPSA id c23-20020aa78817000000b00687933946ddsm1214837pfo.23.2023.08.04.02.38.24 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 04 Aug 2023 02:38:27 -0700 (PDT) From: Chuyi Zhou To: hannes@cmpxchg.org, mhocko@kernel.org, roman.gushchin@linux.dev, ast@kernel.org, daniel@iogearbox.net, andrii@kernel.org, muchun.song@linux.dev Cc: bpf@vger.kernel.org, linux-kernel@vger.kernel.org, wuyun.abel@bytedance.com, robin.lu@bytedance.com, Chuyi Zhou Subject: [RFC PATCH 2/2] bpf: Add OOM policy test Date: Fri, 4 Aug 2023 17:38:04 +0800 Message-Id: <20230804093804.47039-3-zhouchuyi@bytedance.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20230804093804.47039-1-zhouchuyi@bytedance.com> References: <20230804093804.47039-1-zhouchuyi@bytedance.com> Precedence: bulk X-Mailing-List: bpf@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-Spam-Status: No, score=-2.1 required=5.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,DKIM_VALID_EF,RCVD_IN_DNSWL_NONE, SPF_HELO_NONE,SPF_NONE,URIBL_BLOCKED autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on lindbergh.monkeyblade.net X-Patchwork-Delegate: bpf@iogearbox.net X-Patchwork-State: RFC This patch adds a test which implements a priority-based policy through bpf_select_task. The BPF program, oom_policy.c, compares the cgroup priority of two tasks and select the lower one. The userspace program test_oom_policy.c maintains a priority map by using cgroup id as the keys and priority as the values. We could protect certain cgroups from oom-killer by setting higher priority. Signed-off-by: Chuyi Zhou --- .../bpf/prog_tests/test_oom_policy.c | 140 ++++++++++++++++++ .../testing/selftests/bpf/progs/oom_policy.c | 77 ++++++++++ 2 files changed, 217 insertions(+) create mode 100644 tools/testing/selftests/bpf/prog_tests/test_oom_policy.c create mode 100644 tools/testing/selftests/bpf/progs/oom_policy.c diff --git a/tools/testing/selftests/bpf/prog_tests/test_oom_policy.c b/tools/testing/selftests/bpf/prog_tests/test_oom_policy.c new file mode 100644 index 000000000000..2400cc48ba83 --- /dev/null +++ b/tools/testing/selftests/bpf/prog_tests/test_oom_policy.c @@ -0,0 +1,140 @@ +// SPDX-License-Identifier: GPL-2.0-only +#define _GNU_SOURCE + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "cgroup_helpers.h" +#include "oom_policy.skel.h" + +static int map_fd; +static int cg_nr; +struct { + const char *path; + int fd; + unsigned long long id; +} cgs[] = { + { "/cg1" }, + { "/cg2" }, +}; + + +static struct oom_policy *open_load_oom_policy_skel(void) +{ + struct oom_policy *skel; + int err; + + skel = oom_policy__open(); + if (!ASSERT_OK_PTR(skel, "skel_open")) + return NULL; + + err = oom_policy__load(skel); + if (!ASSERT_OK(err, "skel_load")) + goto cleanup; + + return skel; + +cleanup: + oom_policy__destroy(skel); + return NULL; +} + +static void run_memory_consume(unsigned long long consume_size, int idx) +{ + char *buf; + + join_parent_cgroup(cgs[idx].path); + buf = malloc(consume_size); + memset(buf, 0, consume_size); + sleep(2); + exit(0); +} + +static int set_cgroup_prio(unsigned long long cg_id, int prio) +{ + int err; + + err = bpf_map_update_elem(map_fd, &cg_id, &prio, BPF_ANY); + ASSERT_EQ(err, 0, "update_map"); + return err; +} + +static int prepare_cgroup_environment(void) +{ + int err; + + err = setup_cgroup_environment(); + if (err) + goto clean_cg_env; + for (int i = 0; i < cg_nr; i++) { + err = cgs[i].fd = create_and_get_cgroup(cgs[i].path); + if (!ASSERT_GE(cgs[i].fd, 0, "cg_create")) + goto clean_cg_env; + cgs[i].id = get_cgroup_id(cgs[i].path); + } + return 0; +clean_cg_env: + cleanup_cgroup_environment(); + return err; +} + +void test_oom_policy(void) +{ + struct oom_policy *skel; + struct bpf_link *link; + int err; + int victim_pid; + unsigned long long victim_cg_id; + + link = NULL; + cg_nr = ARRAY_SIZE(cgs); + + skel = open_load_oom_policy_skel(); + err = oom_policy__attach(skel); + if (!ASSERT_OK(err, "oom_policy__attach")) + goto cleanup; + + map_fd = bpf_object__find_map_fd_by_name(skel->obj, "cg_map"); + if (!ASSERT_GE(map_fd, 0, "find map")) + goto cleanup; + + err = prepare_cgroup_environment(); + if (!ASSERT_EQ(err, 0, "prepare cgroup env")) + goto cleanup; + + write_cgroup_file("/", "memory.max", "10M"); + + /* + * Set higher priority to cg2 and lower to cg1, so we would select + * task under cg1 as victim.(see oom_policy.c) + */ + set_cgroup_prio(cgs[0].id, 10); + set_cgroup_prio(cgs[1].id, 50); + + victim_cg_id = cgs[0].id; + victim_pid = fork(); + + if (victim_pid == 0) + run_memory_consume(1024 * 1024 * 4, 0); + + if (fork() == 0) + run_memory_consume(1024 * 1024 * 8, 1); + + while (wait(NULL) > 0) + ; + + ASSERT_EQ(skel->bss->victim_pid, victim_pid, "victim_pid"); + ASSERT_EQ(skel->bss->victim_cg_id, victim_cg_id, "victim_cgid"); + +cleanup: + bpf_link__destroy(link); + oom_policy__destroy(skel); + cleanup_cgroup_environment(); +} diff --git a/tools/testing/selftests/bpf/progs/oom_policy.c b/tools/testing/selftests/bpf/progs/oom_policy.c new file mode 100644 index 000000000000..d269ea52bcb2 --- /dev/null +++ b/tools/testing/selftests/bpf/progs/oom_policy.c @@ -0,0 +1,77 @@ +// SPDX-License-Identifier: GPL-2.0-only +#include +#include +#include + +char _license[] SEC("license") = "GPL"; +struct { + __uint(type, BPF_MAP_TYPE_HASH); + __type(key, int); + __type(value, int); + __uint(max_entries, 24); +} cg_map SEC(".maps"); + +unsigned int victim_pid; +u64 victim_cg_id; + +enum bpf_select_ret { + BPF_SELECT_DISABLE, + BPF_SELECT_TASK, + BPF_SELECT_CHOSEN, +}; + +static __always_inline u64 task_cgroup_id(struct task_struct *task) +{ + struct kernfs_node *node; + struct task_group *tg; + + if (!task) + return 0; + + tg = task->sched_task_group; + node = tg->css.cgroup->kn; + + return node->id; +} + +SEC("fentry/oom_kill_process") +int BPF_PROG(oom_kill_process_k, struct oom_control *oc, const char *message) +{ + struct task_struct *victim = oc->chosen; + + if (!victim) + return 0; + + victim_pid = victim->pid; + victim_cg_id = task_cgroup_id(victim); + return 0; +} + +SEC("fmod_ret/bpf_select_task") +int BPF_PROG(select_task_test, struct oom_control *oc, struct task_struct *task, long points) +{ + u64 chosen_cg_id, task_cg_id; + int chosen_cg_prio, task_cg_prio; + struct task_struct *chosen; + int *val; + + chosen = oc->chosen; + chosen_cg_id = task_cgroup_id(chosen); + task_cg_id = task_cgroup_id(task); + chosen_cg_prio = task_cg_prio = 0; + val = bpf_map_lookup_elem(&cg_map, &chosen_cg_id); + if (val) + chosen_cg_prio = *val; + val = bpf_map_lookup_elem(&cg_map, &task_cg_id); + if (val) + task_cg_prio = *val; + + if (chosen_cg_prio > task_cg_prio) + return BPF_SELECT_TASK; + if (chosen_cg_prio < task_cg_prio) + return BPF_SELECT_CHOSEN; + + return BPF_SELECT_DISABLE; + +} +