diff options
Diffstat (limited to 'kernel/cgroup')
| -rw-r--r-- | kernel/cgroup/Makefile | 2 | ||||
| -rw-r--r-- | kernel/cgroup/cgroup-internal.h | 9 | ||||
| -rw-r--r-- | kernel/cgroup/cgroup-v1.c | 27 | ||||
| -rw-r--r-- | kernel/cgroup/cgroup.c | 389 | ||||
| -rw-r--r-- | kernel/cgroup/cpuset-internal.h | 306 | ||||
| -rw-r--r-- | kernel/cgroup/cpuset-v1.c | 611 | ||||
| -rw-r--r-- | kernel/cgroup/cpuset.c | 2161 | ||||
| -rw-r--r-- | kernel/cgroup/dmem.c | 829 | ||||
| -rw-r--r-- | kernel/cgroup/freezer.c | 97 | ||||
| -rw-r--r-- | kernel/cgroup/legacy_freezer.c | 11 | ||||
| -rw-r--r-- | kernel/cgroup/misc.c | 100 | ||||
| -rw-r--r-- | kernel/cgroup/pids.c | 141 | ||||
| -rw-r--r-- | kernel/cgroup/rstat.c | 616 |
13 files changed, 3404 insertions, 1895 deletions
diff --git a/kernel/cgroup/Makefile b/kernel/cgroup/Makefile index 12f8457ad1f9..ede31601a363 100644 --- a/kernel/cgroup/Makefile +++ b/kernel/cgroup/Makefile @@ -5,5 +5,7 @@ obj-$(CONFIG_CGROUP_FREEZER) += legacy_freezer.o obj-$(CONFIG_CGROUP_PIDS) += pids.o obj-$(CONFIG_CGROUP_RDMA) += rdma.o obj-$(CONFIG_CPUSETS) += cpuset.o +obj-$(CONFIG_CPUSETS_V1) += cpuset-v1.o obj-$(CONFIG_CGROUP_MISC) += misc.o +obj-$(CONFIG_CGROUP_DMEM) += dmem.o obj-$(CONFIG_CGROUP_DEBUG) += debug.o diff --git a/kernel/cgroup/cgroup-internal.h b/kernel/cgroup/cgroup-internal.h index 520b90dd97ec..b14e61c64a34 100644 --- a/kernel/cgroup/cgroup-internal.h +++ b/kernel/cgroup/cgroup-internal.h @@ -81,6 +81,8 @@ struct cgroup_file_ctx { struct { struct cgroup_pidlist *pidlist; } procs1; + + struct cgroup_of_peak peak; }; /* @@ -166,6 +168,7 @@ struct cgroup_mgctx { extern struct cgroup_subsys *cgroup_subsys[]; extern struct list_head cgroup_roots; +extern bool cgrp_dfl_visible; /* iterate across the hierarchies */ #define for_each_root(root) \ @@ -267,9 +270,9 @@ int cgroup_task_count(const struct cgroup *cgrp); /* * rstat.c */ -int cgroup_rstat_init(struct cgroup *cgrp); -void cgroup_rstat_exit(struct cgroup *cgrp); -void cgroup_rstat_boot(void); +int css_rstat_init(struct cgroup_subsys_state *css); +void css_rstat_exit(struct cgroup_subsys_state *css); +int ss_rstat_init(struct cgroup_subsys *ss); void cgroup_base_stat_cputime_show(struct seq_file *seq); /* diff --git a/kernel/cgroup/cgroup-v1.c b/kernel/cgroup/cgroup-v1.c index 520a11cb12f4..fa24c032ed6f 100644 --- a/kernel/cgroup/cgroup-v1.c +++ b/kernel/cgroup/cgroup-v1.c @@ -46,6 +46,12 @@ bool cgroup1_ssid_disabled(int ssid) return cgroup_no_v1_mask & (1 << ssid); } +static bool cgroup1_subsys_absent(struct cgroup_subsys *ss) +{ + /* Check also dfl_cftypes for file-less controllers, i.e. perf_event */ + return ss->legacy_cftypes == NULL && ss->dfl_cftypes; +} + /** * cgroup_attach_task_all - attach task 'tsk' to all cgroups of task 'from' * @from: attach to all cgroups of a given task @@ -667,6 +673,7 @@ struct cftype cgroup1_base_files[] = { int proc_cgroupstats_show(struct seq_file *m, void *v) { struct cgroup_subsys *ss; + bool cgrp_v1_visible = false; int i; seq_puts(m, "#subsys_name\thierarchy\tnum_cgroups\tenabled\n"); @@ -675,11 +682,20 @@ int proc_cgroupstats_show(struct seq_file *m, void *v) * cgroup_mutex contention. */ - for_each_subsys(ss, i) + for_each_subsys(ss, i) { + if (cgroup1_subsys_absent(ss)) + continue; + cgrp_v1_visible |= ss->root != &cgrp_dfl_root; + seq_printf(m, "%s\t%d\t%d\t%d\n", ss->legacy_name, ss->root->hierarchy_id, atomic_read(&ss->root->nr_cgrps), cgroup_ssid_enabled(i)); + } + + if (cgrp_dfl_visible && !cgrp_v1_visible) + pr_info_once("/proc/cgroups lists only v1 controllers, use cgroup.controllers of root cgroup for v2 info\n"); + return 0; } @@ -835,7 +851,7 @@ static int cgroup1_rename(struct kernfs_node *kn, struct kernfs_node *new_parent if (kernfs_type(kn) != KERNFS_DIR) return -ENOTDIR; - if (kn->parent != new_parent) + if (rcu_access_pointer(kn->__parent) != new_parent) return -EIO; /* @@ -932,7 +948,8 @@ int cgroup1_parse_param(struct fs_context *fc, struct fs_parameter *param) if (ret != -ENOPARAM) return ret; for_each_subsys(ss, i) { - if (strcmp(param->key, ss->legacy_name)) + if (strcmp(param->key, ss->legacy_name) || + cgroup1_subsys_absent(ss)) continue; if (!cgroup_ssid_enabled(i) || cgroup1_ssid_disabled(i)) return invalfc(fc, "Disabled controller '%s'", @@ -1024,7 +1041,8 @@ static int check_cgroupfs_options(struct fs_context *fc) mask = ~((u16)1 << cpuset_cgrp_id); #endif for_each_subsys(ss, i) - if (cgroup_ssid_enabled(i) && !cgroup1_ssid_disabled(i)) + if (cgroup_ssid_enabled(i) && !cgroup1_ssid_disabled(i) && + !cgroup1_subsys_absent(ss)) enabled |= 1 << i; ctx->subsys_mask &= enabled; @@ -1335,6 +1353,7 @@ static int __init cgroup_no_v1(char *str) continue; cgroup_no_v1_mask |= 1 << i; + break; } } return 1; diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c index a66c088c851c..a723b7dc6e4e 100644 --- a/kernel/cgroup/cgroup.c +++ b/kernel/cgroup/cgroup.c @@ -90,11 +90,14 @@ DEFINE_MUTEX(cgroup_mutex); DEFINE_SPINLOCK(css_set_lock); -#ifdef CONFIG_PROVE_RCU +#if (defined CONFIG_PROVE_RCU || defined CONFIG_LOCKDEP) EXPORT_SYMBOL_GPL(cgroup_mutex); EXPORT_SYMBOL_GPL(css_set_lock); #endif +struct blocking_notifier_head cgroup_lifetime_notifier = + BLOCKING_NOTIFIER_INIT(cgroup_lifetime_notifier); + DEFINE_SPINLOCK(trace_cgroup_path_lock); char trace_cgroup_path[TRACE_CGROUP_PATH_LEN]; static bool cgroup_debug __read_mostly; @@ -161,17 +164,21 @@ static struct static_key_true *cgroup_subsys_on_dfl_key[] = { }; #undef SUBSYS -static DEFINE_PER_CPU(struct cgroup_rstat_cpu, cgrp_dfl_root_rstat_cpu); +static DEFINE_PER_CPU(struct css_rstat_cpu, root_rstat_cpu); +static DEFINE_PER_CPU(struct cgroup_rstat_base_cpu, root_rstat_base_cpu); /* the default hierarchy */ -struct cgroup_root cgrp_dfl_root = { .cgrp.rstat_cpu = &cgrp_dfl_root_rstat_cpu }; +struct cgroup_root cgrp_dfl_root = { + .cgrp.self.rstat_cpu = &root_rstat_cpu, + .cgrp.rstat_base_cpu = &root_rstat_base_cpu, +}; EXPORT_SYMBOL_GPL(cgrp_dfl_root); /* * The default hierarchy always exists but is hidden until mounted for the * first time. This is for backward compatibility. */ -static bool cgrp_dfl_visible; +bool cgrp_dfl_visible; /* some controllers are not supported in the default hierarchy */ static u16 cgrp_dfl_inhibit_ss_mask; @@ -633,9 +640,22 @@ int cgroup_task_count(const struct cgroup *cgrp) return count; } +static struct cgroup *kn_priv(struct kernfs_node *kn) +{ + struct kernfs_node *parent; + /* + * The parent can not be replaced due to KERNFS_ROOT_INVARIANT_PARENT. + * Therefore it is always safe to dereference this pointer outside of a + * RCU section. + */ + parent = rcu_dereference_check(kn->__parent, + kernfs_root_flags(kn) & KERNFS_ROOT_INVARIANT_PARENT); + return parent->priv; +} + struct cgroup_subsys_state *of_css(struct kernfs_open_file *of) { - struct cgroup *cgrp = of->kn->parent->priv; + struct cgroup *cgrp = kn_priv(of->kn); struct cftype *cft = of_cft(of); /* @@ -1322,6 +1342,7 @@ static void cgroup_destroy_root(struct cgroup_root *root) { struct cgroup *cgrp = &root->cgrp; struct cgrp_cset_link *link, *tmp_link; + int ret; trace_cgroup_destroy_root(root); @@ -1330,6 +1351,10 @@ static void cgroup_destroy_root(struct cgroup_root *root) BUG_ON(atomic_read(&root->nr_cgrps)); BUG_ON(!list_empty(&cgrp->self.children)); + ret = blocking_notifier_call_chain(&cgroup_lifetime_notifier, + CGROUP_LIFETIME_OFFLINE, cgrp); + WARN_ON_ONCE(notifier_to_errno(ret)); + /* Rebind all subsystems back to the default hierarchy */ WARN_ON(rebind_subsystems(&cgrp_dfl_root, root->subsys_mask)); @@ -1358,7 +1383,6 @@ static void cgroup_destroy_root(struct cgroup_root *root) cgroup_unlock(); - cgroup_rstat_exit(cgrp); kernfs_destroy_root(root->kf_root); cgroup_free_root(root); } @@ -1612,7 +1636,7 @@ void cgroup_kn_unlock(struct kernfs_node *kn) if (kernfs_type(kn) == KERNFS_DIR) cgrp = kn->priv; else - cgrp = kn->parent->priv; + cgrp = kn_priv(kn); cgroup_unlock(); @@ -1644,7 +1668,7 @@ struct cgroup *cgroup_kn_lock_live(struct kernfs_node *kn, bool drain_offline) if (kernfs_type(kn) == KERNFS_DIR) cgrp = kn->priv; else - cgrp = kn->parent->priv; + cgrp = kn_priv(kn); /* * We're gonna grab cgroup_mutex which nests outside kernfs @@ -1682,7 +1706,7 @@ static void cgroup_rm_file(struct cgroup *cgrp, const struct cftype *cft) cfile->kn = NULL; spin_unlock_irq(&cgroup_file_kn_lock); - del_timer_sync(&cfile->notify_timer); + timer_delete_sync(&cfile->notify_timer); } kernfs_remove_by_name(cgrp->kn, cgroup_file_name(cgrp, cft, name)); @@ -1702,7 +1726,7 @@ static void css_clear_dir(struct cgroup_subsys_state *css) css->flags &= ~CSS_VISIBLE; - if (!css->ss) { + if (css_is_self(css)) { if (cgroup_on_dfl(cgrp)) { cgroup_addrm_files(css, cgrp, cgroup_base_files, false); @@ -1734,7 +1758,7 @@ static int css_populate_dir(struct cgroup_subsys_state *css) if (css->flags & CSS_VISIBLE) return 0; - if (!css->ss) { + if (css_is_self(css)) { if (cgroup_on_dfl(cgrp)) { ret = cgroup_addrm_files(css, cgrp, cgroup_base_files, true); @@ -1744,8 +1768,11 @@ static int css_populate_dir(struct cgroup_subsys_state *css) if (cgroup_psi_enabled()) { ret = cgroup_addrm_files(css, cgrp, cgroup_psi_files, true); - if (ret < 0) + if (ret < 0) { + cgroup_addrm_files(css, cgrp, + cgroup_base_files, false); return ret; + } } } else { ret = cgroup_addrm_files(css, cgrp, @@ -1839,9 +1866,9 @@ int rebind_subsystems(struct cgroup_root *dst_root, u16 ss_mask) RCU_INIT_POINTER(scgrp->subsys[ssid], NULL); rcu_assign_pointer(dcgrp->subsys[ssid], css); ss->root = dst_root; - css->cgroup = dcgrp; spin_lock_irq(&css_set_lock); + css->cgroup = dcgrp; WARN_ON(!list_empty(&dcgrp->e_csets[ss->id])); list_for_each_entry_safe(cset, cset_pos, &scgrp->e_csets[ss->id], e_cset_node[ss->id]) { @@ -1860,13 +1887,6 @@ int rebind_subsystems(struct cgroup_root *dst_root, u16 ss_mask) } spin_unlock_irq(&css_set_lock); - if (ss->css_rstat_flush) { - list_del_rcu(&css->rstat_css_node); - synchronize_rcu(); - list_add_rcu(&css->rstat_css_node, - &dcgrp->rstat_css_list); - } - /* default hierarchy doesn't enable controllers by default */ dst_root->subsys_mask |= 1 << ssid; if (dst_root == &cgrp_dfl_root) { @@ -1922,6 +1942,7 @@ enum cgroup2_param { Opt_memory_localevents, Opt_memory_recursiveprot, Opt_memory_hugetlb_accounting, + Opt_pids_localevents, nr__cgroup2_params }; @@ -1931,6 +1952,7 @@ static const struct fs_parameter_spec cgroup2_fs_parameters[] = { fsparam_flag("memory_localevents", Opt_memory_localevents), fsparam_flag("memory_recursiveprot", Opt_memory_recursiveprot), fsparam_flag("memory_hugetlb_accounting", Opt_memory_hugetlb_accounting), + fsparam_flag("pids_localevents", Opt_pids_localevents), {} }; @@ -1960,10 +1982,20 @@ static int cgroup2_parse_param(struct fs_context *fc, struct fs_parameter *param case Opt_memory_hugetlb_accounting: ctx->flags |= CGRP_ROOT_MEMORY_HUGETLB_ACCOUNTING; return 0; + case Opt_pids_localevents: + ctx->flags |= CGRP_ROOT_PIDS_LOCAL_EVENTS; + return 0; } return -EINVAL; } +struct cgroup_of_peak *of_peak(struct kernfs_open_file *of) +{ + struct cgroup_file_ctx *ctx = of->priv; + + return &ctx->peak; +} + static void apply_cgroup_root_flags(unsigned int root_flags) { if (current->nsproxy->cgroup_ns == &init_cgroup_ns) { @@ -1989,6 +2021,11 @@ static void apply_cgroup_root_flags(unsigned int root_flags) cgrp_dfl_root.flags |= CGRP_ROOT_MEMORY_HUGETLB_ACCOUNTING; else cgrp_dfl_root.flags &= ~CGRP_ROOT_MEMORY_HUGETLB_ACCOUNTING; + + if (root_flags & CGRP_ROOT_PIDS_LOCAL_EVENTS) + cgrp_dfl_root.flags |= CGRP_ROOT_PIDS_LOCAL_EVENTS; + else + cgrp_dfl_root.flags &= ~CGRP_ROOT_PIDS_LOCAL_EVENTS; } } @@ -2004,6 +2041,8 @@ static int cgroup_show_options(struct seq_file *seq, struct kernfs_root *kf_root seq_puts(seq, ",memory_recursiveprot"); if (cgrp_dfl_root.flags & CGRP_ROOT_MEMORY_HUGETLB_ACCOUNTING) seq_puts(seq, ",memory_hugetlb_accounting"); + if (cgrp_dfl_root.flags & CGRP_ROOT_PIDS_LOCAL_EVENTS) + seq_puts(seq, ",pids_localevents"); return 0; } @@ -2030,7 +2069,6 @@ static void init_cgroup_housekeeping(struct cgroup *cgrp) cgrp->dom_cgrp = cgrp; cgrp->max_descendants = INT_MAX; cgrp->max_depth = INT_MAX; - INIT_LIST_HEAD(&cgrp->rstat_css_list); prev_cputime_init(&cgrp->prev_cputime); for_each_subsys(ss, ssid) @@ -2096,7 +2134,8 @@ int cgroup_setup_root(struct cgroup_root *root, u16 ss_mask) root->kf_root = kernfs_create_root(kf_sops, KERNFS_ROOT_CREATE_DEACTIVATED | KERNFS_ROOT_SUPPORT_EXPORTOP | - KERNFS_ROOT_SUPPORT_USER_XATTR, + KERNFS_ROOT_SUPPORT_USER_XATTR | + KERNFS_ROOT_INVARIANT_PARENT, root_cgrp); if (IS_ERR(root->kf_root)) { ret = PTR_ERR(root->kf_root); @@ -2110,7 +2149,7 @@ int cgroup_setup_root(struct cgroup_root *root, u16 ss_mask) if (ret) goto destroy_root; - ret = cgroup_rstat_init(root_cgrp); + ret = css_rstat_init(&root_cgrp->self); if (ret) goto destroy_root; @@ -2118,8 +2157,9 @@ int cgroup_setup_root(struct cgroup_root *root, u16 ss_mask) if (ret) goto exit_stats; - ret = cgroup_bpf_inherit(root_cgrp); - WARN_ON_ONCE(ret); + ret = blocking_notifier_call_chain(&cgroup_lifetime_notifier, + CGROUP_LIFETIME_ONLINE, root_cgrp); + WARN_ON_ONCE(notifier_to_errno(ret)); trace_cgroup_setup_root(root); @@ -2150,7 +2190,7 @@ int cgroup_setup_root(struct cgroup_root *root, u16 ss_mask) goto out; exit_stats: - cgroup_rstat_exit(root_cgrp); + css_rstat_exit(&root_cgrp->self); destroy_root: kernfs_destroy_root(root->kf_root); root->kf_root = NULL; @@ -2292,10 +2332,8 @@ static void cgroup_kill_sb(struct super_block *sb) * And don't kill the default root. */ if (list_empty(&root->cgrp.self.children) && root != &cgrp_dfl_root && - !percpu_ref_is_dying(&root->cgrp.self.refcnt)) { - cgroup_bpf_offline(&root->cgrp); + !percpu_ref_is_dying(&root->cgrp.self.refcnt)) percpu_ref_kill(&root->cgrp.self.refcnt); - } cgroup_put(&root->cgrp); kernfs_kill_sb(sb); } @@ -2316,10 +2354,38 @@ static struct file_system_type cgroup2_fs_type = { .fs_flags = FS_USERNS_MOUNT, }; -#ifdef CONFIG_CPUSETS +#ifdef CONFIG_CPUSETS_V1 +enum cpuset_param { + Opt_cpuset_v2_mode, +}; + +static const struct fs_parameter_spec cpuset_fs_parameters[] = { + fsparam_flag ("cpuset_v2_mode", Opt_cpuset_v2_mode), + {} +}; + +static int cpuset_parse_param(struct fs_context *fc, struct fs_parameter *param) +{ + struct cgroup_fs_context *ctx = cgroup_fc2context(fc); + struct fs_parse_result result; + int opt; + + opt = fs_parse(fc, cpuset_fs_parameters, param, &result); + if (opt < 0) + return opt; + + switch (opt) { + case Opt_cpuset_v2_mode: + ctx->flags |= CGRP_ROOT_CPUSET_V2_MODE; + return 0; + } + return -EINVAL; +} + static const struct fs_context_operations cpuset_fs_context_ops = { .get_tree = cgroup1_get_tree, .free = cgroup_fs_context_free, + .parse_param = cpuset_parse_param, }; /* @@ -2356,6 +2422,7 @@ static int cpuset_init_fs_context(struct fs_context *fc) static struct file_system_type cpuset_fs_type = { .name = "cpuset", .init_fs_context = cpuset_init_fs_context, + .parameters = cpuset_fs_parameters, .fs_flags = FS_USERNS_MOUNT, }; #endif @@ -3654,12 +3721,40 @@ static int cgroup_events_show(struct seq_file *seq, void *v) static int cgroup_stat_show(struct seq_file *seq, void *v) { struct cgroup *cgroup = seq_css(seq)->cgroup; + struct cgroup_subsys_state *css; + int dying_cnt[CGROUP_SUBSYS_COUNT]; + int ssid; seq_printf(seq, "nr_descendants %d\n", cgroup->nr_descendants); + + /* + * Show the number of live and dying csses associated with each of + * non-inhibited cgroup subsystems that is bound to cgroup v2. + * + * Without proper lock protection, racing is possible. So the + * numbers may not be consistent when that happens. + */ + rcu_read_lock(); + for (ssid = 0; ssid < CGROUP_SUBSYS_COUNT; ssid++) { + dying_cnt[ssid] = -1; + if ((BIT(ssid) & cgrp_dfl_inhibit_ss_mask) || + (cgroup_subsys[ssid]->root != &cgrp_dfl_root)) + continue; + css = rcu_dereference_raw(cgroup->subsys[ssid]); + dying_cnt[ssid] = cgroup->nr_dying_subsys[ssid]; + seq_printf(seq, "nr_subsys_%s %d\n", cgroup_subsys[ssid]->name, + css ? (css->nr_descendants + 1) : 0); + } + seq_printf(seq, "nr_dying_descendants %d\n", cgroup->nr_dying_descendants); - + for (ssid = 0; ssid < CGROUP_SUBSYS_COUNT; ssid++) { + if (dying_cnt[ssid] >= 0) + seq_printf(seq, "nr_dying_subsys_%s %d\n", + cgroup_subsys[ssid]->name, dying_cnt[ssid]); + } + rcu_read_unlock(); return 0; } @@ -3963,7 +4058,7 @@ static void __cgroup_kill(struct cgroup *cgrp) lockdep_assert_held(&cgroup_mutex); spin_lock_irq(&css_set_lock); - set_bit(CGRP_KILL, &cgrp->flags); + cgrp->kill_seq++; spin_unlock_irq(&css_set_lock); css_task_iter_start(&cgrp->self, CSS_TASK_ITER_PROCS | CSS_TASK_ITER_THREADED, &it); @@ -3979,10 +4074,6 @@ static void __cgroup_kill(struct cgroup *cgrp) send_sig(SIGKILL, task, 0); } css_task_iter_end(&it); - - spin_lock_irq(&css_set_lock); - clear_bit(CGRP_KILL, &cgrp->flags); - spin_unlock_irq(&css_set_lock); } static void cgroup_kill(struct cgroup *cgrp) @@ -4069,7 +4160,7 @@ static ssize_t cgroup_file_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) { struct cgroup_file_ctx *ctx = of->priv; - struct cgroup *cgrp = of->kn->parent->priv; + struct cgroup *cgrp = kn_priv(of->kn); struct cftype *cft = of_cft(of); struct cgroup_subsys_state *css; int ret; @@ -4081,7 +4172,7 @@ static ssize_t cgroup_file_write(struct kernfs_open_file *of, char *buf, * If namespaces are delegation boundaries, disallow writes to * files in an non-init namespace root from inside the namespace * except for the files explicitly marked delegatable - - * cgroup.procs and cgroup.subtree_control. + * eg. cgroup.procs, cgroup.threads and cgroup.subtree_control. */ if ((cgrp->root->flags & CGRP_ROOT_NS_DELEGATE) && !(cft->flags & CFTYPE_NS_DELEGATABLE) && @@ -4401,7 +4492,7 @@ int cgroup_rm_cftypes(struct cftype *cfts) * function currently returns 0 as long as @cfts registration is successful * even if some file creation attempts on existing cgroups fail. */ -static int cgroup_add_cftypes(struct cgroup_subsys *ss, struct cftype *cfts) +int cgroup_add_cftypes(struct cgroup_subsys *ss, struct cftype *cfts) { int ret; @@ -4580,8 +4671,9 @@ struct cgroup_subsys_state *css_next_child(struct cgroup_subsys_state *pos, * * While this function requires cgroup_mutex or RCU read locking, it * doesn't require the whole traversal to be contained in a single critical - * section. This function will return the correct next descendant as long - * as both @pos and @root are accessible and @pos is a descendant of @root. + * section. Additionally, it isn't necessary to hold onto a reference to @pos. + * This function will return the correct next descendant as long as both @pos + * and @root are accessible and @pos is a descendant of @root. * * If a subsystem synchronizes ->css_online() and the start of iteration, a * css which finished ->css_online() is guaranteed to be visible in the @@ -4629,8 +4721,9 @@ EXPORT_SYMBOL_GPL(css_next_descendant_pre); * * While this function requires cgroup_mutex or RCU read locking, it * doesn't require the whole traversal to be contained in a single critical - * section. This function will return the correct rightmost descendant as - * long as @pos is accessible. + * section. Additionally, it isn't necessary to hold onto a reference to @pos. + * This function will return the correct rightmost descendant as long as @pos + * is accessible. */ struct cgroup_subsys_state * css_rightmost_descendant(struct cgroup_subsys_state *pos) @@ -4674,9 +4767,9 @@ css_leftmost_descendant(struct cgroup_subsys_state *pos) * * While this function requires cgroup_mutex or RCU read locking, it * doesn't require the whole traversal to be contained in a single critical - * section. This function will return the correct next descendant as long - * as both @pos and @cgroup are accessible and @pos is a descendant of - * @cgroup. + * section. Additionally, it isn't necessary to hold onto a reference to @pos. + * This function will return the correct next descendant as long as both @pos + * and @cgroup are accessible and @pos is a descendant of @cgroup. * * If a subsystem synchronizes ->css_online() and the start of iteration, a * css which finished ->css_online() is guaranteed to be visible in the @@ -5353,8 +5446,9 @@ static void css_free_rwork_fn(struct work_struct *work) struct cgroup *cgrp = css->cgroup; percpu_ref_exit(&css->refcnt); + css_rstat_exit(css); - if (ss) { + if (!css_is_self(css)) { /* css free path */ struct cgroup_subsys_state *parent = css->parent; int id = css->id; @@ -5368,7 +5462,8 @@ static void css_free_rwork_fn(struct work_struct *work) } else { /* cgroup free path */ atomic_dec(&cgrp->root->nr_cgrps); - cgroup1_pidlist_destroy_all(cgrp); + if (!cgroup_on_dfl(cgrp)) + cgroup1_pidlist_destroy_all(cgrp); cancel_work_sync(&cgrp->release_agent_work); bpf_cgrp_storage_free(cgrp); @@ -5382,7 +5477,6 @@ static void css_free_rwork_fn(struct work_struct *work) cgroup_put(cgroup_parent(cgrp)); kernfs_put(cgrp->kn); psi_cgroup_free(cgrp); - cgroup_rstat_exit(cgrp); kfree(cgrp); } else { /* @@ -5407,23 +5501,36 @@ static void css_release_work_fn(struct work_struct *work) css->flags |= CSS_RELEASED; list_del_rcu(&css->sibling); - if (ss) { - /* css release path */ - if (!list_empty(&css->rstat_css_node)) { - cgroup_rstat_flush(cgrp); - list_del_rcu(&css->rstat_css_node); - } + if (!css_is_self(css)) { + struct cgroup *parent_cgrp; + + css_rstat_flush(css); cgroup_idr_replace(&ss->css_idr, NULL, css->id); if (ss->css_released) ss->css_released(css); + + cgrp->nr_dying_subsys[ss->id]--; + /* + * When a css is released and ready to be freed, its + * nr_descendants must be zero. However, the corresponding + * cgrp->nr_dying_subsys[ss->id] may not be 0 if a subsystem + * is activated and deactivated multiple times with one or + * more of its previous activation leaving behind dying csses. + */ + WARN_ON_ONCE(css->nr_descendants); + parent_cgrp = cgroup_parent(cgrp); + while (parent_cgrp) { + parent_cgrp->nr_dying_subsys[ss->id]--; + parent_cgrp = cgroup_parent(parent_cgrp); + } } else { struct cgroup *tcgrp; /* cgroup release path */ TRACE_CGROUP_PATH(release, cgrp); - cgroup_rstat_flush(cgrp); + css_rstat_flush(&cgrp->self); spin_lock_irq(&css_set_lock); for (tcgrp = cgroup_parent(cgrp); tcgrp; @@ -5471,7 +5578,6 @@ static void init_and_link_css(struct cgroup_subsys_state *css, css->id = -1; INIT_LIST_HEAD(&css->sibling); INIT_LIST_HEAD(&css->children); - INIT_LIST_HEAD(&css->rstat_css_node); css->serial_nr = css_serial_nr_next++; atomic_set(&css->online_cnt, 0); @@ -5480,9 +5586,6 @@ static void init_and_link_css(struct cgroup_subsys_state *css, css_get(css->parent); } - if (ss->css_rstat_flush) - list_add_rcu(&css->rstat_css_node, &cgrp->rstat_css_list); - BUG_ON(cgroup_css(cgrp, ss)); } @@ -5501,8 +5604,11 @@ static int online_css(struct cgroup_subsys_state *css) rcu_assign_pointer(css->cgroup->subsys[ss->id], css); atomic_inc(&css->online_cnt); - if (css->parent) + if (css->parent) { atomic_inc(&css->parent->online_cnt); + while ((css = css->parent)) + css->nr_descendants++; + } } return ret; } @@ -5524,6 +5630,16 @@ static void offline_css(struct cgroup_subsys_state *css) RCU_INIT_POINTER(css->cgroup->subsys[ss->id], NULL); wake_up_all(&css->cgroup->offline_waitq); + + css->cgroup->nr_dying_subsys[ss->id]++; + /* + * Parent css and cgroup cannot be freed until after the freeing + * of child css, see css_free_rwork_fn(). + */ + while ((css = css->parent)) { + css->nr_descendants--; + css->cgroup->nr_dying_subsys[ss->id]++; + } } /** @@ -5562,6 +5678,10 @@ static struct cgroup_subsys_state *css_create(struct cgroup *cgrp, goto err_free_css; css->id = err; + err = css_rstat_init(css); + if (err) + goto err_free_css; + /* @css is ready to be brought online now, make it visible */ list_add_tail_rcu(&css->sibling, &parent_css->children); cgroup_idr_replace(&ss->css_idr, css, css->id); @@ -5575,7 +5695,6 @@ static struct cgroup_subsys_state *css_create(struct cgroup *cgrp, err_list_del: list_del_rcu(&css->sibling); err_free_css: - list_del_rcu(&css->rstat_css_node); INIT_RCU_WORK(&css->destroy_rwork, css_free_rwork_fn); queue_rcu_work(cgroup_destroy_wq, &css->destroy_rwork); return ERR_PTR(err); @@ -5591,7 +5710,7 @@ static struct cgroup *cgroup_create(struct cgroup *parent, const char *name, struct cgroup_root *root = parent->root; struct cgroup *cgrp, *tcgrp; struct kernfs_node *kn; - int level = parent->level + 1; + int i, level = parent->level + 1; int ret; /* allocate the cgroup and its ID, 0 is reserved for the root */ @@ -5603,17 +5722,13 @@ static struct cgroup *cgroup_create(struct cgroup *parent, const char *name, if (ret) goto out_free_cgrp; - ret = cgroup_rstat_init(cgrp); - if (ret) - goto out_cancel_ref; - /* create the directory */ kn = kernfs_create_dir_ns(parent->kn, name, mode, current_fsuid(), current_fsgid(), cgrp, NULL); if (IS_ERR(kn)) { ret = PTR_ERR(kn); - goto out_stat_exit; + goto out_cancel_ref; } cgrp->kn = kn; @@ -5623,13 +5738,20 @@ static struct cgroup *cgroup_create(struct cgroup *parent, const char *name, cgrp->root = root; cgrp->level = level; - ret = psi_cgroup_alloc(cgrp); + /* + * Now that init_cgroup_housekeeping() has been called and cgrp->self + * is setup, it is safe to perform rstat initialization on it. + */ + ret = css_rstat_init(&cgrp->self); if (ret) goto out_kernfs_remove; - ret = cgroup_bpf_inherit(cgrp); + ret = psi_cgroup_alloc(cgrp); if (ret) - goto out_psi_free; + goto out_stat_exit; + + for (tcgrp = cgrp; tcgrp; tcgrp = cgroup_parent(tcgrp)) + cgrp->ancestors[tcgrp->level] = tcgrp; /* * New cgroup inherits effective freeze counter, and @@ -5647,24 +5769,6 @@ static struct cgroup *cgroup_create(struct cgroup *parent, const char *name, set_bit(CGRP_FROZEN, &cgrp->flags); } - spin_lock_irq(&css_set_lock); - for (tcgrp = cgrp; tcgrp; tcgrp = cgroup_parent(tcgrp)) { - cgrp->ancestors[tcgrp->level] = tcgrp; - - if (tcgrp != cgrp) { - tcgrp->nr_descendants++; - - /* - * If the new cgroup is frozen, all ancestor cgroups - * get a new frozen descendant, but their state can't - * change because of this. - */ - if (cgrp->freezer.e_freeze) - tcgrp->freezer.nr_frozen_descendants++; - } - } - spin_unlock_irq(&css_set_lock); - if (notify_on_release(parent)) set_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags); @@ -5673,7 +5777,29 @@ static struct cgroup *cgroup_create(struct cgroup *parent, const char *name, cgrp->self.serial_nr = css_serial_nr_next++; + ret = blocking_notifier_call_chain_robust(&cgroup_lifetime_notifier, + CGROUP_LIFETIME_ONLINE, + CGROUP_LIFETIME_OFFLINE, cgrp); + ret = notifier_to_errno(ret); + if (ret) + goto out_psi_free; + /* allocation complete, commit to creation */ + spin_lock_irq(&css_set_lock); + for (i = 0; i < level; i++) { + tcgrp = cgrp->ancestors[i]; + tcgrp->nr_descendants++; + + /* + * If the new cgroup is frozen, all ancestor cgroups get a new + * frozen descendant, but their state can't change because of + * this. + */ + if (cgrp->freezer.e_freeze) + tcgrp->freezer.nr_frozen_descendants++; + } + spin_unlock_irq(&css_set_lock); + list_add_tail_rcu(&cgrp->self.sibling, &cgroup_parent(cgrp)->self.children); atomic_inc(&root->nr_cgrps); cgroup_get_live(parent); @@ -5691,10 +5817,10 @@ static struct cgroup *cgroup_create(struct cgroup *parent, const char *name, out_psi_free: psi_cgroup_free(cgrp); +out_stat_exit: + css_rstat_exit(&cgrp->self); out_kernfs_remove: kernfs_remove(cgrp->kn); -out_stat_exit: - cgroup_rstat_exit(cgrp); out_cancel_ref: percpu_ref_exit(&cgrp->self.refcnt); out_free_cgrp: @@ -5706,7 +5832,7 @@ static bool cgroup_check_hierarchy_limits(struct cgroup *parent) { struct cgroup *cgroup; int ret = false; - int level = 1; + int level = 0; lockdep_assert_held(&cgroup_mutex); @@ -5714,7 +5840,7 @@ static bool cgroup_check_hierarchy_limits(struct cgroup *parent) if (cgroup->nr_descendants >= cgroup->max_descendants) goto fail; - if (level > cgroup->max_depth) + if (level >= cgroup->max_depth) goto fail; level++; @@ -5750,7 +5876,7 @@ int cgroup_mkdir(struct kernfs_node *parent_kn, const char *name, umode_t mode) } /* - * This extra ref will be put in cgroup_free_fn() and guarantees + * This extra ref will be put in css_free_rwork_fn() and guarantees * that @cgrp->kn is always accessible. */ kernfs_get(cgrp->kn); @@ -5828,6 +5954,12 @@ static void kill_css(struct cgroup_subsys_state *css) if (css->flags & CSS_DYING) return; + /* + * Call css_killed(), if defined, before setting the CSS_DYING flag + */ + if (css->ss->css_killed) + css->ss->css_killed(css); + css->flags |= CSS_DYING; /* @@ -5885,7 +6017,7 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) struct cgroup *tcgrp, *parent = cgroup_parent(cgrp); struct cgroup_subsys_state *css; struct cgrp_cset_link *link; - int ssid; + int ssid, ret; lockdep_assert_held(&cgroup_mutex); @@ -5943,7 +6075,9 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) cgroup1_check_for_release(parent); - cgroup_bpf_offline(cgrp); + ret = blocking_notifier_call_chain(&cgroup_lifetime_notifier, + CGROUP_LIFETIME_OFFLINE, cgrp); + WARN_ON_ONCE(notifier_to_errno(ret)); /* put the base reference */ percpu_ref_kill(&cgrp->self.refcnt); @@ -6005,6 +6139,9 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss, bool early) } else { css->id = cgroup_idr_alloc(&ss->css_idr, css, 1, 2, GFP_KERNEL); BUG_ON(css->id < 0); + + BUG_ON(ss_rstat_init(ss)); + BUG_ON(css_rstat_init(css)); } /* Update the init_css_set to contain a subsys @@ -6053,6 +6190,8 @@ int __init cgroup_init_early(void) ss->id, ss->name); WARN(strlen(cgroup_subsys_name[i]) > MAX_CGROUP_TYPE_NAMELEN, "cgroup_subsys_name %s too long\n", cgroup_subsys_name[i]); + WARN(ss->early_init && ss->css_rstat_flush, + "cgroup rstat cannot be used with early init subsystem\n"); ss->id = i; ss->name = cgroup_subsys_name[i]; @@ -6081,7 +6220,7 @@ int __init cgroup_init(void) BUG_ON(cgroup_init_cftypes(NULL, cgroup_psi_files)); BUG_ON(cgroup_init_cftypes(NULL, cgroup1_base_files)); - cgroup_rstat_boot(); + BUG_ON(ss_rstat_init(NULL)); get_user_ns(init_cgroup_ns.user_ns); @@ -6094,6 +6233,8 @@ int __init cgroup_init(void) hash_add(css_set_table, &init_css_set.hlist, css_set_hash(init_css_set.subsys)); + cgroup_bpf_lifetime_notifier_init(); + BUG_ON(cgroup_setup_root(&cgrp_dfl_root, 0)); cgroup_unlock(); @@ -6162,7 +6303,7 @@ int __init cgroup_init(void) WARN_ON(register_filesystem(&cgroup_fs_type)); WARN_ON(register_filesystem(&cgroup2_fs_type)); WARN_ON(!proc_create_single("cgroups", 0, NULL, proc_cgroupstats_show)); -#ifdef CONFIG_CPUSETS +#ifdef CONFIG_CPUSETS_V1 WARN_ON(register_filesystem(&cpuset_fs_type)); #endif @@ -6393,7 +6534,6 @@ static int cgroup_css_set_fork(struct kernel_clone_args *kargs) struct cgroup *dst_cgrp = NULL; struct css_set *cset; struct super_block *sb; - struct file *f; if (kargs->flags & CLONE_INTO_CGROUP) cgroup_lock(); @@ -6403,6 +6543,10 @@ static int cgroup_css_set_fork(struct kernel_clone_args *kargs) spin_lock_irq(&css_set_lock); cset = task_css_set(current); get_css_set(cset); + if (kargs->cgrp) + kargs->kill_seq = kargs->cgrp->kill_seq; + else + kargs->kill_seq = cset->dfl_cgrp->kill_seq; spin_unlock_irq(&css_set_lock); if (!(kargs->flags & CLONE_INTO_CGROUP)) { @@ -6410,14 +6554,14 @@ static int cgroup_css_set_fork(struct kernel_clone_args *kargs) return 0; } - f = fget_raw(kargs->cgroup); - if (!f) { + CLASS(fd_raw, f)(kargs->cgroup); + if (fd_empty(f)) { ret = -EBADF; goto err; } - sb = f->f_path.dentry->d_sb; + sb = fd_file(f)->f_path.dentry->d_sb; - dst_cgrp = cgroup_get_from_file(f); + dst_cgrp = cgroup_get_from_file(fd_file(f)); if (IS_ERR(dst_cgrp)) { ret = PTR_ERR(dst_cgrp); dst_cgrp = NULL; @@ -6465,15 +6609,12 @@ static int cgroup_css_set_fork(struct kernel_clone_args *kargs) } put_css_set(cset); - fput(f); kargs->cgrp = dst_cgrp; return ret; err: cgroup_threadgroup_change_end(current); cgroup_unlock(); - if (f) - fput(f); if (dst_cgrp) cgroup_put(dst_cgrp); put_css_set(cset); @@ -6586,6 +6727,7 @@ void cgroup_post_fork(struct task_struct *child, struct kernel_clone_args *kargs) __releases(&cgroup_threadgroup_rwsem) __releases(&cgroup_mutex) { + unsigned int cgrp_kill_seq = 0; unsigned long cgrp_flags = 0; bool kill = false; struct cgroup_subsys *ss; @@ -6599,10 +6741,13 @@ void cgroup_post_fork(struct task_struct *child, /* init tasks are special, only link regular threads */ if (likely(child->pid)) { - if (kargs->cgrp) + if (kargs->cgrp) { cgrp_flags = kargs->cgrp->flags; - else + cgrp_kill_seq = kargs->cgrp->kill_seq; + } else { cgrp_flags = cset->dfl_cgrp->flags; + cgrp_kill_seq = cset->dfl_cgrp->kill_seq; + } WARN_ON_ONCE(!list_empty(&child->cg_list)); cset->nr_tasks++; @@ -6637,7 +6782,7 @@ void cgroup_post_fork(struct task_struct *child, * child down right after we finished preparing it for * userspace. */ - kill = test_bit(CGRP_KILL, &cgrp_flags); + kill = kargs->kill_seq != cgrp_kill_seq; } spin_unlock_irq(&css_set_lock); @@ -6685,8 +6830,10 @@ void cgroup_exit(struct task_struct *tsk) WARN_ON_ONCE(list_empty(&tsk->cg_list)); cset = task_css_set(tsk); css_set_move_task(tsk, cset, NULL, false); - list_add_tail(&tsk->cg_list, &cset->dying_tasks); cset->nr_tasks--; + /* matches the signal->live check in css_task_iter_advance() */ + if (thread_group_leader(tsk) && atomic_read(&tsk->signal->live)) + list_add_tail(&tsk->cg_list, &cset->dying_tasks); if (dl_task(tsk)) dec_dl_tasks_cs(tsk); @@ -6713,10 +6860,12 @@ void cgroup_release(struct task_struct *task) ss->release(task); } while_each_subsys_mask(); - spin_lock_irq(&css_set_lock); - css_set_skip_task_iters(task_css_set(task), task); - list_del_init(&task->cg_list); - spin_unlock_irq(&css_set_lock); + if (!list_empty(&task->cg_list)) { + spin_lock_irq(&css_set_lock); + css_set_skip_task_iters(task_css_set(task), task); + list_del_init(&task->cg_list); + spin_unlock_irq(&css_set_lock); + } } void cgroup_free(struct task_struct *task) @@ -6879,14 +7028,11 @@ EXPORT_SYMBOL_GPL(cgroup_get_from_path); */ struct cgroup *cgroup_v1v2_get_from_fd(int fd) { - struct cgroup *cgrp; - struct fd f = fdget_raw(fd); - if (!f.file) + CLASS(fd_raw, f)(fd); + if (fd_empty(f)) return ERR_PTR(-EBADF); - cgrp = cgroup_v1v2_get_from_file(f.file); - fdput(f); - return cgrp; + return cgroup_v1v2_get_from_file(fd_file(f)); } /** @@ -7061,7 +7207,8 @@ static ssize_t features_show(struct kobject *kobj, struct kobj_attribute *attr, "favordynmods\n" "memory_localevents\n" "memory_recursiveprot\n" - "memory_hugetlb_accounting\n"); + "memory_hugetlb_accounting\n" + "pids_localevents\n"); } static struct kobj_attribute cgroup_features_attr = __ATTR_RO(features); diff --git a/kernel/cgroup/cpuset-internal.h b/kernel/cgroup/cpuset-internal.h new file mode 100644 index 000000000000..383963e28ac6 --- /dev/null +++ b/kernel/cgroup/cpuset-internal.h @@ -0,0 +1,306 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ + +#ifndef __CPUSET_INTERNAL_H +#define __CPUSET_INTERNAL_H + +#include <linux/cgroup.h> +#include <linux/cpu.h> +#include <linux/cpumask.h> +#include <linux/cpuset.h> +#include <linux/spinlock.h> +#include <linux/union_find.h> + +/* See "Frequency meter" comments, below. */ + +struct fmeter { + int cnt; /* unprocessed events count */ + int val; /* most recent output value */ + time64_t time; /* clock (secs) when val computed */ + spinlock_t lock; /* guards read or write of above */ +}; + +/* + * Invalid partition error code + */ +enum prs_errcode { + PERR_NONE = 0, + PERR_INVCPUS, + PERR_INVPARENT, + PERR_NOTPART, + PERR_NOTEXCL, + PERR_NOCPUS, + PERR_HOTPLUG, + PERR_CPUSEMPTY, + PERR_HKEEPING, + PERR_ACCESS, + PERR_REMOTE, +}; + +/* bits in struct cpuset flags field */ +typedef enum { + CS_ONLINE, + CS_CPU_EXCLUSIVE, + CS_MEM_EXCLUSIVE, + CS_MEM_HARDWALL, + CS_MEMORY_MIGRATE, + CS_SCHED_LOAD_BALANCE, + CS_SPREAD_PAGE, + CS_SPREAD_SLAB, +} cpuset_flagbits_t; + +/* The various types of files and directories in a cpuset file system */ + +typedef enum { + FILE_MEMORY_MIGRATE, + FILE_CPULIST, + FILE_MEMLIST, + FILE_EFFECTIVE_CPULIST, + FILE_EFFECTIVE_MEMLIST, + FILE_SUBPARTS_CPULIST, + FILE_EXCLUSIVE_CPULIST, + FILE_EFFECTIVE_XCPULIST, + FILE_ISOLATED_CPULIST, + FILE_CPU_EXCLUSIVE, + FILE_MEM_EXCLUSIVE, + FILE_MEM_HARDWALL, + FILE_SCHED_LOAD_BALANCE, + FILE_PARTITION_ROOT, + FILE_SCHED_RELAX_DOMAIN_LEVEL, + FILE_MEMORY_PRESSURE_ENABLED, + FILE_MEMORY_PRESSURE, + FILE_SPREAD_PAGE, + FILE_SPREAD_SLAB, +} cpuset_filetype_t; + +struct cpuset { + struct cgroup_subsys_state css; + + unsigned long flags; /* "unsigned long" so bitops work */ + + /* + * On default hierarchy: + * + * The user-configured masks can only be changed by writing to + * cpuset.cpus and cpuset.mems, and won't be limited by the + * parent masks. + * + * The effective masks is the real masks that apply to the tasks + * in the cpuset. They may be changed if the configured masks are + * changed or hotplug happens. + * + * effective_mask == configured_mask & parent's effective_mask, + * and if it ends up empty, it will inherit the parent's mask. + * + * + * On legacy hierarchy: + * + * The user-configured masks are always the same with effective masks. + */ + + /* user-configured CPUs and Memory Nodes allow to tasks */ + cpumask_var_t cpus_allowed; + nodemask_t mems_allowed; + + /* effective CPUs and Memory Nodes allow to tasks */ + cpumask_var_t effective_cpus; + nodemask_t effective_mems; + + /* + * Exclusive CPUs dedicated to current cgroup (default hierarchy only) + * + * The effective_cpus of a valid partition root comes solely from its + * effective_xcpus and some of the effective_xcpus may be distributed + * to sub-partitions below & hence excluded from its effective_cpus. + * For a valid partition root, its effective_cpus have no relationship + * with cpus_allowed unless its exclusive_cpus isn't set. + * + * This value will only be set if either exclusive_cpus is set or + * when this cpuset becomes a local partition root. + */ + cpumask_var_t effective_xcpus; + + /* + * Exclusive CPUs as requested by the user (default hierarchy only) + * + * Its value is independent of cpus_allowed and designates the set of + * CPUs that can be granted to the current cpuset or its children when + * it becomes a valid partition root. The effective set of exclusive + * CPUs granted (effective_xcpus) depends on whether those exclusive + * CPUs are passed down by its ancestors and not yet taken up by + * another sibling partition root along the way. + * + * If its value isn't set, it defaults to cpus_allowed. + */ + cpumask_var_t exclusive_cpus; + + /* + * This is old Memory Nodes tasks took on. + * + * - top_cpuset.old_mems_allowed is initialized to mems_allowed. + * - A new cpuset's old_mems_allowed is initialized when some + * task is moved into it. + * - old_mems_allowed is used in cpuset_migrate_mm() when we change + * cpuset.mems_allowed and have tasks' nodemask updated, and + * then old_mems_allowed is updated to mems_allowed. + */ + nodemask_t old_mems_allowed; + + struct fmeter fmeter; /* memory_pressure filter */ + + /* + * Tasks are being attached to this cpuset. Used to prevent + * zeroing cpus/mems_allowed between ->can_attach() and ->attach(). + */ + int attach_in_progress; + + /* for custom sched domain */ + int relax_domain_level; + + /* number of valid local child partitions */ + int nr_subparts; + + /* partition root state */ + int partition_root_state; + + /* + * number of SCHED_DEADLINE tasks attached to this cpuset, so that we + * know when to rebuild associated root domain bandwidth information. + */ + int nr_deadline_tasks; + int nr_migrate_dl_tasks; + u64 sum_migrate_dl_bw; + + /* Invalid partition error code, not lock protected */ + enum prs_errcode prs_err; + + /* Handle for cpuset.cpus.partition */ + struct cgroup_file partition_file; + + /* Remote partition silbling list anchored at remote_children */ + struct list_head remote_sibling; + + /* Used to merge intersecting subsets for generate_sched_domains */ + struct uf_node node; +}; + +static inline struct cpuset *css_cs(struct cgroup_subsys_state *css) +{ + return css ? container_of(css, struct cpuset, css) : NULL; +} + +/* Retrieve the cpuset for a task */ +static inline struct cpuset *task_cs(struct task_struct *task) +{ + return css_cs(task_css(task, cpuset_cgrp_id)); +} + +static inline struct cpuset *parent_cs(struct cpuset *cs) +{ + return css_cs(cs->css.parent); +} + +/* convenient tests for these bits */ +static inline bool is_cpuset_online(struct cpuset *cs) +{ + return test_bit(CS_ONLINE, &cs->flags) && !css_is_dying(&cs->css); +} + +static inline int is_cpu_exclusive(const struct cpuset *cs) +{ + return test_bit(CS_CPU_EXCLUSIVE, &cs->flags); +} + +static inline int is_mem_exclusive(const struct cpuset *cs) +{ + return test_bit(CS_MEM_EXCLUSIVE, &cs->flags); +} + +static inline int is_mem_hardwall(const struct cpuset *cs) +{ + return test_bit(CS_MEM_HARDWALL, &cs->flags); +} + +static inline int is_sched_load_balance(const struct cpuset *cs) +{ + return test_bit(CS_SCHED_LOAD_BALANCE, &cs->flags); +} + +static inline int is_memory_migrate(const struct cpuset *cs) +{ + return test_bit(CS_MEMORY_MIGRATE, &cs->flags); +} + +static inline int is_spread_page(const struct cpuset *cs) +{ + return test_bit(CS_SPREAD_PAGE, &cs->flags); +} + +static inline int is_spread_slab(const struct cpuset *cs) +{ + return test_bit(CS_SPREAD_SLAB, &cs->flags); +} + +/** + * cpuset_for_each_child - traverse online children of a cpuset + * @child_cs: loop cursor pointing to the current child + * @pos_css: used for iteration + * @parent_cs: target cpuset to walk children of + * + * Walk @child_cs through the online children of @parent_cs. Must be used + * with RCU read locked. + */ +#define cpuset_for_each_child(child_cs, pos_css, parent_cs) \ + css_for_each_child((pos_css), &(parent_cs)->css) \ + if (is_cpuset_online(((child_cs) = css_cs((pos_css))))) + +/** + * cpuset_for_each_descendant_pre - pre-order walk of a cpuset's descendants + * @des_cs: loop cursor pointing to the current descendant + * @pos_css: used for iteration + * @root_cs: target cpuset to walk ancestor of + * + * Walk @des_cs through the online descendants of @root_cs. Must be used + * with RCU read locked. The caller may modify @pos_css by calling + * css_rightmost_descendant() to skip subtree. @root_cs is included in the + * iteration and the first node to be visited. + */ +#define cpuset_for_each_descendant_pre(des_cs, pos_css, root_cs) \ + css_for_each_descendant_pre((pos_css), &(root_cs)->css) \ + if (is_cpuset_online(((des_cs) = css_cs((pos_css))))) + +void rebuild_sched_domains_locked(void); +void cpuset_callback_lock_irq(void); +void cpuset_callback_unlock_irq(void); +void cpuset_update_tasks_cpumask(struct cpuset *cs, struct cpumask *new_cpus); +void cpuset_update_tasks_nodemask(struct cpuset *cs); +int cpuset_update_flag(cpuset_flagbits_t bit, struct cpuset *cs, int turning_on); +ssize_t cpuset_write_resmask(struct kernfs_open_file *of, + char *buf, size_t nbytes, loff_t off); +int cpuset_common_seq_show(struct seq_file *sf, void *v); + +/* + * cpuset-v1.c + */ +#ifdef CONFIG_CPUSETS_V1 +extern struct cftype cpuset1_files[]; +void fmeter_init(struct fmeter *fmp); +void cpuset1_update_task_spread_flags(struct cpuset *cs, + struct task_struct *tsk); +void cpuset1_update_tasks_flags(struct cpuset *cs); +void cpuset1_hotplug_update_tasks(struct cpuset *cs, + struct cpumask *new_cpus, nodemask_t *new_mems, + bool cpus_updated, bool mems_updated); +int cpuset1_validate_change(struct cpuset *cur, struct cpuset *trial); +#else +static inline void fmeter_init(struct fmeter *fmp) {} +static inline void cpuset1_update_task_spread_flags(struct cpuset *cs, + struct task_struct *tsk) {} +static inline void cpuset1_update_tasks_flags(struct cpuset *cs) {} +static inline void cpuset1_hotplug_update_tasks(struct cpuset *cs, + struct cpumask *new_cpus, nodemask_t *new_mems, + bool cpus_updated, bool mems_updated) {} +static inline int cpuset1_validate_change(struct cpuset *cur, + struct cpuset *trial) { return 0; } +#endif /* CONFIG_CPUSETS_V1 */ + +#endif /* __CPUSET_INTERNAL_H */ diff --git a/kernel/cgroup/cpuset-v1.c b/kernel/cgroup/cpuset-v1.c new file mode 100644 index 000000000000..b69a7db67090 --- /dev/null +++ b/kernel/cgroup/cpuset-v1.c @@ -0,0 +1,611 @@ +// SPDX-License-Identifier: GPL-2.0-or-later + +#include "cgroup-internal.h" +#include "cpuset-internal.h" + +/* + * Legacy hierarchy call to cgroup_transfer_tasks() is handled asynchrously + */ +struct cpuset_remove_tasks_struct { + struct work_struct work; + struct cpuset *cs; +}; + +/* + * Frequency meter - How fast is some event occurring? + * + * These routines manage a digitally filtered, constant time based, + * event frequency meter. There are four routines: + * fmeter_init() - initialize a frequency meter. + * fmeter_markevent() - called each time the event happens. + * fmeter_getrate() - returns the recent rate of such events. + * fmeter_update() - internal routine used to update fmeter. + * + * A common data structure is passed to each of these routines, + * which is used to keep track of the state required to manage the + * frequency meter and its digital filter. + * + * The filter works on the number of events marked per unit time. + * The filter is single-pole low-pass recursive (IIR). The time unit + * is 1 second. Arithmetic is done using 32-bit integers scaled to + * simulate 3 decimal digits of precision (multiplied by 1000). + * + * With an FM_COEF of 933, and a time base of 1 second, the filter + * has a half-life of 10 seconds, meaning that if the events quit + * happening, then the rate returned from the fmeter_getrate() + * will be cut in half each 10 seconds, until it converges to zero. + * + * It is not worth doing a real infinitely recursive filter. If more + * than FM_MAXTICKS ticks have elapsed since the last filter event, + * just compute FM_MAXTICKS ticks worth, by which point the level + * will be stable. + * + * Limit the count of unprocessed events to FM_MAXCNT, so as to avoid + * arithmetic overflow in the fmeter_update() routine. + * + * Given the simple 32 bit integer arithmetic used, this meter works + * best for reporting rates between one per millisecond (msec) and + * one per 32 (approx) seconds. At constant rates faster than one + * per msec it maxes out at values just under 1,000,000. At constant + * rates between one per msec, and one per second it will stabilize + * to a value N*1000, where N is the rate of events per second. + * At constant rates between one per second and one per 32 seconds, + * it will be choppy, moving up on the seconds that have an event, + * and then decaying until the next event. At rates slower than + * about one in 32 seconds, it decays all the way back to zero between + * each event. + */ + +#define FM_COEF 933 /* coefficient for half-life of 10 secs */ +#define FM_MAXTICKS ((u32)99) /* useless computing more ticks than this */ +#define FM_MAXCNT 1000000 /* limit cnt to avoid overflow */ +#define FM_SCALE 1000 /* faux fixed point scale */ + +/* Initialize a frequency meter */ +void fmeter_init(struct fmeter *fmp) +{ + fmp->cnt = 0; + fmp->val = 0; + fmp->time = 0; + spin_lock_init(&fmp->lock); +} + +/* Internal meter update - process cnt events and update value */ +static void fmeter_update(struct fmeter *fmp) +{ + time64_t now; + u32 ticks; + + now = ktime_get_seconds(); + ticks = now - fmp->time; + + if (ticks == 0) + return; + + ticks = min(FM_MAXTICKS, ticks); + while (ticks-- > 0) + fmp->val = (FM_COEF * fmp->val) / FM_SCALE; + fmp->time = now; + + fmp->val += ((FM_SCALE - FM_COEF) * fmp->cnt) / FM_SCALE; + fmp->cnt = 0; +} + +/* Process any previous ticks, then bump cnt by one (times scale). */ +static void fmeter_markevent(struct fmeter *fmp) +{ + spin_lock(&fmp->lock); + fmeter_update(fmp); + fmp->cnt = min(FM_MAXCNT, fmp->cnt + FM_SCALE); + spin_unlock(&fmp->lock); +} + +/* Process any previous ticks, then return current value. */ +static int fmeter_getrate(struct fmeter *fmp) +{ + int val; + + spin_lock(&fmp->lock); + fmeter_update(fmp); + val = fmp->val; + spin_unlock(&fmp->lock); + return val; +} + +/* + * Collection of memory_pressure is suppressed unless + * this flag is enabled by writing "1" to the special + * cpuset file 'memory_pressure_enabled' in the root cpuset. + */ + +int cpuset_memory_pressure_enabled __read_mostly; + +/* + * __cpuset_memory_pressure_bump - keep stats of per-cpuset reclaims. + * + * Keep a running average of the rate of synchronous (direct) + * page reclaim efforts initiated by tasks in each cpuset. + * + * This represents the rate at which some task in the cpuset + * ran low on memory on all nodes it was allowed to use, and + * had to enter the kernels page reclaim code in an effort to + * create more free memory by tossing clean pages or swapping + * or writing dirty pages. + * + * Display to user space in the per-cpuset read-only file + * "memory_pressure". Value displayed is an integer + * representing the recent rate of entry into the synchronous + * (direct) page reclaim by any task attached to the cpuset. + */ + +void __cpuset_memory_pressure_bump(void) +{ + rcu_read_lock(); + fmeter_markevent(&task_cs(current)->fmeter); + rcu_read_unlock(); +} + +static int update_relax_domain_level(struct cpuset *cs, s64 val) +{ +#ifdef CONFIG_SMP + if (val < -1 || val > sched_domain_level_max + 1) + return -EINVAL; +#endif + + if (val != cs->relax_domain_level) { + cs->relax_domain_level = val; + if (!cpumask_empty(cs->cpus_allowed) && + is_sched_load_balance(cs)) + rebuild_sched_domains_locked(); + } + + return 0; +} + +static int cpuset_write_s64(struct cgroup_subsys_state *css, struct cftype *cft, + s64 val) +{ + struct cpuset *cs = css_cs(css); + cpuset_filetype_t type = cft->private; + int retval = -ENODEV; + + cpus_read_lock(); + cpuset_lock(); + if (!is_cpuset_online(cs)) + goto out_unlock; + + switch (type) { + case FILE_SCHED_RELAX_DOMAIN_LEVEL: + pr_info_once("cpuset.%s is deprecated\n", cft->name); + retval = update_relax_domain_level(cs, val); + break; + default: + retval = -EINVAL; + break; + } +out_unlock: + cpuset_unlock(); + cpus_read_unlock(); + return retval; +} + +static s64 cpuset_read_s64(struct cgroup_subsys_state *css, struct cftype *cft) +{ + struct cpuset *cs = css_cs(css); + cpuset_filetype_t type = cft->private; + + switch (type) { + case FILE_SCHED_RELAX_DOMAIN_LEVEL: + return cs->relax_domain_level; + default: + BUG(); + } + + /* Unreachable but makes gcc happy */ + return 0; +} + +/* + * update task's spread flag if cpuset's page/slab spread flag is set + * + * Call with callback_lock or cpuset_mutex held. The check can be skipped + * if on default hierarchy. + */ +void cpuset1_update_task_spread_flags(struct cpuset *cs, + struct task_struct *tsk) +{ + if (cgroup_subsys_on_dfl(cpuset_cgrp_subsys)) + return; + + if (is_spread_page(cs)) + task_set_spread_page(tsk); + else + task_clear_spread_page(tsk); + + if (is_spread_slab(cs)) + task_set_spread_slab(tsk); + else + task_clear_spread_slab(tsk); +} + +/** + * cpuset1_update_tasks_flags - update the spread flags of tasks in the cpuset. + * @cs: the cpuset in which each task's spread flags needs to be changed + * + * Iterate through each task of @cs updating its spread flags. As this + * function is called with cpuset_mutex held, cpuset membership stays + * stable. + */ +void cpuset1_update_tasks_flags(struct cpuset *cs) +{ + struct css_task_iter it; + struct task_struct *task; + + css_task_iter_start(&cs->css, 0, &it); + while ((task = css_task_iter_next(&it))) + cpuset1_update_task_spread_flags(cs, task); + css_task_iter_end(&it); +} + +/* + * If CPU and/or memory hotplug handlers, below, unplug any CPUs + * or memory nodes, we need to walk over the cpuset hierarchy, + * removing that CPU or node from all cpusets. If this removes the + * last CPU or node from a cpuset, then move the tasks in the empty + * cpuset to its next-highest non-empty parent. + */ +static void remove_tasks_in_empty_cpuset(struct cpuset *cs) +{ + struct cpuset *parent; + + /* + * Find its next-highest non-empty parent, (top cpuset + * has online cpus, so can't be empty). + */ + parent = parent_cs(cs); + while (cpumask_empty(parent->cpus_allowed) || + nodes_empty(parent->mems_allowed)) + parent = parent_cs(parent); + + if (cgroup_transfer_tasks(parent->css.cgroup, cs->css.cgroup)) { + pr_err("cpuset: failed to transfer tasks out of empty cpuset "); + pr_cont_cgroup_name(cs->css.cgroup); + pr_cont("\n"); + } +} + +static void cpuset_migrate_tasks_workfn(struct work_struct *work) +{ + struct cpuset_remove_tasks_struct *s; + + s = container_of(work, struct cpuset_remove_tasks_struct, work); + remove_tasks_in_empty_cpuset(s->cs); + css_put(&s->cs->css); + kfree(s); +} + +void cpuset1_hotplug_update_tasks(struct cpuset *cs, + struct cpumask *new_cpus, nodemask_t *new_mems, + bool cpus_updated, bool mems_updated) +{ + bool is_empty; + + cpuset_callback_lock_irq(); + cpumask_copy(cs->cpus_allowed, new_cpus); + cpumask_copy(cs->effective_cpus, new_cpus); + cs->mems_allowed = *new_mems; + cs->effective_mems = *new_mems; + cpuset_callback_unlock_irq(); + + /* + * Don't call cpuset_update_tasks_cpumask() if the cpuset becomes empty, + * as the tasks will be migrated to an ancestor. + */ + if (cpus_updated && !cpumask_empty(cs->cpus_allowed)) + cpuset_update_tasks_cpumask(cs, new_cpus); + if (mems_updated && !nodes_empty(cs->mems_allowed)) + cpuset_update_tasks_nodemask(cs); + + is_empty = cpumask_empty(cs->cpus_allowed) || + nodes_empty(cs->mems_allowed); + + /* + * Move tasks to the nearest ancestor with execution resources, + * This is full cgroup operation which will also call back into + * cpuset. Execute it asynchronously using workqueue. + */ + if (is_empty && cs->css.cgroup->nr_populated_csets && + css_tryget_online(&cs->css)) { + struct cpuset_remove_tasks_struct *s; + + s = kzalloc(sizeof(*s), GFP_KERNEL); + if (WARN_ON_ONCE(!s)) { + css_put(&cs->css); + return; + } + + s->cs = cs; + INIT_WORK(&s->work, cpuset_migrate_tasks_workfn); + schedule_work(&s->work); + } +} + +/* + * is_cpuset_subset(p, q) - Is cpuset p a subset of cpuset q? + * + * One cpuset is a subset of another if all its allowed CPUs and + * Memory Nodes are a subset of the other, and its exclusive flags + * are only set if the other's are set. Call holding cpuset_mutex. + */ + +static int is_cpuset_subset(const struct cpuset *p, const struct cpuset *q) +{ + return cpumask_subset(p->cpus_allowed, q->cpus_allowed) && + nodes_subset(p->mems_allowed, q->mems_allowed) && + is_cpu_exclusive(p) <= is_cpu_exclusive(q) && + is_mem_exclusive(p) <= is_mem_exclusive(q); +} + +/* + * cpuset1_validate_change() - Validate conditions specific to legacy (v1) + * behavior. + */ +int cpuset1_validate_change(struct cpuset *cur, struct cpuset *trial) +{ + struct cgroup_subsys_state *css; + struct cpuset *c, *par; + int ret; + + WARN_ON_ONCE(!rcu_read_lock_held()); + + /* Each of our child cpusets must be a subset of us */ + ret = -EBUSY; + cpuset_for_each_child(c, css, cur) + if (!is_cpuset_subset(c, trial)) + goto out; + + /* On legacy hierarchy, we must be a subset of our parent cpuset. */ + ret = -EACCES; + par = parent_cs(cur); + if (par && !is_cpuset_subset(trial, par)) + goto out; + + ret = 0; +out: + return ret; +} + +#ifdef CONFIG_PROC_PID_CPUSET +/* + * proc_cpuset_show() + * - Print tasks cpuset path into seq_file. + * - Used for /proc/<pid>/cpuset. + */ +int proc_cpuset_show(struct seq_file *m, struct pid_namespace *ns, + struct pid *pid, struct task_struct *tsk) +{ + char *buf; + struct cgroup_subsys_state *css; + int retval; + + retval = -ENOMEM; + buf = kmalloc(PATH_MAX, GFP_KERNEL); + if (!buf) + goto out; + + rcu_read_lock(); + spin_lock_irq(&css_set_lock); + css = task_css(tsk, cpuset_cgrp_id); + retval = cgroup_path_ns_locked(css->cgroup, buf, PATH_MAX, + current->nsproxy->cgroup_ns); + spin_unlock_irq(&css_set_lock); + rcu_read_unlock(); + + if (retval == -E2BIG) + retval = -ENAMETOOLONG; + if (retval < 0) + goto out_free; + seq_puts(m, buf); + seq_putc(m, '\n'); + retval = 0; +out_free: + kfree(buf); +out: + return retval; +} +#endif /* CONFIG_PROC_PID_CPUSET */ + +static u64 cpuset_read_u64(struct cgroup_subsys_state *css, struct cftype *cft) +{ + struct cpuset *cs = css_cs(css); + cpuset_filetype_t type = cft->private; + + switch (type) { + case FILE_CPU_EXCLUSIVE: + return is_cpu_exclusive(cs); + case FILE_MEM_EXCLUSIVE: + return is_mem_exclusive(cs); + case FILE_MEM_HARDWALL: + return is_mem_hardwall(cs); + case FILE_SCHED_LOAD_BALANCE: + return is_sched_load_balance(cs); + case FILE_MEMORY_MIGRATE: + return is_memory_migrate(cs); + case FILE_MEMORY_PRESSURE_ENABLED: + return cpuset_memory_pressure_enabled; + case FILE_MEMORY_PRESSURE: + return fmeter_getrate(&cs->fmeter); + case FILE_SPREAD_PAGE: + return is_spread_page(cs); + case FILE_SPREAD_SLAB: + return is_spread_slab(cs); + default: + BUG(); + } + + /* Unreachable but makes gcc happy */ + return 0; +} + +static int cpuset_write_u64(struct cgroup_subsys_state *css, struct cftype *cft, + u64 val) +{ + struct cpuset *cs = css_cs(css); + cpuset_filetype_t type = cft->private; + int retval = 0; + + cpus_read_lock(); + cpuset_lock(); + if (!is_cpuset_online(cs)) { + retval = -ENODEV; + goto out_unlock; + } + + switch (type) { + case FILE_CPU_EXCLUSIVE: + retval = cpuset_update_flag(CS_CPU_EXCLUSIVE, cs, val); + break; + case FILE_MEM_EXCLUSIVE: + pr_info_once("cpuset.%s is deprecated\n", cft->name); + retval = cpuset_update_flag(CS_MEM_EXCLUSIVE, cs, val); + break; + case FILE_MEM_HARDWALL: + pr_info_once("cpuset.%s is deprecated\n", cft->name); + retval = cpuset_update_flag(CS_MEM_HARDWALL, cs, val); + break; + case FILE_SCHED_LOAD_BALANCE: + pr_info_once("cpuset.%s is deprecated, use cpuset.cpus.partition instead\n", cft->name); + retval = cpuset_update_flag(CS_SCHED_LOAD_BALANCE, cs, val); + break; + case FILE_MEMORY_MIGRATE: + pr_info_once("cpuset.%s is deprecated\n", cft->name); + retval = cpuset_update_flag(CS_MEMORY_MIGRATE, cs, val); + break; + case FILE_MEMORY_PRESSURE_ENABLED: + pr_info_once("cpuset.%s is deprecated, use memory.pressure with CONFIG_PSI instead\n", cft->name); + cpuset_memory_pressure_enabled = !!val; + break; + case FILE_SPREAD_PAGE: + pr_info_once("cpuset.%s is deprecated\n", cft->name); + retval = cpuset_update_flag(CS_SPREAD_PAGE, cs, val); + break; + case FILE_SPREAD_SLAB: + pr_warn_once("cpuset.%s is deprecated\n", cft->name); + retval = cpuset_update_flag(CS_SPREAD_SLAB, cs, val); + break; + default: + retval = -EINVAL; + break; + } +out_unlock: + cpuset_unlock(); + cpus_read_unlock(); + return retval; +} + +/* + * for the common functions, 'private' gives the type of file + */ + +struct cftype cpuset1_files[] = { + { + .name = "cpus", + .seq_show = cpuset_common_seq_show, + .write = cpuset_write_resmask, + .max_write_len = (100U + 6 * NR_CPUS), + .private = FILE_CPULIST, + }, + + { + .name = "mems", + .seq_show = cpuset_common_seq_show, + .write = cpuset_write_resmask, + .max_write_len = (100U + 6 * MAX_NUMNODES), + .private = FILE_MEMLIST, + }, + + { + .name = "effective_cpus", + .seq_show = cpuset_common_seq_show, + .private = FILE_EFFECTIVE_CPULIST, + }, + + { + .name = "effective_mems", + .seq_show = cpuset_common_seq_show, + .private = FILE_EFFECTIVE_MEMLIST, + }, + + { + .name = "cpu_exclusive", + .read_u64 = cpuset_read_u64, + .write_u64 = cpuset_write_u64, + .private = FILE_CPU_EXCLUSIVE, + }, + + { + .name = "mem_exclusive", + .read_u64 = cpuset_read_u64, + .write_u64 = cpuset_write_u64, + .private = FILE_MEM_EXCLUSIVE, + }, + + { + .name = "mem_hardwall", + .read_u64 = cpuset_read_u64, + .write_u64 = cpuset_write_u64, + .private = FILE_MEM_HARDWALL, + }, + + { + .name = "sched_load_balance", + .read_u64 = cpuset_read_u64, + .write_u64 = cpuset_write_u64, + .private = FILE_SCHED_LOAD_BALANCE, + }, + + { + .name = "sched_relax_domain_level", + .read_s64 = cpuset_read_s64, + .write_s64 = cpuset_write_s64, + .private = FILE_SCHED_RELAX_DOMAIN_LEVEL, + }, + + { + .name = "memory_migrate", + .read_u64 = cpuset_read_u64, + .write_u64 = cpuset_write_u64, + .private = FILE_MEMORY_MIGRATE, + }, + + { + .name = "memory_pressure", + .read_u64 = cpuset_read_u64, + .private = FILE_MEMORY_PRESSURE, + }, + + { + .name = "memory_spread_page", + .read_u64 = cpuset_read_u64, + .write_u64 = cpuset_write_u64, + .private = FILE_SPREAD_PAGE, + }, + + { + /* obsolete, may be removed in the future */ + .name = "memory_spread_slab", + .read_u64 = cpuset_read_u64, + .write_u64 = cpuset_write_u64, + .private = FILE_SPREAD_SLAB, + }, + + { + .name = "memory_pressure_enabled", + .flags = CFTYPE_ONLY_ON_ROOT, + .read_u64 = cpuset_read_u64, + .write_u64 = cpuset_write_u64, + .private = FILE_MEMORY_PRESSURE_ENABLED, + }, + + { } /* terminate */ +}; diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c index 4237c8748715..3bc4301466f3 100644 --- a/kernel/cgroup/cpuset.c +++ b/kernel/cgroup/cpuset.c @@ -21,11 +21,8 @@ * License. See the file COPYING in the main directory of the Linux * distribution for more details. */ +#include "cpuset-internal.h" -#include <linux/cpu.h> -#include <linux/cpumask.h> -#include <linux/cpuset.h> -#include <linux/delay.h> #include <linux/init.h> #include <linux/interrupt.h> #include <linux/kernel.h> @@ -39,10 +36,8 @@ #include <linux/sched/mm.h> #include <linux/sched/task.h> #include <linux/security.h> -#include <linux/spinlock.h> #include <linux/oom.h> #include <linux/sched/isolation.h> -#include <linux/cgroup.h> #include <linux/wait.h> #include <linux/workqueue.h> @@ -56,30 +51,6 @@ DEFINE_STATIC_KEY_FALSE(cpusets_enabled_key); */ DEFINE_STATIC_KEY_FALSE(cpusets_insane_config_key); -/* See "Frequency meter" comments, below. */ - -struct fmeter { - int cnt; /* unprocessed events count */ - int val; /* most recent output value */ - time64_t time; /* clock (secs) when val computed */ - spinlock_t lock; /* guards read or write of above */ -}; - -/* - * Invalid partition error code - */ -enum prs_errcode { - PERR_NONE = 0, - PERR_INVCPUS, - PERR_INVPARENT, - PERR_NOTPART, - PERR_NOTEXCL, - PERR_NOCPUS, - PERR_HOTPLUG, - PERR_CPUSEMPTY, - PERR_HKEEPING, -}; - static const char * const perr_strings[] = { [PERR_INVCPUS] = "Invalid cpu list in cpuset.cpus.exclusive", [PERR_INVPARENT] = "Parent is an invalid partition root", @@ -87,122 +58,20 @@ static const char * const perr_strings[] = { [PERR_NOTEXCL] = "Cpu list in cpuset.cpus not exclusive", [PERR_NOCPUS] = "Parent unable to distribute cpu downstream", [PERR_HOTPLUG] = "No cpu available due to hotplug", - [PERR_CPUSEMPTY] = "cpuset.cpus is empty", + [PERR_CPUSEMPTY] = "cpuset.cpus and cpuset.cpus.exclusive are empty", [PERR_HKEEPING] = "partition config conflicts with housekeeping setup", + [PERR_ACCESS] = "Enable partition not permitted", + [PERR_REMOTE] = "Have remote partition underneath", }; -struct cpuset { - struct cgroup_subsys_state css; - - unsigned long flags; /* "unsigned long" so bitops work */ - - /* - * On default hierarchy: - * - * The user-configured masks can only be changed by writing to - * cpuset.cpus and cpuset.mems, and won't be limited by the - * parent masks. - * - * The effective masks is the real masks that apply to the tasks - * in the cpuset. They may be changed if the configured masks are - * changed or hotplug happens. - * - * effective_mask == configured_mask & parent's effective_mask, - * and if it ends up empty, it will inherit the parent's mask. - * - * - * On legacy hierarchy: - * - * The user-configured masks are always the same with effective masks. - */ - - /* user-configured CPUs and Memory Nodes allow to tasks */ - cpumask_var_t cpus_allowed; - nodemask_t mems_allowed; - - /* effective CPUs and Memory Nodes allow to tasks */ - cpumask_var_t effective_cpus; - nodemask_t effective_mems; - - /* - * Exclusive CPUs dedicated to current cgroup (default hierarchy only) - * - * This exclusive CPUs must be a subset of cpus_allowed. A parent - * cgroup can only grant exclusive CPUs to one of its children. - * - * When the cgroup becomes a valid partition root, effective_xcpus - * defaults to cpus_allowed if not set. The effective_cpus of a valid - * partition root comes solely from its effective_xcpus and some of the - * effective_xcpus may be distributed to sub-partitions below & hence - * excluded from its effective_cpus. - */ - cpumask_var_t effective_xcpus; - - /* - * Exclusive CPUs as requested by the user (default hierarchy only) - */ - cpumask_var_t exclusive_cpus; - - /* - * This is old Memory Nodes tasks took on. - * - * - top_cpuset.old_mems_allowed is initialized to mems_allowed. - * - A new cpuset's old_mems_allowed is initialized when some - * task is moved into it. - * - old_mems_allowed is used in cpuset_migrate_mm() when we change - * cpuset.mems_allowed and have tasks' nodemask updated, and - * then old_mems_allowed is updated to mems_allowed. - */ - nodemask_t old_mems_allowed; - - struct fmeter fmeter; /* memory_pressure filter */ - - /* - * Tasks are being attached to this cpuset. Used to prevent - * zeroing cpus/mems_allowed between ->can_attach() and ->attach(). - */ - int attach_in_progress; - - /* partition number for rebuild_sched_domains() */ - int pn; - - /* for custom sched domain */ - int relax_domain_level; - - /* number of valid sub-partitions */ - int nr_subparts; - - /* partition root state */ - int partition_root_state; - - /* - * Default hierarchy only: - * use_parent_ecpus - set if using parent's effective_cpus - * child_ecpus_count - # of children with use_parent_ecpus set - */ - int use_parent_ecpus; - int child_ecpus_count; - - /* - * number of SCHED_DEADLINE tasks attached to this cpuset, so that we - * know when to rebuild associated root domain bandwidth information. - */ - int nr_deadline_tasks; - int nr_migrate_dl_tasks; - u64 sum_migrate_dl_bw; - - /* Invalid partition error code, not lock protected */ - enum prs_errcode prs_err; - - /* Handle for cpuset.cpus.partition */ - struct cgroup_file partition_file; - - /* Remote partition silbling list anchored at remote_children */ - struct list_head remote_sibling; -}; - /* - * Exclusive CPUs distributed out to sub-partitions of top_cpuset + * For local partitions, update to subpartitions_cpus & isolated_cpus is done + * in update_parent_effective_cpumask(). For remote partitions, it is done in + * the remote_partition_*() and remote_cpus_update() helpers. + */ +/* + * Exclusive CPUs distributed out to local or remote sub-partitions of + * top_cpuset */ static cpumask_var_t subpartitions_cpus; @@ -211,10 +80,32 @@ static cpumask_var_t subpartitions_cpus; */ static cpumask_var_t isolated_cpus; +/* + * Housekeeping (HK_TYPE_DOMAIN) CPUs at boot + */ +static cpumask_var_t boot_hk_cpus; +static bool have_boot_isolcpus; + /* List of remote partition root children */ static struct list_head remote_children; /* + * A flag to force sched domain rebuild at the end of an operation. + * It can be set in + * - update_partition_sd_lb() + * - update_cpumasks_hier() + * - cpuset_update_flag() + * - cpuset_hotplug_update_tasks() + * - cpuset_handle_hotplug() + * + * Protected by cpuset_mutex (with cpus_read_lock held) or cpus_write_lock. + * + * Note that update_relax_domain_level() in cpuset-v1.c can still call + * rebuild_sched_domains_locked() directly without using this flag. + */ +static bool force_sd_rebuild; + +/* * Partition root states: * * 0 - member (not a partition root) @@ -222,6 +113,17 @@ static struct list_head remote_children; * 2 - partition root without load balancing (isolated) * -1 - invalid partition root * -2 - invalid isolated partition root + * + * There are 2 types of partitions - local or remote. Local partitions are + * those whose parents are partition root themselves. Setting of + * cpuset.cpus.exclusive are optional in setting up local partitions. + * Remote partitions are those whose parents are not partition roots. Passing + * down exclusive CPUs by setting cpuset.cpus.exclusive along its ancestor + * nodes are mandatory in creating a remote partition. + * + * For simplicity, a local partition can be created under a local or remote + * partition but a remote partition cannot have any partition root in its + * ancestor chain except the cgroup root. */ #define PRS_MEMBER 0 #define PRS_ROOT 1 @@ -243,22 +145,6 @@ struct tmpmasks { cpumask_var_t new_cpus; /* For update_cpumasks_hier() */ }; -static inline struct cpuset *css_cs(struct cgroup_subsys_state *css) -{ - return css ? container_of(css, struct cpuset, css) : NULL; -} - -/* Retrieve the cpuset for a task */ -static inline struct cpuset *task_cs(struct task_struct *task) -{ - return css_cs(task_css(task, cpuset_cgrp_id)); -} - -static inline struct cpuset *parent_cs(struct cpuset *cs) -{ - return css_cs(cs->css.parent); -} - void inc_dl_tasks_cs(struct task_struct *p) { struct cpuset *cs = task_cs(p); @@ -273,59 +159,6 @@ void dec_dl_tasks_cs(struct task_struct *p) cs->nr_deadline_tasks--; } -/* bits in struct cpuset flags field */ -typedef enum { - CS_ONLINE, - CS_CPU_EXCLUSIVE, - CS_MEM_EXCLUSIVE, - CS_MEM_HARDWALL, - CS_MEMORY_MIGRATE, - CS_SCHED_LOAD_BALANCE, - CS_SPREAD_PAGE, - CS_SPREAD_SLAB, -} cpuset_flagbits_t; - -/* convenient tests for these bits */ -static inline bool is_cpuset_online(struct cpuset *cs) -{ - return test_bit(CS_ONLINE, &cs->flags) && !css_is_dying(&cs->css); -} - -static inline int is_cpu_exclusive(const struct cpuset *cs) -{ - return test_bit(CS_CPU_EXCLUSIVE, &cs->flags); -} - -static inline int is_mem_exclusive(const struct cpuset *cs) -{ - return test_bit(CS_MEM_EXCLUSIVE, &cs->flags); -} - -static inline int is_mem_hardwall(const struct cpuset *cs) -{ - return test_bit(CS_MEM_HARDWALL, &cs->flags); -} - -static inline int is_sched_load_balance(const struct cpuset *cs) -{ - return test_bit(CS_SCHED_LOAD_BALANCE, &cs->flags); -} - -static inline int is_memory_migrate(const struct cpuset *cs) -{ - return test_bit(CS_MEMORY_MIGRATE, &cs->flags); -} - -static inline int is_spread_page(const struct cpuset *cs) -{ - return test_bit(CS_SPREAD_PAGE, &cs->flags); -} - -static inline int is_spread_slab(const struct cpuset *cs) -{ - return test_bit(CS_SPREAD_SLAB, &cs->flags); -} - static inline int is_partition_valid(const struct cpuset *cs) { return cs->partition_root_state > 0; @@ -359,47 +192,32 @@ static inline void notify_partition_change(struct cpuset *cs, int old_prs) WRITE_ONCE(cs->prs_err, PERR_NONE); } +/* + * The top_cpuset is always synchronized to cpu_active_mask and we should avoid + * using cpu_online_mask as much as possible. An active CPU is always an online + * CPU, but not vice versa. cpu_active_mask and cpu_online_mask can differ + * during hotplug operations. A CPU is marked active at the last stage of CPU + * bringup (CPUHP_AP_ACTIVE). It is also the stage where cpuset hotplug code + * will be called to update the sched domains so that the scheduler can move + * a normal task to a newly active CPU or remove tasks away from a newly + * inactivated CPU. The online bit is set much earlier in the CPU bringup + * process and cleared much later in CPU teardown. + * + * If cpu_online_mask is used while a hotunplug operation is happening in + * parallel, we may leave an offline CPU in cpu_allowed or some other masks. + */ static struct cpuset top_cpuset = { - .flags = ((1 << CS_ONLINE) | (1 << CS_CPU_EXCLUSIVE) | - (1 << CS_MEM_EXCLUSIVE)), + .flags = BIT(CS_ONLINE) | BIT(CS_CPU_EXCLUSIVE) | + BIT(CS_MEM_EXCLUSIVE) | BIT(CS_SCHED_LOAD_BALANCE), .partition_root_state = PRS_ROOT, + .relax_domain_level = -1, .remote_sibling = LIST_HEAD_INIT(top_cpuset.remote_sibling), }; -/** - * cpuset_for_each_child - traverse online children of a cpuset - * @child_cs: loop cursor pointing to the current child - * @pos_css: used for iteration - * @parent_cs: target cpuset to walk children of - * - * Walk @child_cs through the online children of @parent_cs. Must be used - * with RCU read locked. - */ -#define cpuset_for_each_child(child_cs, pos_css, parent_cs) \ - css_for_each_child((pos_css), &(parent_cs)->css) \ - if (is_cpuset_online(((child_cs) = css_cs((pos_css))))) - -/** - * cpuset_for_each_descendant_pre - pre-order walk of a cpuset's descendants - * @des_cs: loop cursor pointing to the current descendant - * @pos_css: used for iteration - * @root_cs: target cpuset to walk ancestor of - * - * Walk @des_cs through the online descendants of @root_cs. Must be used - * with RCU read locked. The caller may modify @pos_css by calling - * css_rightmost_descendant() to skip subtree. @root_cs is included in the - * iteration and the first node to be visited. - */ -#define cpuset_for_each_descendant_pre(des_cs, pos_css, root_cs) \ - css_for_each_descendant_pre((pos_css), &(root_cs)->css) \ - if (is_cpuset_online(((des_cs) = css_cs((pos_css))))) - /* * There are two global locks guarding cpuset structures - cpuset_mutex and - * callback_lock. We also require taking task_lock() when dereferencing a - * task's cpuset pointer. See "The task_lock() exception", at the end of this - * comment. The cpuset code uses only cpuset_mutex. Other kernel subsystems - * can use cpuset_lock()/cpuset_unlock() to prevent change to cpuset + * callback_lock. The cpuset code uses only cpuset_mutex. Other kernel + * subsystems can use cpuset_lock()/cpuset_unlock() to prevent change to cpuset * structures. Note that cpuset_mutex needs to be a mutex as it is used in * paths that rely on priority inheritance (e.g. scheduler - on RT) for * correctness. @@ -425,12 +243,9 @@ static struct cpuset top_cpuset = { * by other task, we use alloc_lock in the task_struct fields to protect * them. * - * The cpuset_common_file_read() handlers only hold callback_lock across + * The cpuset_common_seq_show() handlers only hold callback_lock across * small pieces of code, such as when reading out possibly multi-word * cpumasks and nodemasks. - * - * Accessing a task's cpuset should be done in accordance with the - * guidelines for accessing subsystem state in kernel/cgroup.c */ static DEFINE_MUTEX(cpuset_mutex); @@ -447,13 +262,17 @@ void cpuset_unlock(void) static DEFINE_SPINLOCK(callback_lock); -static struct workqueue_struct *cpuset_migrate_mm_wq; +void cpuset_callback_lock_irq(void) +{ + spin_lock_irq(&callback_lock); +} -/* - * CPU / memory hotplug is handled asynchronously. - */ -static void cpuset_hotplug_workfn(struct work_struct *work); -static DECLARE_WORK(cpuset_hotplug_work, cpuset_hotplug_workfn); +void cpuset_callback_unlock_irq(void) +{ + spin_unlock_irq(&callback_lock); +} + +static struct workqueue_struct *cpuset_migrate_mm_wq; static DECLARE_WAIT_QUEUE_HEAD(cpuset_attach_wq); @@ -469,6 +288,32 @@ static inline void check_insane_mems_config(nodemask_t *nodes) } /* + * decrease cs->attach_in_progress. + * wake_up cpuset_attach_wq if cs->attach_in_progress==0. + */ +static inline void dec_attach_in_progress_locked(struct cpuset *cs) +{ + lockdep_assert_held(&cpuset_mutex); + + cs->attach_in_progress--; + if (!cs->attach_in_progress) + wake_up(&cpuset_attach_wq); +} + +static inline void dec_attach_in_progress(struct cpuset *cs) +{ + mutex_lock(&cpuset_mutex); + dec_attach_in_progress_locked(cs); + mutex_unlock(&cpuset_mutex); +} + +static inline bool cpuset_v2(void) +{ + return !IS_ENABLED(CONFIG_CPUSETS_V1) || + cgroup_subsys_on_dfl(cpuset_cgrp_subsys); +} + +/* * Cgroup v2 behavior is used on the "cpus" and "mems" control files when * on default hierarchy or when the cpuset_v2_mode flag is set by mounting * the v1 cpuset cgroup filesystem with the "cpuset_v2_mode" mount option. @@ -478,7 +323,7 @@ static inline void check_insane_mems_config(nodemask_t *nodes) */ static inline bool is_in_v2_mode(void) { - return cgroup_subsys_on_dfl(cpuset_cgrp_subsys) || + return cpuset_v2() || (cpuset_cgrp_subsys.root->flags & CGRP_ROOT_CPUSET_V2_MODE); } @@ -524,38 +369,26 @@ static inline bool partition_is_populated(struct cpuset *cs, * appropriate cpus. * * One way or another, we guarantee to return some non-empty subset - * of cpu_online_mask. + * of cpu_active_mask. * * Call with callback_lock or cpuset_mutex held. */ -static void guarantee_online_cpus(struct task_struct *tsk, +static void guarantee_active_cpus(struct task_struct *tsk, struct cpumask *pmask) { const struct cpumask *possible_mask = task_cpu_possible_mask(tsk); struct cpuset *cs; - if (WARN_ON(!cpumask_and(pmask, possible_mask, cpu_online_mask))) - cpumask_copy(pmask, cpu_online_mask); + if (WARN_ON(!cpumask_and(pmask, possible_mask, cpu_active_mask))) + cpumask_copy(pmask, cpu_active_mask); rcu_read_lock(); cs = task_cs(tsk); - while (!cpumask_intersects(cs->effective_cpus, pmask)) { + while (!cpumask_intersects(cs->effective_cpus, pmask)) cs = parent_cs(cs); - if (unlikely(!cs)) { - /* - * The top cpuset doesn't have any online cpu as a - * consequence of a race between cpuset_hotplug_work - * and cpu hotplug notifier. But we know the top - * cpuset's effective_cpus is on its way to be - * identical to cpu_online_mask. - */ - goto out_unlock; - } - } - cpumask_and(pmask, pmask, cs->effective_cpus); -out_unlock: + cpumask_and(pmask, pmask, cs->effective_cpus); rcu_read_unlock(); } @@ -577,45 +410,6 @@ static void guarantee_online_mems(struct cpuset *cs, nodemask_t *pmask) nodes_and(*pmask, cs->effective_mems, node_states[N_MEMORY]); } -/* - * update task's spread flag if cpuset's page/slab spread flag is set - * - * Call with callback_lock or cpuset_mutex held. The check can be skipped - * if on default hierarchy. - */ -static void cpuset_update_task_spread_flags(struct cpuset *cs, - struct task_struct *tsk) -{ - if (cgroup_subsys_on_dfl(cpuset_cgrp_subsys)) - return; - - if (is_spread_page(cs)) - task_set_spread_page(tsk); - else - task_clear_spread_page(tsk); - - if (is_spread_slab(cs)) - task_set_spread_slab(tsk); - else - task_clear_spread_slab(tsk); -} - -/* - * is_cpuset_subset(p, q) - Is cpuset p a subset of cpuset q? - * - * One cpuset is a subset of another if all its allowed CPUs and - * Memory Nodes are a subset of the other, and its exclusive flags - * are only set if the other's are set. Call holding cpuset_mutex. - */ - -static int is_cpuset_subset(const struct cpuset *p, const struct cpuset *q) -{ - return cpumask_subset(p->cpus_allowed, q->cpus_allowed) && - nodes_subset(p->mems_allowed, q->mems_allowed) && - is_cpu_exclusive(p) <= is_cpu_exclusive(q) && - is_mem_exclusive(p) <= is_mem_exclusive(q); -} - /** * alloc_cpumasks - allocate three cpumasks for cpuset * @cs: the cpuset that have cpumasks to be allocated. @@ -718,11 +512,17 @@ static inline void free_cpuset(struct cpuset *cs) kfree(cs); } -static inline struct cpumask *fetch_xcpus(struct cpuset *cs) +/* Return user specified exclusive CPUs */ +static inline struct cpumask *user_xcpus(struct cpuset *cs) { - return !cpumask_empty(cs->exclusive_cpus) ? cs->exclusive_cpus : - cpumask_empty(cs->effective_xcpus) ? cs->cpus_allowed - : cs->effective_xcpus; + return cpumask_empty(cs->exclusive_cpus) ? cs->cpus_allowed + : cs->exclusive_cpus; +} + +static inline bool xcpus_empty(struct cpuset *cs) +{ + return cpumask_empty(cs->cpus_allowed) && + cpumask_empty(cs->exclusive_cpus); } /* @@ -732,8 +532,8 @@ static inline struct cpumask *fetch_xcpus(struct cpuset *cs) */ static inline bool cpusets_are_exclusive(struct cpuset *cs1, struct cpuset *cs2) { - struct cpumask *xcpus1 = fetch_xcpus(cs1); - struct cpumask *xcpus2 = fetch_xcpus(cs2); + struct cpumask *xcpus1 = user_xcpus(cs1); + struct cpumask *xcpus2 = user_xcpus(cs2); if (cpumask_intersects(xcpus1, xcpus2)) return false; @@ -741,35 +541,6 @@ static inline bool cpusets_are_exclusive(struct cpuset *cs1, struct cpuset *cs2) } /* - * validate_change_legacy() - Validate conditions specific to legacy (v1) - * behavior. - */ -static int validate_change_legacy(struct cpuset *cur, struct cpuset *trial) -{ - struct cgroup_subsys_state *css; - struct cpuset *c, *par; - int ret; - - WARN_ON_ONCE(!rcu_read_lock_held()); - - /* Each of our child cpusets must be a subset of us */ - ret = -EBUSY; - cpuset_for_each_child(c, css, cur) - if (!is_cpuset_subset(c, trial)) - goto out; - - /* On legacy hierarchy, we must be a subset of our parent cpuset. */ - ret = -EACCES; - par = parent_cs(cur); - if (par && !is_cpuset_subset(trial, par)) - goto out; - - ret = 0; -out: - return ret; -} - -/* * validate_change() - Used to validate that any proposed cpuset change * follows the structural rules for cpusets. * @@ -798,7 +569,7 @@ static int validate_change(struct cpuset *cur, struct cpuset *trial) rcu_read_lock(); if (!is_in_v2_mode()) - ret = validate_change_legacy(cur, trial); + ret = cpuset1_validate_change(cur, trial); if (ret) goto out; @@ -824,27 +595,63 @@ static int validate_change(struct cpuset *cur, struct cpuset *trial) /* * We can't shrink if we won't have enough room for SCHED_DEADLINE - * tasks. + * tasks. This check is not done when scheduling is disabled as the + * users should know what they are doing. + * + * For v1, effective_cpus == cpus_allowed & user_xcpus() returns + * cpus_allowed. + * + * For v2, is_cpu_exclusive() & is_sched_load_balance() are true only + * for non-isolated partition root. At this point, the target + * effective_cpus isn't computed yet. user_xcpus() is the best + * approximation. + * + * TBD: May need to precompute the real effective_cpus here in case + * incorrect scheduling of SCHED_DEADLINE tasks in a partition + * becomes an issue. */ ret = -EBUSY; - if (is_cpu_exclusive(cur) && - !cpuset_cpumask_can_shrink(cur->cpus_allowed, - trial->cpus_allowed)) + if (is_cpu_exclusive(cur) && is_sched_load_balance(cur) && + !cpuset_cpumask_can_shrink(cur->effective_cpus, user_xcpus(trial))) goto out; /* * If either I or some sibling (!= me) is exclusive, we can't - * overlap + * overlap. exclusive_cpus cannot overlap with each other if set. */ ret = -EINVAL; cpuset_for_each_child(c, css, par) { - if ((is_cpu_exclusive(trial) || is_cpu_exclusive(c)) && - c != cur) { + bool txset, cxset; /* Are exclusive_cpus set? */ + + if (c == cur) + continue; + + txset = !cpumask_empty(trial->exclusive_cpus); + cxset = !cpumask_empty(c->exclusive_cpus); + if (is_cpu_exclusive(trial) || is_cpu_exclusive(c) || + (txset && cxset)) { if (!cpusets_are_exclusive(trial, c)) goto out; + } else if (txset || cxset) { + struct cpumask *xcpus, *acpus; + + /* + * When just one of the exclusive_cpus's is set, + * cpus_allowed of the other cpuset, if set, cannot be + * a subset of it or none of those CPUs will be + * available if these exclusive CPUs are activated. + */ + if (txset) { + xcpus = trial->exclusive_cpus; + acpus = c->cpus_allowed; + } else { + xcpus = c->exclusive_cpus; + acpus = trial->cpus_allowed; + } + if (!cpumask_empty(acpus) && cpumask_subset(acpus, xcpus)) + goto out; } if ((is_mem_exclusive(trial) || is_mem_exclusive(c)) && - c != cur && nodes_intersects(trial->mems_allowed, c->mems_allowed)) goto out; } @@ -940,18 +747,15 @@ static inline int nr_cpusets(void) * were changed (added or removed.) * * Finding the best partition (set of domains): - * The triple nested loops below over i, j, k scan over the - * load balanced cpusets (using the array of cpuset pointers in - * csa[]) looking for pairs of cpusets that have overlapping - * cpus_allowed, but which don't have the same 'pn' partition - * number and gives them in the same partition number. It keeps - * looping on the 'restart' label until it can no longer find - * any such pairs. + * The double nested loops below over i, j scan over the load + * balanced cpusets (using the array of cpuset pointers in csa[]) + * looking for pairs of cpusets that have overlapping cpus_allowed + * and merging them using a union-find algorithm. + * + * The union of the cpus_allowed masks from the set of all cpusets + * having the same root then form the one element of the partition + * (one sched domain) to be passed to partition_sched_domains(). * - * The union of the cpus_allowed masks from the set of - * all cpusets having the same 'pn' value then form the one - * element of the partition (one sched domain) to be passed to - * partition_sched_domains(). */ static int generate_sched_domains(cpumask_var_t **domains, struct sched_domain_attr **attributes) @@ -959,20 +763,23 @@ static int generate_sched_domains(cpumask_var_t **domains, struct cpuset *cp; /* top-down scan of cpusets */ struct cpuset **csa; /* array of all cpuset ptrs */ int csn; /* how many cpuset ptrs in csa so far */ - int i, j, k; /* indices for partition finding loops */ + int i, j; /* indices for partition finding loops */ cpumask_var_t *doms; /* resulting partition; i.e. sched domains */ struct sched_domain_attr *dattr; /* attributes for custom domains */ int ndoms = 0; /* number of sched domains in result */ int nslot; /* next empty doms[] struct cpumask slot */ struct cgroup_subsys_state *pos_css; bool root_load_balance = is_sched_load_balance(&top_cpuset); + bool cgrpv2 = cpuset_v2(); + int nslot_update; doms = NULL; dattr = NULL; csa = NULL; /* Special case for the 99% of systems with one, full, sched domain */ - if (root_load_balance && !top_cpuset.nr_subparts) { + if (root_load_balance && cpumask_empty(subpartitions_cpus)) { +single_root_domain: ndoms = 1; doms = alloc_sched_domains(ndoms); if (!doms) @@ -1000,16 +807,18 @@ static int generate_sched_domains(cpumask_var_t **domains, cpuset_for_each_descendant_pre(cp, pos_css, &top_cpuset) { if (cp == &top_cpuset) continue; + + if (cgrpv2) + goto v2; + /* + * v1: * Continue traversing beyond @cp iff @cp has some CPUs and * isn't load balancing. The former is obvious. The * latter: All child cpusets contain a subset of the * parent's cpus, so just skip them, and then we call * update_domain_attr_tree() to calc relax_domain_level of * the corresponding sched domain. - * - * If root is load-balancing, we can skip @cp if it - * is a subset of the root's effective_cpus. */ if (!cpumask_empty(cp->cpus_allowed) && !(is_sched_load_balance(cp) && @@ -1017,47 +826,62 @@ static int generate_sched_domains(cpumask_var_t **domains, housekeeping_cpumask(HK_TYPE_DOMAIN)))) continue; - if (root_load_balance && - cpumask_subset(cp->cpus_allowed, top_cpuset.effective_cpus)) - continue; - if (is_sched_load_balance(cp) && !cpumask_empty(cp->effective_cpus)) csa[csn++] = cp; - /* skip @cp's subtree if not a partition root */ - if (!is_partition_valid(cp)) + /* skip @cp's subtree */ + pos_css = css_rightmost_descendant(pos_css); + continue; + +v2: + /* + * Only valid partition roots that are not isolated and with + * non-empty effective_cpus will be saved into csn[]. + */ + if ((cp->partition_root_state == PRS_ROOT) && + !cpumask_empty(cp->effective_cpus)) + csa[csn++] = cp; + + /* + * Skip @cp's subtree if not a partition root and has no + * exclusive CPUs to be granted to child cpusets. + */ + if (!is_partition_valid(cp) && cpumask_empty(cp->exclusive_cpus)) pos_css = css_rightmost_descendant(pos_css); } rcu_read_unlock(); + /* + * If there are only isolated partitions underneath the cgroup root, + * we can optimize out unneeded sched domains scanning. + */ + if (root_load_balance && (csn == 1)) + goto single_root_domain; + for (i = 0; i < csn; i++) - csa[i]->pn = i; - ndoms = csn; + uf_node_init(&csa[i]->node); -restart: - /* Find the best partition (set of sched domains) */ + /* Merge overlapping cpusets */ for (i = 0; i < csn; i++) { - struct cpuset *a = csa[i]; - int apn = a->pn; - - for (j = 0; j < csn; j++) { - struct cpuset *b = csa[j]; - int bpn = b->pn; - - if (apn != bpn && cpusets_overlap(a, b)) { - for (k = 0; k < csn; k++) { - struct cpuset *c = csa[k]; - - if (c->pn == bpn) - c->pn = apn; - } - ndoms--; /* one less element */ - goto restart; + for (j = i + 1; j < csn; j++) { + if (cpusets_overlap(csa[i], csa[j])) { + /* + * Cgroup v2 shouldn't pass down overlapping + * partition root cpusets. + */ + WARN_ON_ONCE(cgrpv2); + uf_union(&csa[i]->node, &csa[j]->node); } } } + /* Count the total number of domains */ + for (i = 0; i < csn; i++) { + if (uf_find(&csa[i]->node) == &csa[i]->node) + ndoms++; + } + /* * Now we know how many domains to create. * Convert <csn, csa> to <ndoms, doms> and populate cpu masks. @@ -1073,45 +897,48 @@ restart: dattr = kmalloc_array(ndoms, sizeof(struct sched_domain_attr), GFP_KERNEL); - for (nslot = 0, i = 0; i < csn; i++) { - struct cpuset *a = csa[i]; - struct cpumask *dp; - int apn = a->pn; - - if (apn < 0) { - /* Skip completed partitions */ - continue; - } - - dp = doms[nslot]; - - if (nslot == ndoms) { - static int warnings = 10; - if (warnings) { - pr_warn("rebuild_sched_domains confused: nslot %d, ndoms %d, csn %d, i %d, apn %d\n", - nslot, ndoms, csn, i, apn); - warnings--; - } - continue; + /* + * Cgroup v2 doesn't support domain attributes, just set all of them + * to SD_ATTR_INIT. Also non-isolating partition root CPUs are a + * subset of HK_TYPE_DOMAIN housekeeping CPUs. + */ + if (cgrpv2) { + for (i = 0; i < ndoms; i++) { + /* + * The top cpuset may contain some boot time isolated + * CPUs that need to be excluded from the sched domain. + */ + if (csa[i] == &top_cpuset) + cpumask_and(doms[i], csa[i]->effective_cpus, + housekeeping_cpumask(HK_TYPE_DOMAIN)); + else + cpumask_copy(doms[i], csa[i]->effective_cpus); + if (dattr) + dattr[i] = SD_ATTR_INIT; } + goto done; + } - cpumask_clear(dp); - if (dattr) - *(dattr + nslot) = SD_ATTR_INIT; + for (nslot = 0, i = 0; i < csn; i++) { + nslot_update = 0; for (j = i; j < csn; j++) { - struct cpuset *b = csa[j]; - - if (apn == b->pn) { - cpumask_or(dp, dp, b->effective_cpus); + if (uf_find(&csa[j]->node) == &csa[i]->node) { + struct cpumask *dp = doms[nslot]; + + if (i == j) { + nslot_update = 1; + cpumask_clear(dp); + if (dattr) + *(dattr + nslot) = SD_ATTR_INIT; + } + cpumask_or(dp, dp, csa[j]->effective_cpus); cpumask_and(dp, dp, housekeeping_cpumask(HK_TYPE_DOMAIN)); if (dattr) - update_domain_attr_tree(dattr + nslot, b); - - /* Done with this partition */ - b->pn = -1; + update_domain_attr_tree(dattr + nslot, csa[j]); } } - nslot++; + if (nslot_update) + nslot++; } BUG_ON(nslot != ndoms); @@ -1146,10 +973,12 @@ static void dl_update_tasks_root_domain(struct cpuset *cs) css_task_iter_end(&it); } -static void dl_rebuild_rd_accounting(void) +void dl_rebuild_rd_accounting(void) { struct cpuset *cs = NULL; struct cgroup_subsys_state *pos_css; + int cpu; + u64 cookie = ++dl_cookie; lockdep_assert_held(&cpuset_mutex); lockdep_assert_cpus_held(); @@ -1157,11 +986,12 @@ static void dl_rebuild_rd_accounting(void) rcu_read_lock(); - /* - * Clear default root domain DL accounting, it will be computed again - * if a task belongs to it. - */ - dl_clear_root_domain(&def_root_domain); + for_each_possible_cpu(cpu) { + if (dl_bw_visited(cpu, cookie)) + continue; + + dl_clear_root_domain_cpu(cpu); + } cpuset_for_each_descendant_pre(cs, pos_css, &top_cpuset) { @@ -1182,16 +1012,6 @@ static void dl_rebuild_rd_accounting(void) rcu_read_unlock(); } -static void -partition_and_rebuild_sched_domains(int ndoms_new, cpumask_var_t doms_new[], - struct sched_domain_attr *dattr_new) -{ - mutex_lock(&sched_domains_mutex); - partition_sched_domains_locked(ndoms_new, doms_new, dattr_new); - dl_rebuild_rd_accounting(); - mutex_unlock(&sched_domains_mutex); -} - /* * Rebuild scheduler domains. * @@ -1203,7 +1023,7 @@ partition_and_rebuild_sched_domains(int ndoms_new, cpumask_var_t doms_new[], * * Call with cpuset_mutex held. Takes cpus_read_lock(). */ -static void rebuild_sched_domains_locked(void) +void rebuild_sched_domains_locked(void) { struct cgroup_subsys_state *pos_css; struct sched_domain_attr *attr; @@ -1213,11 +1033,12 @@ static void rebuild_sched_domains_locked(void) lockdep_assert_cpus_held(); lockdep_assert_held(&cpuset_mutex); + force_sd_rebuild = false; /* * If we have raced with CPU hotplug, return early to avoid * passing doms with offlined cpu to partition_sched_domains(). - * Anyways, cpuset_hotplug_workfn() will rebuild sched domains. + * Anyways, cpuset_handle_hotplug() will rebuild sched domains. * * With no CPUs in any subpartitions, top_cpuset's effective CPUs * should be the same as the active CPUs, so checking only top_cpuset @@ -1232,7 +1053,7 @@ static void rebuild_sched_domains_locked(void) * root should be only a subset of the active CPUs. Since a CPU in any * partition root could be offlined, all must be checked. */ - if (top_cpuset.nr_subparts) { + if (!cpumask_empty(subpartitions_cpus)) { rcu_read_lock(); cpuset_for_each_descendant_pre(cs, pos_css, &top_cpuset) { if (!is_partition_valid(cs)) { @@ -1252,35 +1073,52 @@ static void rebuild_sched_domains_locked(void) ndoms = generate_sched_domains(&doms, &attr); /* Have scheduler rebuild the domains */ - partition_and_rebuild_sched_domains(ndoms, doms, attr); + partition_sched_domains(ndoms, doms, attr); } #else /* !CONFIG_SMP */ -static void rebuild_sched_domains_locked(void) +void rebuild_sched_domains_locked(void) { } #endif /* CONFIG_SMP */ -void rebuild_sched_domains(void) +static void rebuild_sched_domains_cpuslocked(void) { - cpus_read_lock(); mutex_lock(&cpuset_mutex); rebuild_sched_domains_locked(); mutex_unlock(&cpuset_mutex); +} + +void rebuild_sched_domains(void) +{ + cpus_read_lock(); + rebuild_sched_domains_cpuslocked(); cpus_read_unlock(); } +void cpuset_reset_sched_domains(void) +{ + mutex_lock(&cpuset_mutex); + partition_sched_domains(1, NULL, NULL); + mutex_unlock(&cpuset_mutex); +} + /** - * update_tasks_cpumask - Update the cpumasks of tasks in the cpuset. + * cpuset_update_tasks_cpumask - Update the cpumasks of tasks in the cpuset. * @cs: the cpuset in which each task's cpus_allowed mask needs to be changed * @new_cpus: the temp variable for the new effective_cpus mask * * Iterate through each task of @cs updating its cpus_allowed to the * effective cpuset's. As this function is called with cpuset_mutex held, - * cpuset membership stays stable. For top_cpuset, task_cpu_possible_mask() - * is used instead of effective_cpus to make sure all offline CPUs are also - * included as hotplug code won't update cpumasks for tasks in top_cpuset. + * cpuset membership stays stable. + * + * For top_cpuset, task_cpu_possible_mask() is used instead of effective_cpus + * to make sure all offline CPUs are also included as hotplug code won't + * update cpumasks for tasks in top_cpuset. + * + * As task_cpu_possible_mask() can be task dependent in arm64, we have to + * do cpu masking per task instead of doing it once for all. */ -static void update_tasks_cpumask(struct cpuset *cs, struct cpumask *new_cpus) +void cpuset_update_tasks_cpumask(struct cpuset *cs, struct cpumask *new_cpus) { struct css_task_iter it; struct task_struct *task; @@ -1292,9 +1130,11 @@ static void update_tasks_cpumask(struct cpuset *cs, struct cpumask *new_cpus) if (top_cs) { /* - * Percpu kthreads in top_cpuset are ignored + * PF_NO_SETAFFINITY tasks are ignored. + * All per cpu kthreads should have PF_NO_SETAFFINITY + * flag set, see kthread_set_per_cpu(). */ - if (kthread_is_per_cpu(task)) + if (task->flags & PF_NO_SETAFFINITY) continue; cpumask_andnot(new_cpus, possible_mask, subpartitions_cpus); } else { @@ -1330,8 +1170,6 @@ enum partition_cmd { partcmd_invalidate, /* Make partition invalid */ }; -static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, - int turning_on); static void update_sibling_cpumasks(struct cpuset *parent, struct cpuset *cs, struct tmpmasks *tmp); @@ -1340,16 +1178,16 @@ static void update_sibling_cpumasks(struct cpuset *parent, struct cpuset *cs, * * Return: 0 if successful, an error code otherwise */ -static int update_partition_exclusive(struct cpuset *cs, int new_prs) +static int update_partition_exclusive_flag(struct cpuset *cs, int new_prs) { - bool exclusive = (new_prs > 0); + bool exclusive = (new_prs > PRS_MEMBER); if (exclusive && !is_cpu_exclusive(cs)) { - if (update_flag(CS_CPU_EXCLUSIVE, cs, 1)) + if (cpuset_update_flag(CS_CPU_EXCLUSIVE, cs, 1)) return PERR_NOTEXCL; } else if (!exclusive && is_cpu_exclusive(cs)) { /* Turning off CS_CPU_EXCLUSIVE will not return error */ - update_flag(CS_CPU_EXCLUSIVE, cs, 0); + cpuset_update_flag(CS_CPU_EXCLUSIVE, cs, 0); } return 0; } @@ -1385,7 +1223,7 @@ static void update_partition_sd_lb(struct cpuset *cs, int old_prs) } if (rebuild_domains) - rebuild_sched_domains_locked(); + cpuset_force_rebuild(); } /* @@ -1407,7 +1245,7 @@ static void reset_partition_data(struct cpuset *cs) { struct cpuset *parent = parent_cs(cs); - if (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys)) + if (!cpuset_v2()) return; lockdep_assert_held(&callback_lock); @@ -1418,21 +1256,17 @@ static void reset_partition_data(struct cpuset *cs) if (is_cpu_exclusive(cs)) clear_bit(CS_CPU_EXCLUSIVE, &cs->flags); } - if (!cpumask_and(cs->effective_cpus, - parent->effective_cpus, cs->cpus_allowed)) { - cs->use_parent_ecpus = true; - parent->child_ecpus_count++; + if (!cpumask_and(cs->effective_cpus, parent->effective_cpus, cs->cpus_allowed)) cpumask_copy(cs->effective_cpus, parent->effective_cpus); - } } /* - * partition_xcpus_newstate - Exclusive CPUs state change + * isolated_cpus_update - Update the isolated_cpus mask * @old_prs: old partition_root_state * @new_prs: new partition_root_state * @xcpus: exclusive CPUs with state change */ -static void partition_xcpus_newstate(int old_prs, int new_prs, struct cpumask *xcpus) +static void isolated_cpus_update(int old_prs, int new_prs, struct cpumask *xcpus) { WARN_ON_ONCE(old_prs == new_prs); if (new_prs == PRS_ISOLATED) @@ -1466,8 +1300,8 @@ static bool partition_xcpus_add(int new_prs, struct cpuset *parent, isolcpus_updated = (new_prs != parent->partition_root_state); if (isolcpus_updated) - partition_xcpus_newstate(parent->partition_root_state, new_prs, - xcpus); + isolated_cpus_update(parent->partition_root_state, new_prs, + xcpus); cpumask_andnot(parent->effective_cpus, parent->effective_cpus, xcpus); return isolcpus_updated; @@ -1497,8 +1331,8 @@ static bool partition_xcpus_del(int old_prs, struct cpuset *parent, isolcpus_updated = (old_prs != parent->partition_root_state); if (isolcpus_updated) - partition_xcpus_newstate(old_prs, parent->partition_root_state, - xcpus); + isolated_cpus_update(old_prs, parent->partition_root_state, + xcpus); cpumask_and(xcpus, xcpus, cpu_active_mask); cpumask_or(parent->effective_cpus, parent->effective_cpus, xcpus); @@ -1533,25 +1367,55 @@ EXPORT_SYMBOL_GPL(cpuset_cpu_is_isolated); * compute_effective_exclusive_cpumask - compute effective exclusive CPUs * @cs: cpuset * @xcpus: effective exclusive CPUs value to be set - * Return: true if xcpus is not empty, false otherwise. + * @real_cs: the real cpuset (can be NULL) + * Return: 0 if there is no sibling conflict, > 0 otherwise * - * Starting with exclusive_cpus (cpus_allowed if exclusive_cpus is not set), - * it must be a subset of cpus_allowed and parent's effective_xcpus. + * If exclusive_cpus isn't explicitly set or a real_cs is provided, we have to + * scan the sibling cpusets and exclude their exclusive_cpus or effective_xcpus + * as well. The provision of real_cs means that a cpumask is being changed and + * the given cs is a trial one. */ -static bool compute_effective_exclusive_cpumask(struct cpuset *cs, - struct cpumask *xcpus) +static int compute_effective_exclusive_cpumask(struct cpuset *cs, + struct cpumask *xcpus, + struct cpuset *real_cs) { + struct cgroup_subsys_state *css; struct cpuset *parent = parent_cs(cs); + struct cpuset *sibling; + int retval = 0; if (!xcpus) xcpus = cs->effective_xcpus; - if (!cpumask_empty(cs->exclusive_cpus)) - cpumask_and(xcpus, cs->exclusive_cpus, cs->cpus_allowed); - else - cpumask_copy(xcpus, cs->cpus_allowed); + cpumask_and(xcpus, user_xcpus(cs), parent->effective_xcpus); + + if (!real_cs) { + if (!cpumask_empty(cs->exclusive_cpus)) + return 0; + } else { + cs = real_cs; + } + + /* + * Exclude exclusive CPUs from siblings + */ + rcu_read_lock(); + cpuset_for_each_child(sibling, css, parent) { + if (sibling == cs) + continue; - return cpumask_and(xcpus, xcpus, parent->effective_xcpus); + if (cpumask_intersects(xcpus, sibling->exclusive_cpus)) { + cpumask_andnot(xcpus, xcpus, sibling->exclusive_cpus); + retval++; + continue; + } + if (cpumask_intersects(xcpus, sibling->effective_xcpus)) { + cpumask_andnot(xcpus, xcpus, sibling->effective_xcpus); + retval++; + } + } + rcu_read_unlock(); + return retval; } static inline bool is_remote_partition(struct cpuset *cs) @@ -1568,8 +1432,8 @@ static inline bool is_local_partition(struct cpuset *cs) * remote_partition_enable - Enable current cpuset as a remote partition root * @cs: the cpuset to update * @new_prs: new partition_root_state - * @tmp: temparary masks - * Return: 1 if successful, 0 if error + * @tmp: temporary masks + * Return: 0 if successful, errcode if error * * Enable the current cpuset to become a remote partition root taking CPUs * directly from the top cpuset. cpuset_mutex must be held by the caller. @@ -1583,46 +1447,45 @@ static int remote_partition_enable(struct cpuset *cs, int new_prs, * The user must have sysadmin privilege. */ if (!capable(CAP_SYS_ADMIN)) - return 0; + return PERR_ACCESS; /* * The requested exclusive_cpus must not be allocated to other * partitions and it can't use up all the root's effective_cpus. * - * Note that if there is any local partition root above it or - * remote partition root underneath it, its exclusive_cpus must - * have overlapped with subpartitions_cpus. + * The effective_xcpus mask can contain offline CPUs, but there must + * be at least one or more online CPUs present before it can be enabled. + * + * Note that creating a remote partition with any local partition root + * above it or remote partition root underneath it is not allowed. */ - compute_effective_exclusive_cpumask(cs, tmp->new_cpus); - if (cpumask_empty(tmp->new_cpus) || - cpumask_intersects(tmp->new_cpus, subpartitions_cpus) || + compute_effective_exclusive_cpumask(cs, tmp->new_cpus, NULL); + WARN_ON_ONCE(cpumask_intersects(tmp->new_cpus, subpartitions_cpus)); + if (!cpumask_intersects(tmp->new_cpus, cpu_active_mask) || cpumask_subset(top_cpuset.effective_cpus, tmp->new_cpus)) - return 0; + return PERR_INVCPUS; spin_lock_irq(&callback_lock); isolcpus_updated = partition_xcpus_add(new_prs, NULL, tmp->new_cpus); list_add(&cs->remote_sibling, &remote_children); - if (cs->use_parent_ecpus) { - struct cpuset *parent = parent_cs(cs); - - cs->use_parent_ecpus = false; - parent->child_ecpus_count--; - } + cpumask_copy(cs->effective_xcpus, tmp->new_cpus); spin_unlock_irq(&callback_lock); update_unbound_workqueue_cpumask(isolcpus_updated); + cpuset_force_rebuild(); + cs->prs_err = 0; /* - * Proprogate changes in top_cpuset's effective_cpus down the hierarchy. + * Propagate changes in top_cpuset's effective_cpus down the hierarchy. */ - update_tasks_cpumask(&top_cpuset, tmp->new_cpus); + cpuset_update_tasks_cpumask(&top_cpuset, tmp->new_cpus); update_sibling_cpumasks(&top_cpuset, NULL, tmp); - return 1; + return 0; } /* * remote_partition_disable - Remove current cpuset from remote partition list * @cs: the cpuset to update - * @tmp: temparary masks + * @tmp: temporary masks * * The effective_cpus is also updated. * @@ -1632,39 +1495,44 @@ static void remote_partition_disable(struct cpuset *cs, struct tmpmasks *tmp) { bool isolcpus_updated; - compute_effective_exclusive_cpumask(cs, tmp->new_cpus); WARN_ON_ONCE(!is_remote_partition(cs)); - WARN_ON_ONCE(!cpumask_subset(tmp->new_cpus, subpartitions_cpus)); + WARN_ON_ONCE(!cpumask_subset(cs->effective_xcpus, subpartitions_cpus)); spin_lock_irq(&callback_lock); list_del_init(&cs->remote_sibling); isolcpus_updated = partition_xcpus_del(cs->partition_root_state, - NULL, tmp->new_cpus); - cs->partition_root_state = -cs->partition_root_state; - if (!cs->prs_err) - cs->prs_err = PERR_INVCPUS; + NULL, cs->effective_xcpus); + if (cs->prs_err) + cs->partition_root_state = -cs->partition_root_state; + else + cs->partition_root_state = PRS_MEMBER; + + /* effective_xcpus may need to be changed */ + compute_effective_exclusive_cpumask(cs, NULL, NULL); reset_partition_data(cs); spin_unlock_irq(&callback_lock); update_unbound_workqueue_cpumask(isolcpus_updated); + cpuset_force_rebuild(); /* - * Proprogate changes in top_cpuset's effective_cpus down the hierarchy. + * Propagate changes in top_cpuset's effective_cpus down the hierarchy. */ - update_tasks_cpumask(&top_cpuset, tmp->new_cpus); + cpuset_update_tasks_cpumask(&top_cpuset, tmp->new_cpus); update_sibling_cpumasks(&top_cpuset, NULL, tmp); } /* * remote_cpus_update - cpus_exclusive change of remote partition * @cs: the cpuset to be updated - * @newmask: the new effective_xcpus mask - * @tmp: temparary masks + * @xcpus: the new exclusive_cpus mask, if non-NULL + * @excpus: the new effective_xcpus mask + * @tmp: temporary masks * * top_cpuset and subpartitions_cpus will be updated or partition can be * invalidated. */ -static void remote_cpus_update(struct cpuset *cs, struct cpumask *newmask, - struct tmpmasks *tmp) +static void remote_cpus_update(struct cpuset *cs, struct cpumask *xcpus, + struct cpumask *excpus, struct tmpmasks *tmp) { bool adding, deleting; int prs = cs->partition_root_state; @@ -1675,34 +1543,51 @@ static void remote_cpus_update(struct cpuset *cs, struct cpumask *newmask, WARN_ON_ONCE(!cpumask_subset(cs->effective_xcpus, subpartitions_cpus)); - if (cpumask_empty(newmask)) + if (cpumask_empty(excpus)) { + cs->prs_err = PERR_CPUSEMPTY; goto invalidate; + } - adding = cpumask_andnot(tmp->addmask, newmask, cs->effective_xcpus); - deleting = cpumask_andnot(tmp->delmask, cs->effective_xcpus, newmask); + adding = cpumask_andnot(tmp->addmask, excpus, cs->effective_xcpus); + deleting = cpumask_andnot(tmp->delmask, cs->effective_xcpus, excpus); /* * Additions of remote CPUs is only allowed if those CPUs are * not allocated to other partitions and there are effective_cpus * left in the top cpuset. */ - if (adding && (!capable(CAP_SYS_ADMIN) || - cpumask_intersects(tmp->addmask, subpartitions_cpus) || - cpumask_subset(top_cpuset.effective_cpus, tmp->addmask))) - goto invalidate; + if (adding) { + WARN_ON_ONCE(cpumask_intersects(tmp->addmask, subpartitions_cpus)); + if (!capable(CAP_SYS_ADMIN)) + cs->prs_err = PERR_ACCESS; + else if (cpumask_intersects(tmp->addmask, subpartitions_cpus) || + cpumask_subset(top_cpuset.effective_cpus, tmp->addmask)) + cs->prs_err = PERR_NOCPUS; + if (cs->prs_err) + goto invalidate; + } spin_lock_irq(&callback_lock); if (adding) isolcpus_updated += partition_xcpus_add(prs, NULL, tmp->addmask); if (deleting) isolcpus_updated += partition_xcpus_del(prs, NULL, tmp->delmask); + /* + * Need to update effective_xcpus and exclusive_cpus now as + * update_sibling_cpumasks() below may iterate back to the same cs. + */ + cpumask_copy(cs->effective_xcpus, excpus); + if (xcpus) + cpumask_copy(cs->exclusive_cpus, xcpus); spin_unlock_irq(&callback_lock); update_unbound_workqueue_cpumask(isolcpus_updated); + if (adding || deleting) + cpuset_force_rebuild(); /* - * Proprogate changes in top_cpuset's effective_cpus down the hierarchy. + * Propagate changes in top_cpuset's effective_cpus down the hierarchy. */ - update_tasks_cpumask(&top_cpuset, tmp->new_cpus); + cpuset_update_tasks_cpumask(&top_cpuset, tmp->new_cpus); update_sibling_cpumasks(&top_cpuset, NULL, tmp); return; @@ -1711,61 +1596,20 @@ invalidate: } /* - * remote_partition_check - check if a child remote partition needs update - * @cs: the cpuset to be updated - * @newmask: the new effective_xcpus mask - * @delmask: temporary mask for deletion (not in tmp) - * @tmp: temparary masks - * - * This should be called before the given cs has updated its cpus_allowed - * and/or effective_xcpus. - */ -static void remote_partition_check(struct cpuset *cs, struct cpumask *newmask, - struct cpumask *delmask, struct tmpmasks *tmp) -{ - struct cpuset *child, *next; - int disable_cnt = 0; - - /* - * Compute the effective exclusive CPUs that will be deleted. - */ - if (!cpumask_andnot(delmask, cs->effective_xcpus, newmask) || - !cpumask_intersects(delmask, subpartitions_cpus)) - return; /* No deletion of exclusive CPUs in partitions */ - - /* - * Searching the remote children list to look for those that will - * be impacted by the deletion of exclusive CPUs. - * - * Since a cpuset must be removed from the remote children list - * before it can go offline and holding cpuset_mutex will prevent - * any change in cpuset status. RCU read lock isn't needed. - */ - lockdep_assert_held(&cpuset_mutex); - list_for_each_entry_safe(child, next, &remote_children, remote_sibling) - if (cpumask_intersects(child->effective_cpus, delmask)) { - remote_partition_disable(child, tmp); - disable_cnt++; - } - if (disable_cnt) - rebuild_sched_domains_locked(); -} - -/* * prstate_housekeeping_conflict - check for partition & housekeeping conflicts * @prstate: partition root state to be checked * @new_cpus: cpu mask * Return: true if there is conflict, false otherwise * - * CPUs outside of housekeeping_cpumask(HK_TYPE_DOMAIN) can only be used in - * an isolated partition. + * CPUs outside of boot_hk_cpus, if defined, can only be used in an + * isolated partition. */ static bool prstate_housekeeping_conflict(int prstate, struct cpumask *new_cpus) { - const struct cpumask *hk_domain = housekeeping_cpumask(HK_TYPE_DOMAIN); - bool all_in_hk = cpumask_subset(new_cpus, hk_domain); + if (!have_boot_isolcpus) + return false; - if (!all_in_hk && (prstate != PRS_ISOLATED)) + if ((prstate != PRS_ISOLATED) && !cpumask_subset(new_cpus, boot_hk_cpus)) return true; return false; @@ -1805,7 +1649,7 @@ static bool prstate_housekeeping_conflict(int prstate, struct cpumask *new_cpus) * The partcmd_update command is used by update_cpumasks_hier() with newmask * NULL and update_cpumask() with newmask set. The partcmd_invalidate is used * by update_cpumask() with NULL newmask. In both cases, the callers won't - * check for error and so partition_root_state and prs_error will be updated + * check for error and so partition_root_state and prs_err will be updated * directly. */ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd, @@ -1818,11 +1662,12 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd, int old_prs, new_prs; int part_error = PERR_NONE; /* Partition error? */ int subparts_delta = 0; - struct cpumask *xcpus; /* cs effective_xcpus */ int isolcpus_updated = 0; + struct cpumask *xcpus = user_xcpus(cs); bool nocpu; lockdep_assert_held(&cpuset_mutex); + WARN_ON_ONCE(is_remote_partition(cs)); /* For local partition only */ /* * new_prs will only be changed for the partcmd_update and @@ -1830,8 +1675,6 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd, */ adding = deleting = false; old_prs = new_prs = cs->partition_root_state; - xcpus = !cpumask_empty(cs->exclusive_cpus) - ? cs->effective_xcpus : cs->cpus_allowed; if (cmd == partcmd_invalidate) { if (is_prs_invalid(old_prs)) @@ -1859,19 +1702,26 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd, return is_partition_invalid(parent) ? PERR_INVPARENT : PERR_NOTPART; } - if (!newmask && cpumask_empty(cs->cpus_allowed)) + if (!newmask && xcpus_empty(cs)) return PERR_CPUSEMPTY; nocpu = tasks_nocpu_error(parent, cs, xcpus); if ((cmd == partcmd_enable) || (cmd == partcmd_enablei)) { /* + * Need to call compute_effective_exclusive_cpumask() in case + * exclusive_cpus not set. Sibling conflict should only happen + * if exclusive_cpus isn't set. + */ + xcpus = tmp->delmask; + if (compute_effective_exclusive_cpumask(cs, xcpus, NULL)) + WARN_ON_ONCE(!cpumask_empty(cs->exclusive_cpus)); + + /* * Enabling partition root is not allowed if its - * effective_xcpus is empty or doesn't overlap with - * parent's effective_xcpus. + * effective_xcpus is empty. */ - if (cpumask_empty(xcpus) || - !cpumask_intersects(xcpus, parent->effective_xcpus)) + if (cpumask_empty(xcpus)) return PERR_INVCPUS; if (prstate_housekeeping_conflict(new_prs, xcpus)) @@ -1884,19 +1734,33 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd, if (nocpu) return PERR_NOCPUS; - cpumask_copy(tmp->delmask, xcpus); + /* + * This function will only be called when all the preliminary + * checks have passed. At this point, the following condition + * should hold. + * + * (cs->effective_xcpus & cpu_active_mask) ⊆ parent->effective_cpus + * + * Warn if it is not the case. + */ + cpumask_and(tmp->new_cpus, xcpus, cpu_active_mask); + WARN_ON_ONCE(!cpumask_subset(tmp->new_cpus, parent->effective_cpus)); + deleting = true; subparts_delta++; new_prs = (cmd == partcmd_enable) ? PRS_ROOT : PRS_ISOLATED; } else if (cmd == partcmd_disable) { /* - * May need to add cpus to parent's effective_cpus for - * valid partition root. + * May need to add cpus back to parent's effective_cpus + * (and maybe removed from subpartitions_cpus/isolated_cpus) + * for valid partition root. xcpus may contain CPUs that + * shouldn't be removed from the two global cpumasks. */ - adding = !is_prs_invalid(old_prs) && - cpumask_and(tmp->addmask, xcpus, parent->effective_xcpus); - if (adding) + if (is_partition_valid(cs)) { + cpumask_copy(tmp->addmask, cs->effective_xcpus); + adding = true; subparts_delta--; + } new_prs = PRS_MEMBER; } else if (newmask) { /* @@ -1907,6 +1771,9 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd, goto write_error; } + /* Check newmask again, whether cpus are available for parent/cs */ + nocpu |= tasks_nocpu_error(parent, cs, newmask); + /* * partcmd_update with newmask: * @@ -1935,6 +1802,15 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd, parent->effective_xcpus); } /* + * The new CPUs to be removed from parent's effective CPUs + * must be present. + */ + if (deleting) { + cpumask_and(tmp->new_cpus, tmp->delmask, cpu_active_mask); + WARN_ON_ONCE(!cpumask_subset(tmp->new_cpus, parent->effective_cpus)); + } + + /* * Make partition invalid if parent's effective_cpus could * become empty and there are tasks in the parent. */ @@ -2032,7 +1908,7 @@ write_error: * CPU lists in cs haven't been updated yet. So defer it to later. */ if ((old_prs != new_prs) && (cmd != partcmd_update)) { - int err = update_partition_exclusive(cs, new_prs); + int err = update_partition_exclusive_flag(cs, new_prs); if (err) return err; @@ -2070,23 +1946,20 @@ write_error: update_unbound_workqueue_cpumask(isolcpus_updated); if ((old_prs != new_prs) && (cmd == partcmd_update)) - update_partition_exclusive(cs, new_prs); + update_partition_exclusive_flag(cs, new_prs); if (adding || deleting) { - update_tasks_cpumask(parent, tmp->addmask); + cpuset_update_tasks_cpumask(parent, tmp->addmask); update_sibling_cpumasks(parent, cs, tmp); } /* * For partcmd_update without newmask, it is being called from - * cpuset_hotplug_workfn() where cpus_read_lock() wasn't taken. - * Update the load balance flag and scheduling domain if - * cpus_read_trylock() is successful. + * cpuset_handle_hotplug(). Update the load balance flag and + * scheduling domain accordingly. */ - if ((cmd == partcmd_update) && !newmask && cpus_read_trylock()) { + if ((cmd == partcmd_update) && !newmask) update_partition_sd_lb(cs, old_prs); - cpus_read_unlock(); - } notify_partition_change(cs, old_prs); return 0; @@ -2123,7 +1996,7 @@ static void compute_partition_effective_cpumask(struct cpuset *cs, * 2) All the effective_cpus will be used up and cp * has tasks */ - compute_effective_exclusive_cpumask(cs, new_ecpus); + compute_effective_exclusive_cpumask(cs, new_ecpus, NULL); cpumask_and(new_ecpus, new_ecpus, cpu_active_mask); rcu_read_lock(); @@ -2131,6 +2004,11 @@ static void compute_partition_effective_cpumask(struct cpuset *cs, if (!is_partition_valid(child)) continue; + /* + * There shouldn't be a remote partition underneath another + * partition root. + */ + WARN_ON_ONCE(is_remote_partition(child)); child->prs_err = 0; if (!cpumask_subset(child->effective_xcpus, cs->effective_xcpus)) @@ -2160,12 +2038,6 @@ static void compute_partition_effective_cpumask(struct cpuset *cs, } /* - * update_cpumasks_hier() flags - */ -#define HIER_CHECKALL 0x01 /* Check all cpusets with no skipping */ -#define HIER_NO_SD_REBUILD 0x02 /* Don't rebuild sched domains */ - -/* * update_cpumasks_hier - Update effective cpumasks and tasks in the subtree * @cs: the cpuset to consider * @tmp: temp variables for calculating effective_cpus & partition setup @@ -2179,7 +2051,7 @@ static void compute_partition_effective_cpumask(struct cpuset *cs, * Called with cpuset_mutex held */ static void update_cpumasks_hier(struct cpuset *cs, struct tmpmasks *tmp, - int flags) + bool force) { struct cpuset *cp; struct cgroup_subsys_state *pos_css; @@ -2192,32 +2064,39 @@ static void update_cpumasks_hier(struct cpuset *cs, struct tmpmasks *tmp, bool remote = is_remote_partition(cp); bool update_parent = false; + old_prs = new_prs = cp->partition_root_state; + /* - * Skip descendent remote partition that acquires CPUs - * directly from top cpuset unless it is cs. + * For child remote partition root (!= cs), we need to call + * remote_cpus_update() if effective_xcpus will be changed. + * Otherwise, we can skip the whole subtree. + * + * remote_cpus_update() will reuse tmp->new_cpus only after + * its value is being processed. */ if (remote && (cp != cs)) { - pos_css = css_rightmost_descendant(pos_css); - continue; - } + compute_effective_exclusive_cpumask(cp, tmp->new_cpus, NULL); + if (cpumask_equal(cp->effective_xcpus, tmp->new_cpus)) { + pos_css = css_rightmost_descendant(pos_css); + continue; + } + rcu_read_unlock(); + remote_cpus_update(cp, NULL, tmp->new_cpus, tmp); + rcu_read_lock(); - /* - * Update effective_xcpus if exclusive_cpus set. - * The case when exclusive_cpus isn't set is handled later. - */ - if (!cpumask_empty(cp->exclusive_cpus) && (cp != cs)) { - spin_lock_irq(&callback_lock); - compute_effective_exclusive_cpumask(cp, NULL); - spin_unlock_irq(&callback_lock); + /* Remote partition may be invalidated */ + new_prs = cp->partition_root_state; + remote = (new_prs == old_prs); } - old_prs = new_prs = cp->partition_root_state; - if (remote || (is_partition_valid(parent) && - is_partition_valid(cp))) + if (remote || (is_partition_valid(parent) && is_partition_valid(cp))) compute_partition_effective_cpumask(cp, tmp->new_cpus); else compute_effective_cpumask(tmp->new_cpus, cp, parent); + if (remote) + goto get_css; /* Ready to update cpuset data */ + /* * A partition with no effective_cpus is allowed as long as * there is no task associated with it. Call @@ -2234,31 +2113,19 @@ static void update_cpumasks_hier(struct cpuset *cs, struct tmpmasks *tmp, * it is a partition root that has explicitly distributed * out all its CPUs. */ - if (is_in_v2_mode() && !remote && cpumask_empty(tmp->new_cpus)) { + if (is_in_v2_mode() && !remote && cpumask_empty(tmp->new_cpus)) cpumask_copy(tmp->new_cpus, parent->effective_cpus); - if (!cp->use_parent_ecpus) { - cp->use_parent_ecpus = true; - parent->child_ecpus_count++; - } - } else if (cp->use_parent_ecpus) { - cp->use_parent_ecpus = false; - WARN_ON_ONCE(!parent->child_ecpus_count); - parent->child_ecpus_count--; - } - - if (remote) - goto get_css; /* * Skip the whole subtree if * 1) the cpumask remains the same, * 2) has no partition root state, - * 3) HIER_CHECKALL flag not set, and + * 3) force flag not set, and * 4) for v2 load balance state same as its parent. */ - if (!cp->partition_root_state && !(flags & HIER_CHECKALL) && + if (!cp->partition_root_state && !force && cpumask_equal(tmp->new_cpus, cp->effective_cpus) && - (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys) || + (!cpuset_v2() || (is_sched_load_balance(parent) == is_sched_load_balance(cp)))) { pos_css = css_rightmost_descendant(pos_css); continue; @@ -2268,7 +2135,7 @@ update_parent_effective: /* * update_parent_effective_cpumask() should have been called * for cs already in update_cpumask(). We should also call - * update_tasks_cpumask() again for tasks in the parent + * cpuset_update_tasks_cpumask() again for tasks in the parent * cpuset if the parent's effective_cpus changes. */ if ((cp != cs) && old_prs) { @@ -2309,6 +2176,9 @@ get_css: spin_lock_irq(&callback_lock); cpumask_copy(cp->effective_cpus, tmp->new_cpus); cp->partition_root_state = new_prs; + if (!cpumask_empty(cp->exclusive_cpus) && (cp != cs)) + compute_effective_exclusive_cpumask(cp, NULL, NULL); + /* * Make sure effective_xcpus is properly set for a valid * partition root. @@ -2325,15 +2195,14 @@ get_css: WARN_ON(!is_in_v2_mode() && !cpumask_equal(cp->cpus_allowed, cp->effective_cpus)); - update_tasks_cpumask(cp, cp->effective_cpus); + cpuset_update_tasks_cpumask(cp, cp->effective_cpus); /* * On default hierarchy, inherit the CS_SCHED_LOAD_BALANCE * from parent if current cpuset isn't a valid partition root * and their load balance states differ. */ - if (cgroup_subsys_on_dfl(cpuset_cgrp_subsys) && - !is_partition_valid(cp) && + if (cpuset_v2() && !is_partition_valid(cp) && (is_sched_load_balance(parent) != is_sched_load_balance(cp))) { if (is_sched_load_balance(parent)) set_bit(CS_SCHED_LOAD_BALANCE, &cp->flags); @@ -2349,8 +2218,7 @@ get_css: */ if (!cpumask_empty(cp->cpus_allowed) && is_sched_load_balance(cp) && - (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys) || - is_partition_valid(cp))) + (!cpuset_v2() || is_partition_valid(cp))) need_rebuild_sched_domains = true; rcu_read_lock(); @@ -2358,8 +2226,8 @@ get_css: } rcu_read_unlock(); - if (need_rebuild_sched_domains && !(flags & HIER_NO_SD_REBUILD)) - rebuild_sched_domains_locked(); + if (need_rebuild_sched_domains) + cpuset_force_rebuild(); } /** @@ -2380,33 +2248,36 @@ static void update_sibling_cpumasks(struct cpuset *parent, struct cpuset *cs, * Check all its siblings and call update_cpumasks_hier() * if their effective_cpus will need to be changed. * - * With the addition of effective_xcpus which is a subset of - * cpus_allowed. It is possible a change in parent's effective_cpus + * It is possible a change in parent's effective_cpus * due to a change in a child partition's effective_xcpus will impact * its siblings even if they do not inherit parent's effective_cpus * directly. * * The update_cpumasks_hier() function may sleep. So we have to - * release the RCU read lock before calling it. HIER_NO_SD_REBUILD - * flag is used to suppress rebuild of sched domains as the callers - * will take care of that. + * release the RCU read lock before calling it. */ rcu_read_lock(); cpuset_for_each_child(sibling, pos_css, parent) { if (sibling == cs) continue; - if (!sibling->use_parent_ecpus && - !is_partition_valid(sibling)) { + if (!is_partition_valid(sibling)) { compute_effective_cpumask(tmp->new_cpus, sibling, parent); if (cpumask_equal(tmp->new_cpus, sibling->effective_cpus)) continue; + } else if (is_remote_partition(sibling)) { + /* + * Change in a sibling cpuset won't affect a remote + * partition root. + */ + continue; } + if (!css_tryget_online(&sibling->css)) continue; rcu_read_unlock(); - update_cpumasks_hier(sibling, tmp, HIER_NO_SD_REBUILD); + update_cpumasks_hier(sibling, tmp, false); rcu_read_lock(); css_put(&sibling->css); } @@ -2426,10 +2297,10 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, struct tmpmasks tmp; struct cpuset *parent = parent_cs(cs); bool invalidate = false; - int hier_flags = 0; + bool force = false; int old_prs = cs->partition_root_state; - /* top_cpuset.cpus_allowed tracks cpu_online_mask; it's read-only */ + /* top_cpuset.cpus_allowed tracks cpu_active_mask; it's read-only */ if (cs == &top_cpuset) return -EACCES; @@ -2441,7 +2312,8 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, */ if (!*buf) { cpumask_clear(trialcs->cpus_allowed); - cpumask_clear(trialcs->effective_xcpus); + if (cpumask_empty(trialcs->exclusive_cpus)) + cpumask_clear(trialcs->effective_xcpus); } else { retval = cpulist_parse(buf, trialcs->cpus_allowed); if (retval < 0) @@ -2452,13 +2324,14 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, return -EINVAL; /* - * When exclusive_cpus isn't explicitly set, it is constrainted + * When exclusive_cpus isn't explicitly set, it is constrained * by cpus_allowed and parent's effective_xcpus. Otherwise, * trialcs->effective_xcpus is used as a temporary cpumask * for checking validity of the partition root. */ + trialcs->partition_root_state = PRS_MEMBER; if (!cpumask_empty(trialcs->exclusive_cpus) || is_partition_valid(cs)) - compute_effective_exclusive_cpumask(trialcs, NULL); + compute_effective_exclusive_cpumask(trialcs, NULL, cs); } /* Nothing to do if the cpus didn't change */ @@ -2486,12 +2359,11 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, * Check all the descendants in update_cpumasks_hier() if * effective_xcpus is to be changed. */ - if (!cpumask_equal(cs->effective_xcpus, trialcs->effective_xcpus)) - hier_flags = HIER_CHECKALL; + force = !cpumask_equal(cs->effective_xcpus, trialcs->effective_xcpus); retval = validate_change(cs, trialcs); - if ((retval == -EINVAL) && cgroup_subsys_on_dfl(cpuset_cgrp_subsys)) { + if ((retval == -EINVAL) && cpuset_v2()) { struct cgroup_subsys_state *css; struct cpuset *cp; @@ -2505,7 +2377,7 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, invalidate = true; rcu_read_lock(); cpuset_for_each_child(cp, css, parent) { - struct cpumask *xcpus = fetch_xcpus(trialcs); + struct cpumask *xcpus = user_xcpus(trialcs); if (is_partition_valid(cp) && cpumask_intersects(xcpus, cp->effective_xcpus)) { @@ -2532,19 +2404,13 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, * Call remote_cpus_update() to handle valid remote partition */ if (is_remote_partition(cs)) - remote_cpus_update(cs, xcpus, &tmp); + remote_cpus_update(cs, NULL, xcpus, &tmp); else if (invalidate) update_parent_effective_cpumask(cs, partcmd_invalidate, NULL, &tmp); else update_parent_effective_cpumask(cs, partcmd_update, xcpus, &tmp); - } else if (!cpumask_empty(cs->exclusive_cpus)) { - /* - * Use trialcs->effective_cpus as a temp cpumask - */ - remote_partition_check(cs, trialcs->effective_xcpus, - trialcs->effective_cpus, &tmp); } spin_lock_irq(&callback_lock); @@ -2555,7 +2421,7 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, spin_unlock_irq(&callback_lock); /* effective_cpus/effective_xcpus will be updated here */ - update_cpumasks_hier(cs, &tmp, hier_flags); + update_cpumasks_hier(cs, &tmp, force); /* Update CS_SCHED_LOAD_BALANCE and/or sched_domains, if necessary */ if (cs->partition_root_state) @@ -2580,7 +2446,7 @@ static int update_exclusive_cpumask(struct cpuset *cs, struct cpuset *trialcs, struct tmpmasks tmp; struct cpuset *parent = parent_cs(cs); bool invalidate = false; - int hier_flags = 0; + bool force = false; int old_prs = cs->partition_root_state; if (!*buf) { @@ -2590,23 +2456,27 @@ static int update_exclusive_cpumask(struct cpuset *cs, struct cpuset *trialcs, retval = cpulist_parse(buf, trialcs->exclusive_cpus); if (retval < 0) return retval; - if (!is_cpu_exclusive(cs)) - set_bit(CS_CPU_EXCLUSIVE, &trialcs->flags); } /* Nothing to do if the CPUs didn't change */ if (cpumask_equal(cs->exclusive_cpus, trialcs->exclusive_cpus)) return 0; - if (*buf) - compute_effective_exclusive_cpumask(trialcs, NULL); + if (*buf) { + trialcs->partition_root_state = PRS_MEMBER; + /* + * Reject the change if there is exclusive CPUs conflict with + * the siblings. + */ + if (compute_effective_exclusive_cpumask(trialcs, NULL, cs)) + return -EINVAL; + } /* * Check all the descendants in update_cpumasks_hier() if * effective_xcpus is to be changed. */ - if (!cpumask_equal(cs->effective_xcpus, trialcs->effective_xcpus)) - hier_flags = HIER_CHECKALL; + force = !cpumask_equal(cs->effective_xcpus, trialcs->effective_xcpus); retval = validate_change(cs, trialcs); if (retval) @@ -2631,8 +2501,8 @@ static int update_exclusive_cpumask(struct cpuset *cs, struct cpuset *trialcs, if (invalidate) remote_partition_disable(cs, &tmp); else - remote_cpus_update(cs, trialcs->effective_xcpus, - &tmp); + remote_cpus_update(cs, trialcs->exclusive_cpus, + trialcs->effective_xcpus, &tmp); } else if (invalidate) { update_parent_effective_cpumask(cs, partcmd_invalidate, NULL, &tmp); @@ -2640,12 +2510,6 @@ static int update_exclusive_cpumask(struct cpuset *cs, struct cpuset *trialcs, update_parent_effective_cpumask(cs, partcmd_update, trialcs->effective_xcpus, &tmp); } - } else if (!cpumask_empty(trialcs->exclusive_cpus)) { - /* - * Use trialcs->effective_cpus as a temp cpumask - */ - remote_partition_check(cs, trialcs->effective_xcpus, - trialcs->effective_cpus, &tmp); } spin_lock_irq(&callback_lock); cpumask_copy(cs->exclusive_cpus, trialcs->exclusive_cpus); @@ -2659,8 +2523,8 @@ static int update_exclusive_cpumask(struct cpuset *cs, struct cpuset *trialcs, * of the subtree when it is a valid partition root or effective_xcpus * is updated. */ - if (is_partition_valid(cs) || hier_flags) - update_cpumasks_hier(cs, &tmp, hier_flags); + if (is_partition_valid(cs) || force) + update_cpumasks_hier(cs, &tmp, force); /* Update CS_SCHED_LOAD_BALANCE and/or sched_domains, if necessary */ if (cs->partition_root_state) @@ -2754,14 +2618,14 @@ static void cpuset_change_task_nodemask(struct task_struct *tsk, static void *cpuset_being_rebound; /** - * update_tasks_nodemask - Update the nodemasks of tasks in the cpuset. + * cpuset_update_tasks_nodemask - Update the nodemasks of tasks in the cpuset. * @cs: the cpuset in which each task's mems_allowed mask needs to be changed * * Iterate through each task of @cs updating its mems_allowed to the * effective cpuset's. As this function is called with cpuset_mutex held, * cpuset membership stays stable. */ -static void update_tasks_nodemask(struct cpuset *cs) +void cpuset_update_tasks_nodemask(struct cpuset *cs) { static nodemask_t newmems; /* protected by cpuset_mutex */ struct css_task_iter it; @@ -2859,7 +2723,7 @@ static void update_nodemasks_hier(struct cpuset *cs, nodemask_t *new_mems) WARN_ON(!is_in_v2_mode() && !nodes_equal(cp->mems_allowed, cp->effective_mems)); - update_tasks_nodemask(cp); + cpuset_update_tasks_nodemask(cp); rcu_read_lock(); css_put(&cp->css); @@ -2945,44 +2809,8 @@ bool current_cpuset_is_being_rebound(void) return ret; } -static int update_relax_domain_level(struct cpuset *cs, s64 val) -{ -#ifdef CONFIG_SMP - if (val < -1 || val >= sched_domain_level_max) - return -EINVAL; -#endif - - if (val != cs->relax_domain_level) { - cs->relax_domain_level = val; - if (!cpumask_empty(cs->cpus_allowed) && - is_sched_load_balance(cs)) - rebuild_sched_domains_locked(); - } - - return 0; -} - -/** - * update_tasks_flags - update the spread flags of tasks in the cpuset. - * @cs: the cpuset in which each task's spread flags needs to be changed - * - * Iterate through each task of @cs updating its spread flags. As this - * function is called with cpuset_mutex held, cpuset membership stays - * stable. - */ -static void update_tasks_flags(struct cpuset *cs) -{ - struct css_task_iter it; - struct task_struct *task; - - css_task_iter_start(&cs->css, 0, &it); - while ((task = css_task_iter_next(&it))) - cpuset_update_task_spread_flags(cs, task); - css_task_iter_end(&it); -} - /* - * update_flag - read a 0 or a 1 in a file and update associated flag + * cpuset_update_flag - read a 0 or a 1 in a file and update associated flag * bit: the bit to update (see cpuset_flagbits_t) * cs: the cpuset to update * turning_on: whether the flag is being set or cleared @@ -2990,7 +2818,7 @@ static void update_tasks_flags(struct cpuset *cs) * Call with cpuset_mutex held. */ -static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, +int cpuset_update_flag(cpuset_flagbits_t bit, struct cpuset *cs, int turning_on) { struct cpuset *trialcs; @@ -3021,11 +2849,15 @@ static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, cs->flags = trialcs->flags; spin_unlock_irq(&callback_lock); - if (!cpumask_empty(trialcs->cpus_allowed) && balance_flag_changed) - rebuild_sched_domains_locked(); + if (!cpumask_empty(trialcs->cpus_allowed) && balance_flag_changed) { + if (cpuset_v2()) + cpuset_force_rebuild(); + else + rebuild_sched_domains_locked(); + } if (spread_flag_changed) - update_tasks_flags(cs); + cpuset1_update_tasks_flags(cs); out: free_cpuset(trialcs); return err; @@ -3044,7 +2876,7 @@ static int update_prstate(struct cpuset *cs, int new_prs) int err = PERR_NONE, old_prs = cs->partition_root_state; struct cpuset *parent = parent_cs(cs); struct tmpmasks tmpmask; - bool new_xcpus_state = false; + bool isolcpus_updated = false; if (old_prs == new_prs) return 0; @@ -3058,45 +2890,50 @@ static int update_prstate(struct cpuset *cs, int new_prs) if (alloc_cpumasks(NULL, &tmpmask)) return -ENOMEM; - /* - * Setup effective_xcpus if not properly set yet, it will be cleared - * later if partition becomes invalid. - */ - if ((new_prs > 0) && cpumask_empty(cs->exclusive_cpus)) { - spin_lock_irq(&callback_lock); - cpumask_and(cs->effective_xcpus, - cs->cpus_allowed, parent->effective_xcpus); - spin_unlock_irq(&callback_lock); - } - - err = update_partition_exclusive(cs, new_prs); + err = update_partition_exclusive_flag(cs, new_prs); if (err) goto out; if (!old_prs) { - enum partition_cmd cmd = (new_prs == PRS_ROOT) - ? partcmd_enable : partcmd_enablei; - /* - * cpus_allowed cannot be empty. + * cpus_allowed and exclusive_cpus cannot be both empty. */ - if (cpumask_empty(cs->cpus_allowed)) { + if (xcpus_empty(cs)) { err = PERR_CPUSEMPTY; goto out; } - err = update_parent_effective_cpumask(cs, cmd, NULL, &tmpmask); /* - * If an attempt to become local partition root fails, - * try to become a remote partition root instead. + * We don't support the creation of a new local partition with + * a remote partition underneath it. This unsupported + * setting can happen only if parent is the top_cpuset because + * a remote partition cannot be created underneath an existing + * local or remote partition. */ - if (err && remote_partition_enable(cs, new_prs, &tmpmask)) - err = 0; + if ((parent == &top_cpuset) && + cpumask_intersects(cs->exclusive_cpus, subpartitions_cpus)) { + err = PERR_REMOTE; + goto out; + } + + /* + * If parent is valid partition, enable local partiion. + * Otherwise, enable a remote partition. + */ + if (is_partition_valid(parent)) { + enum partition_cmd cmd = (new_prs == PRS_ROOT) + ? partcmd_enable : partcmd_enablei; + + err = update_parent_effective_cpumask(cs, cmd, NULL, &tmpmask); + } else { + err = remote_partition_enable(cs, new_prs, &tmpmask); + } } else if (old_prs && new_prs) { /* * A change in load balance state only, no change in cpumasks. + * Need to update isolated_cpus. */ - new_xcpus_state = true; + isolcpus_updated = true; } else { /* * Switching back to member is always allowed even if it @@ -3120,7 +2957,7 @@ out: */ if (err) { new_prs = -new_prs; - update_partition_exclusive(cs, new_prs); + update_partition_exclusive_flag(cs, new_prs); } spin_lock_irq(&callback_lock); @@ -3128,123 +2965,28 @@ out: WRITE_ONCE(cs->prs_err, err); if (!is_partition_valid(cs)) reset_partition_data(cs); - else if (new_xcpus_state) - partition_xcpus_newstate(old_prs, new_prs, cs->effective_xcpus); + else if (isolcpus_updated) + isolated_cpus_update(old_prs, new_prs, cs->effective_xcpus); spin_unlock_irq(&callback_lock); - update_unbound_workqueue_cpumask(new_xcpus_state); + update_unbound_workqueue_cpumask(isolcpus_updated); - /* Force update if switching back to member */ - update_cpumasks_hier(cs, &tmpmask, !new_prs ? HIER_CHECKALL : 0); + /* Force update if switching back to member & update effective_xcpus */ + update_cpumasks_hier(cs, &tmpmask, !new_prs); + + /* A newly created partition must have effective_xcpus set */ + WARN_ON_ONCE(!old_prs && (new_prs > 0) + && cpumask_empty(cs->effective_xcpus)); /* Update sched domains and load balance flag */ update_partition_sd_lb(cs, old_prs); notify_partition_change(cs, old_prs); + if (force_sd_rebuild) + rebuild_sched_domains_locked(); free_cpumasks(NULL, &tmpmask); return 0; } -/* - * Frequency meter - How fast is some event occurring? - * - * These routines manage a digitally filtered, constant time based, - * event frequency meter. There are four routines: - * fmeter_init() - initialize a frequency meter. - * fmeter_markevent() - called each time the event happens. - * fmeter_getrate() - returns the recent rate of such events. - * fmeter_update() - internal routine used to update fmeter. - * - * A common data structure is passed to each of these routines, - * which is used to keep track of the state required to manage the - * frequency meter and its digital filter. - * - * The filter works on the number of events marked per unit time. - * The filter is single-pole low-pass recursive (IIR). The time unit - * is 1 second. Arithmetic is done using 32-bit integers scaled to - * simulate 3 decimal digits of precision (multiplied by 1000). - * - * With an FM_COEF of 933, and a time base of 1 second, the filter - * has a half-life of 10 seconds, meaning that if the events quit - * happening, then the rate returned from the fmeter_getrate() - * will be cut in half each 10 seconds, until it converges to zero. - * - * It is not worth doing a real infinitely recursive filter. If more - * than FM_MAXTICKS ticks have elapsed since the last filter event, - * just compute FM_MAXTICKS ticks worth, by which point the level - * will be stable. - * - * Limit the count of unprocessed events to FM_MAXCNT, so as to avoid - * arithmetic overflow in the fmeter_update() routine. - * - * Given the simple 32 bit integer arithmetic used, this meter works - * best for reporting rates between one per millisecond (msec) and - * one per 32 (approx) seconds. At constant rates faster than one - * per msec it maxes out at values just under 1,000,000. At constant - * rates between one per msec, and one per second it will stabilize - * to a value N*1000, where N is the rate of events per second. - * At constant rates between one per second and one per 32 seconds, - * it will be choppy, moving up on the seconds that have an event, - * and then decaying until the next event. At rates slower than - * about one in 32 seconds, it decays all the way back to zero between - * each event. - */ - -#define FM_COEF 933 /* coefficient for half-life of 10 secs */ -#define FM_MAXTICKS ((u32)99) /* useless computing more ticks than this */ -#define FM_MAXCNT 1000000 /* limit cnt to avoid overflow */ -#define FM_SCALE 1000 /* faux fixed point scale */ - -/* Initialize a frequency meter */ -static void fmeter_init(struct fmeter *fmp) -{ - fmp->cnt = 0; - fmp->val = 0; - fmp->time = 0; - spin_lock_init(&fmp->lock); -} - -/* Internal meter update - process cnt events and update value */ -static void fmeter_update(struct fmeter *fmp) -{ - time64_t now; - u32 ticks; - - now = ktime_get_seconds(); - ticks = now - fmp->time; - - if (ticks == 0) - return; - - ticks = min(FM_MAXTICKS, ticks); - while (ticks-- > 0) - fmp->val = (FM_COEF * fmp->val) / FM_SCALE; - fmp->time = now; - - fmp->val += ((FM_SCALE - FM_COEF) * fmp->cnt) / FM_SCALE; - fmp->cnt = 0; -} - -/* Process any previous ticks, then bump cnt by one (times scale). */ -static void fmeter_markevent(struct fmeter *fmp) -{ - spin_lock(&fmp->lock); - fmeter_update(fmp); - fmp->cnt = min(FM_MAXCNT, fmp->cnt + FM_SCALE); - spin_unlock(&fmp->lock); -} - -/* Process any previous ticks, then return current value. */ -static int fmeter_getrate(struct fmeter *fmp) -{ - int val; - - spin_lock(&fmp->lock); - fmeter_update(fmp); - val = fmp->val; - spin_unlock(&fmp->lock); - return val; -} - static struct cpuset *cpuset_attach_old_cs; /* @@ -3301,8 +3043,7 @@ static int cpuset_can_attach(struct cgroup_taskset *tset) * migration permission derives from hierarchy ownership in * cgroup_procs_write_permission()). */ - if (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys) || - (cpus_updated || mems_updated)) { + if (!cpuset_v2() || (cpus_updated || mems_updated)) { ret = security_task_setscheduler(task); if (ret) goto out_unlock; @@ -3353,9 +3094,7 @@ static void cpuset_cancel_attach(struct cgroup_taskset *tset) cs = css_cs(css); mutex_lock(&cpuset_mutex); - cs->attach_in_progress--; - if (!cs->attach_in_progress) - wake_up(&cpuset_attach_wq); + dec_attach_in_progress_locked(cs); if (cs->nr_migrate_dl_tasks) { int cpu = cpumask_any(cs->effective_cpus); @@ -3380,7 +3119,7 @@ static void cpuset_attach_task(struct cpuset *cs, struct task_struct *task) lockdep_assert_held(&cpuset_mutex); if (cs != &top_cpuset) - guarantee_online_cpus(task, cpus_attach); + guarantee_active_cpus(task, cpus_attach); else cpumask_andnot(cpus_attach, task_cpu_possible_mask(task), subpartitions_cpus); @@ -3391,7 +3130,7 @@ static void cpuset_attach_task(struct cpuset *cs, struct task_struct *task) WARN_ON_ONCE(set_cpus_allowed_ptr(task, cpus_attach)); cpuset_change_task_nodemask(task, &cpuset_attach_nodemask_to); - cpuset_update_task_spread_flags(cs, task); + cpuset1_update_task_spread_flags(cs, task); } static void cpuset_attach(struct cgroup_taskset *tset) @@ -3418,8 +3157,7 @@ static void cpuset_attach(struct cgroup_taskset *tset) * in effective cpus and mems. In that case, we can optimize out * by skipping the task iteration and update. */ - if (cgroup_subsys_on_dfl(cpuset_cgrp_subsys) && - !cpus_updated && !mems_updated) { + if (cpuset_v2() && !cpus_updated && !mems_updated) { cpuset_attach_nodemask_to = cs->effective_mems; goto out; } @@ -3470,116 +3208,15 @@ out: reset_migrate_dl_data(cs); } - cs->attach_in_progress--; - if (!cs->attach_in_progress) - wake_up(&cpuset_attach_wq); + dec_attach_in_progress_locked(cs); mutex_unlock(&cpuset_mutex); } -/* The various types of files and directories in a cpuset file system */ - -typedef enum { - FILE_MEMORY_MIGRATE, - FILE_CPULIST, - FILE_MEMLIST, - FILE_EFFECTIVE_CPULIST, - FILE_EFFECTIVE_MEMLIST, - FILE_SUBPARTS_CPULIST, - FILE_EXCLUSIVE_CPULIST, - FILE_EFFECTIVE_XCPULIST, - FILE_ISOLATED_CPULIST, - FILE_CPU_EXCLUSIVE, - FILE_MEM_EXCLUSIVE, - FILE_MEM_HARDWALL, - FILE_SCHED_LOAD_BALANCE, - FILE_PARTITION_ROOT, - FILE_SCHED_RELAX_DOMAIN_LEVEL, - FILE_MEMORY_PRESSURE_ENABLED, - FILE_MEMORY_PRESSURE, - FILE_SPREAD_PAGE, - FILE_SPREAD_SLAB, -} cpuset_filetype_t; - -static int cpuset_write_u64(struct cgroup_subsys_state *css, struct cftype *cft, - u64 val) -{ - struct cpuset *cs = css_cs(css); - cpuset_filetype_t type = cft->private; - int retval = 0; - - cpus_read_lock(); - mutex_lock(&cpuset_mutex); - if (!is_cpuset_online(cs)) { - retval = -ENODEV; - goto out_unlock; - } - - switch (type) { - case FILE_CPU_EXCLUSIVE: - retval = update_flag(CS_CPU_EXCLUSIVE, cs, val); - break; - case FILE_MEM_EXCLUSIVE: - retval = update_flag(CS_MEM_EXCLUSIVE, cs, val); - break; - case FILE_MEM_HARDWALL: - retval = update_flag(CS_MEM_HARDWALL, cs, val); - break; - case FILE_SCHED_LOAD_BALANCE: - retval = update_flag(CS_SCHED_LOAD_BALANCE, cs, val); - break; - case FILE_MEMORY_MIGRATE: - retval = update_flag(CS_MEMORY_MIGRATE, cs, val); - break; - case FILE_MEMORY_PRESSURE_ENABLED: - cpuset_memory_pressure_enabled = !!val; - break; - case FILE_SPREAD_PAGE: - retval = update_flag(CS_SPREAD_PAGE, cs, val); - break; - case FILE_SPREAD_SLAB: - retval = update_flag(CS_SPREAD_SLAB, cs, val); - break; - default: - retval = -EINVAL; - break; - } -out_unlock: - mutex_unlock(&cpuset_mutex); - cpus_read_unlock(); - return retval; -} - -static int cpuset_write_s64(struct cgroup_subsys_state *css, struct cftype *cft, - s64 val) -{ - struct cpuset *cs = css_cs(css); - cpuset_filetype_t type = cft->private; - int retval = -ENODEV; - - cpus_read_lock(); - mutex_lock(&cpuset_mutex); - if (!is_cpuset_online(cs)) - goto out_unlock; - - switch (type) { - case FILE_SCHED_RELAX_DOMAIN_LEVEL: - retval = update_relax_domain_level(cs, val); - break; - default: - retval = -EINVAL; - break; - } -out_unlock: - mutex_unlock(&cpuset_mutex); - cpus_read_unlock(); - return retval; -} - /* * Common handling for a write to a "cpus" or "mems" file. */ -static ssize_t cpuset_write_resmask(struct kernfs_open_file *of, +ssize_t cpuset_write_resmask(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) { struct cpuset *cs = css_cs(of_css(of)); @@ -3587,30 +3224,6 @@ static ssize_t cpuset_write_resmask(struct kernfs_open_file *of, int retval = -ENODEV; buf = strstrip(buf); - - /* - * CPU or memory hotunplug may leave @cs w/o any execution - * resources, in which case the hotplug code asynchronously updates - * configuration and transfers all tasks to the nearest ancestor - * which can execute. - * - * As writes to "cpus" or "mems" may restore @cs's execution - * resources, wait for the previously scheduled operations before - * proceeding, so that we don't end up keep removing tasks added - * after execution capability is restored. - * - * cpuset_hotplug_work calls back into cgroup core via - * cgroup_transfer_tasks() and waiting for it from a cgroupfs - * operation like this one can lead to a deadlock through kernfs - * active_ref protection. Let's break the protection. Losing the - * protection is okay as we check whether @cs is online after - * grabbing cpuset_mutex anyway. This only happens on the legacy - * hierarchies. - */ - css_get(&cs->css); - kernfs_break_active_protection(of->kn); - flush_work(&cpuset_hotplug_work); - cpus_read_lock(); mutex_lock(&cpuset_mutex); if (!is_cpuset_online(cs)) @@ -3638,11 +3251,11 @@ static ssize_t cpuset_write_resmask(struct kernfs_open_file *of, } free_cpuset(trialcs); + if (force_sd_rebuild) + rebuild_sched_domains_locked(); out_unlock: mutex_unlock(&cpuset_mutex); cpus_read_unlock(); - kernfs_unbreak_active_protection(of->kn); - css_put(&cs->css); flush_workqueue(cpuset_migrate_mm_wq); return retval ?: nbytes; } @@ -3655,7 +3268,7 @@ out_unlock: * and since these maps can change value dynamically, one could read * gibberish by doing partial reads while a list was changing. */ -static int cpuset_common_seq_show(struct seq_file *sf, void *v) +int cpuset_common_seq_show(struct seq_file *sf, void *v) { struct cpuset *cs = css_cs(seq_css(sf)); cpuset_filetype_t type = seq_cft(sf)->private; @@ -3696,53 +3309,7 @@ static int cpuset_common_seq_show(struct seq_file *sf, void *v) return ret; } -static u64 cpuset_read_u64(struct cgroup_subsys_state *css, struct cftype *cft) -{ - struct cpuset *cs = css_cs(css); - cpuset_filetype_t type = cft->private; - switch (type) { - case FILE_CPU_EXCLUSIVE: - return is_cpu_exclusive(cs); - case FILE_MEM_EXCLUSIVE: - return is_mem_exclusive(cs); - case FILE_MEM_HARDWALL: - return is_mem_hardwall(cs); - case FILE_SCHED_LOAD_BALANCE: - return is_sched_load_balance(cs); - case FILE_MEMORY_MIGRATE: - return is_memory_migrate(cs); - case FILE_MEMORY_PRESSURE_ENABLED: - return cpuset_memory_pressure_enabled; - case FILE_MEMORY_PRESSURE: - return fmeter_getrate(&cs->fmeter); - case FILE_SPREAD_PAGE: - return is_spread_page(cs); - case FILE_SPREAD_SLAB: - return is_spread_slab(cs); - default: - BUG(); - } - - /* Unreachable but makes gcc happy */ - return 0; -} - -static s64 cpuset_read_s64(struct cgroup_subsys_state *css, struct cftype *cft) -{ - struct cpuset *cs = css_cs(css); - cpuset_filetype_t type = cft->private; - switch (type) { - case FILE_SCHED_RELAX_DOMAIN_LEVEL: - return cs->relax_domain_level; - default: - BUG(); - } - - /* Unreachable but makes gcc happy */ - return 0; -} - -static int sched_partition_show(struct seq_file *seq, void *v) +static int cpuset_partition_show(struct seq_file *seq, void *v) { struct cpuset *cs = css_cs(seq_css(seq)); const char *err, *type = NULL; @@ -3773,7 +3340,7 @@ static int sched_partition_show(struct seq_file *seq, void *v) return 0; } -static ssize_t sched_partition_write(struct kernfs_open_file *of, char *buf, +static ssize_t cpuset_partition_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) { struct cpuset *cs = css_cs(of_css(of)); @@ -3782,9 +3349,6 @@ static ssize_t sched_partition_write(struct kernfs_open_file *of, char *buf, buf = strstrip(buf); - /* - * Convert "root" to ENABLED, and convert "member" to DISABLED. - */ if (!strcmp(buf, "root")) val = PRS_ROOT; else if (!strcmp(buf, "member")) @@ -3797,11 +3361,8 @@ static ssize_t sched_partition_write(struct kernfs_open_file *of, char *buf, css_get(&cs->css); cpus_read_lock(); mutex_lock(&cpuset_mutex); - if (!is_cpuset_online(cs)) - goto out_unlock; - - retval = update_prstate(cs, val); -out_unlock: + if (is_cpuset_online(cs)) + retval = update_prstate(cs, val); mutex_unlock(&cpuset_mutex); cpus_read_unlock(); css_put(&cs->css); @@ -3809,113 +3370,6 @@ out_unlock: } /* - * for the common functions, 'private' gives the type of file - */ - -static struct cftype legacy_files[] = { - { - .name = "cpus", - .seq_show = cpuset_common_seq_show, - .write = cpuset_write_resmask, - .max_write_len = (100U + 6 * NR_CPUS), - .private = FILE_CPULIST, - }, - - { - .name = "mems", - .seq_show = cpuset_common_seq_show, - .write = cpuset_write_resmask, - .max_write_len = (100U + 6 * MAX_NUMNODES), - .private = FILE_MEMLIST, - }, - - { - .name = "effective_cpus", - .seq_show = cpuset_common_seq_show, - .private = FILE_EFFECTIVE_CPULIST, - }, - - { - .name = "effective_mems", - .seq_show = cpuset_common_seq_show, - .private = FILE_EFFECTIVE_MEMLIST, - }, - - { - .name = "cpu_exclusive", - .read_u64 = cpuset_read_u64, - .write_u64 = cpuset_write_u64, - .private = FILE_CPU_EXCLUSIVE, - }, - - { - .name = "mem_exclusive", - .read_u64 = cpuset_read_u64, - .write_u64 = cpuset_write_u64, - .private = FILE_MEM_EXCLUSIVE, - }, - - { - .name = "mem_hardwall", - .read_u64 = cpuset_read_u64, - .write_u64 = cpuset_write_u64, - .private = FILE_MEM_HARDWALL, - }, - - { - .name = "sched_load_balance", - .read_u64 = cpuset_read_u64, - .write_u64 = cpuset_write_u64, - .private = FILE_SCHED_LOAD_BALANCE, - }, - - { - .name = "sched_relax_domain_level", - .read_s64 = cpuset_read_s64, - .write_s64 = cpuset_write_s64, - .private = FILE_SCHED_RELAX_DOMAIN_LEVEL, - }, - - { - .name = "memory_migrate", - .read_u64 = cpuset_read_u64, - .write_u64 = cpuset_write_u64, - .private = FILE_MEMORY_MIGRATE, - }, - - { - .name = "memory_pressure", - .read_u64 = cpuset_read_u64, - .private = FILE_MEMORY_PRESSURE, - }, - - { - .name = "memory_spread_page", - .read_u64 = cpuset_read_u64, - .write_u64 = cpuset_write_u64, - .private = FILE_SPREAD_PAGE, - }, - - { - /* obsolete, may be removed in the future */ - .name = "memory_spread_slab", - .read_u64 = cpuset_read_u64, - .write_u64 = cpuset_write_u64, - .private = FILE_SPREAD_SLAB, - }, - - { - .name = "memory_pressure_enabled", - .flags = CFTYPE_ONLY_ON_ROOT, - .read_u64 = cpuset_read_u64, - .write_u64 = cpuset_write_u64, - .private = FILE_MEMORY_PRESSURE_ENABLED, - }, - - { } /* terminate */ -}; - -/* * This is currently a minimal set for the default hierarchy. It can be * expanded later on by migrating more features and control files from v1. */ @@ -3952,8 +3406,8 @@ static struct cftype dfl_files[] = { { .name = "cpus.partition", - .seq_show = sched_partition_show, - .write = sched_partition_write, + .seq_show = cpuset_partition_show, + .write = cpuset_partition_write, .private = FILE_PARTITION_ROOT, .flags = CFTYPE_NOT_ON_ROOT, .file_offset = offsetof(struct cpuset, partition_file), @@ -4020,14 +3474,12 @@ cpuset_css_alloc(struct cgroup_subsys_state *parent_css) } __set_bit(CS_SCHED_LOAD_BALANCE, &cs->flags); - nodes_clear(cs->mems_allowed); - nodes_clear(cs->effective_mems); fmeter_init(&cs->fmeter); cs->relax_domain_level = -1; INIT_LIST_HEAD(&cs->remote_sibling); /* Set CS_MEMORY_MIGRATE for default hierarchy */ - if (cgroup_subsys_on_dfl(cpuset_cgrp_subsys)) + if (cpuset_v2()) __set_bit(CS_MEMORY_MIGRATE, &cs->flags); return &cs->css; @@ -4051,6 +3503,11 @@ static int cpuset_css_online(struct cgroup_subsys_state *css) set_bit(CS_SPREAD_PAGE, &cs->flags); if (is_spread_slab(parent)) set_bit(CS_SPREAD_SLAB, &cs->flags); + /* + * For v2, clear CS_SCHED_LOAD_BALANCE if parent is isolated + */ + if (cpuset_v2() && !is_sched_load_balance(parent)) + clear_bit(CS_SCHED_LOAD_BALANCE, &cs->flags); cpuset_inc(); @@ -4058,22 +3515,7 @@ static int cpuset_css_online(struct cgroup_subsys_state *css) if (is_in_v2_mode()) { cpumask_copy(cs->effective_cpus, parent->effective_cpus); cs->effective_mems = parent->effective_mems; - cs->use_parent_ecpus = true; - parent->child_ecpus_count++; - /* - * Clear CS_SCHED_LOAD_BALANCE if parent is isolated - */ - if (!is_sched_load_balance(parent)) - clear_bit(CS_SCHED_LOAD_BALANCE, &cs->flags); } - - /* - * For v2, clear CS_SCHED_LOAD_BALANCE if parent is isolated - */ - if (cgroup_subsys_on_dfl(cpuset_cgrp_subsys) && - !is_sched_load_balance(parent)) - clear_bit(CS_SCHED_LOAD_BALANCE, &cs->flags); - spin_unlock_irq(&callback_lock); if (!test_bit(CGRP_CPUSET_CLONE_CHILDREN, &css->cgroup->flags)) @@ -4119,11 +3561,7 @@ out_unlock: * will call rebuild_sched_domains_locked(). That is not needed * in the default hierarchy where only changes in partition * will cause repartitioning. - * - * If the cpuset has the 'sched.partition' flag enabled, simulate - * turning 'sched.partition" off. */ - static void cpuset_css_offline(struct cgroup_subsys_state *css) { struct cpuset *cs = css_cs(css); @@ -4131,25 +3569,35 @@ static void cpuset_css_offline(struct cgroup_subsys_state *css) cpus_read_lock(); mutex_lock(&cpuset_mutex); - if (is_partition_valid(cs)) - update_prstate(cs, 0); + if (!cpuset_v2() && is_sched_load_balance(cs)) + cpuset_update_flag(CS_SCHED_LOAD_BALANCE, cs, 0); - if (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys) && - is_sched_load_balance(cs)) - update_flag(CS_SCHED_LOAD_BALANCE, cs, 0); + cpuset_dec(); + clear_bit(CS_ONLINE, &cs->flags); - if (cs->use_parent_ecpus) { - struct cpuset *parent = parent_cs(cs); + mutex_unlock(&cpuset_mutex); + cpus_read_unlock(); +} - cs->use_parent_ecpus = false; - parent->child_ecpus_count--; - } +/* + * If a dying cpuset has the 'cpus.partition' enabled, turn it off by + * changing it back to member to free its exclusive CPUs back to the pool to + * be used by other online cpusets. + */ +static void cpuset_css_killed(struct cgroup_subsys_state *css) +{ + struct cpuset *cs = css_cs(css); - cpuset_dec(); - clear_bit(CS_ONLINE, &cs->flags); + cpus_read_lock(); + mutex_lock(&cpuset_mutex); + + /* Reset valid partition back to member */ + if (is_partition_valid(cs)) + update_prstate(cs, PRS_MEMBER); mutex_unlock(&cpuset_mutex); cpus_read_unlock(); + } static void cpuset_css_free(struct cgroup_subsys_state *css) @@ -4233,11 +3681,7 @@ static void cpuset_cancel_fork(struct task_struct *task, struct css_set *cset) if (same_cs) return; - mutex_lock(&cpuset_mutex); - cs->attach_in_progress--; - if (!cs->attach_in_progress) - wake_up(&cpuset_attach_wq); - mutex_unlock(&cpuset_mutex); + dec_attach_in_progress(cs); } /* @@ -4269,10 +3713,7 @@ static void cpuset_fork(struct task_struct *task) guarantee_online_mems(cs, &cpuset_attach_nodemask_to); cpuset_attach_task(cs, task); - cs->attach_in_progress--; - if (!cs->attach_in_progress) - wake_up(&cpuset_attach_wq); - + dec_attach_in_progress_locked(cs); mutex_unlock(&cpuset_mutex); } @@ -4280,6 +3721,7 @@ struct cgroup_subsys cpuset_cgrp_subsys = { .css_alloc = cpuset_css_alloc, .css_online = cpuset_css_online, .css_offline = cpuset_css_offline, + .css_killed = cpuset_css_killed, .css_free = cpuset_css_free, .can_attach = cpuset_can_attach, .cancel_attach = cpuset_cancel_attach, @@ -4289,7 +3731,9 @@ struct cgroup_subsys cpuset_cgrp_subsys = { .can_fork = cpuset_can_fork, .cancel_fork = cpuset_cancel_fork, .fork = cpuset_fork, - .legacy_cftypes = legacy_files, +#ifdef CONFIG_CPUSETS_V1 + .legacy_cftypes = cpuset1_files, +#endif .dfl_cftypes = dfl_files, .early_init = true, .threaded = true, @@ -4318,78 +3762,18 @@ int __init cpuset_init(void) nodes_setall(top_cpuset.effective_mems); fmeter_init(&top_cpuset.fmeter); - set_bit(CS_SCHED_LOAD_BALANCE, &top_cpuset.flags); - top_cpuset.relax_domain_level = -1; INIT_LIST_HEAD(&remote_children); BUG_ON(!alloc_cpumask_var(&cpus_attach, GFP_KERNEL)); - return 0; -} - -/* - * If CPU and/or memory hotplug handlers, below, unplug any CPUs - * or memory nodes, we need to walk over the cpuset hierarchy, - * removing that CPU or node from all cpusets. If this removes the - * last CPU or node from a cpuset, then move the tasks in the empty - * cpuset to its next-highest non-empty parent. - */ -static void remove_tasks_in_empty_cpuset(struct cpuset *cs) -{ - struct cpuset *parent; - - /* - * Find its next-highest non-empty parent, (top cpuset - * has online cpus, so can't be empty). - */ - parent = parent_cs(cs); - while (cpumask_empty(parent->cpus_allowed) || - nodes_empty(parent->mems_allowed)) - parent = parent_cs(parent); - - if (cgroup_transfer_tasks(parent->css.cgroup, cs->css.cgroup)) { - pr_err("cpuset: failed to transfer tasks out of empty cpuset "); - pr_cont_cgroup_name(cs->css.cgroup); - pr_cont("\n"); + have_boot_isolcpus = housekeeping_enabled(HK_TYPE_DOMAIN); + if (have_boot_isolcpus) { + BUG_ON(!alloc_cpumask_var(&boot_hk_cpus, GFP_KERNEL)); + cpumask_copy(boot_hk_cpus, housekeeping_cpumask(HK_TYPE_DOMAIN)); + cpumask_andnot(isolated_cpus, cpu_possible_mask, boot_hk_cpus); } -} - -static void -hotplug_update_tasks_legacy(struct cpuset *cs, - struct cpumask *new_cpus, nodemask_t *new_mems, - bool cpus_updated, bool mems_updated) -{ - bool is_empty; - spin_lock_irq(&callback_lock); - cpumask_copy(cs->cpus_allowed, new_cpus); - cpumask_copy(cs->effective_cpus, new_cpus); - cs->mems_allowed = *new_mems; - cs->effective_mems = *new_mems; - spin_unlock_irq(&callback_lock); - - /* - * Don't call update_tasks_cpumask() if the cpuset becomes empty, - * as the tasks will be migrated to an ancestor. - */ - if (cpus_updated && !cpumask_empty(cs->cpus_allowed)) - update_tasks_cpumask(cs, new_cpus); - if (mems_updated && !nodes_empty(cs->mems_allowed)) - update_tasks_nodemask(cs); - - is_empty = cpumask_empty(cs->cpus_allowed) || - nodes_empty(cs->mems_allowed); - - /* - * Move tasks to the nearest ancestor with execution resources, - * This is full cgroup operation which will also call back into - * cpuset. Should be done outside any lock. - */ - if (is_empty) { - mutex_unlock(&cpuset_mutex); - remove_tasks_in_empty_cpuset(cs); - mutex_lock(&cpuset_mutex); - } + return 0; } static void @@ -4409,40 +3793,14 @@ hotplug_update_tasks(struct cpuset *cs, spin_unlock_irq(&callback_lock); if (cpus_updated) - update_tasks_cpumask(cs, new_cpus); + cpuset_update_tasks_cpumask(cs, new_cpus); if (mems_updated) - update_tasks_nodemask(cs); + cpuset_update_tasks_nodemask(cs); } -static bool force_rebuild; - void cpuset_force_rebuild(void) { - force_rebuild = true; -} - -/* - * Attempt to acquire a cpus_read_lock while a hotplug operation may be in - * progress. - * Return: true if successful, false otherwise - * - * To avoid circular lock dependency between cpuset_mutex and cpus_read_lock, - * cpus_read_trylock() is used here to acquire the lock. - */ -static bool cpuset_hotplug_cpus_read_trylock(void) -{ - int retries = 0; - - while (!cpus_read_trylock()) { - /* - * CPU hotplug still in progress. Retry 5 times - * with a 10ms wait before bailing out. - */ - if (++retries > 5) - return false; - msleep(10); - } - return true; + force_sd_rebuild = true; } /** @@ -4493,13 +3851,11 @@ retry: compute_partition_effective_cpumask(cs, &new_cpus); if (remote && cpumask_empty(&new_cpus) && - partition_is_populated(cs, NULL) && - cpuset_hotplug_cpus_read_trylock()) { + partition_is_populated(cs, NULL)) { + cs->prs_err = PERR_HOTPLUG; remote_partition_disable(cs, tmp); compute_effective_cpumask(&new_cpus, cs, parent); remote = false; - cpuset_force_rebuild(); - cpus_read_unlock(); } /* @@ -4519,18 +3875,8 @@ retry: else if (is_partition_valid(parent) && is_partition_invalid(cs)) partcmd = partcmd_update; - /* - * cpus_read_lock needs to be held before calling - * update_parent_effective_cpumask(). To avoid circular lock - * dependency between cpuset_mutex and cpus_read_lock, - * cpus_read_trylock() is used here to acquire the lock. - */ if (partcmd >= 0) { - if (!cpuset_hotplug_cpus_read_trylock()) - goto update_tasks; - update_parent_effective_cpumask(cs, partcmd, NULL, tmp); - cpus_read_unlock(); if ((partcmd == partcmd_invalidate) || is_partition_valid(cs)) { compute_partition_effective_cpumask(cs, &new_cpus); cpuset_force_rebuild(); @@ -4550,7 +3896,7 @@ update_tasks: hotplug_update_tasks(cs, &new_cpus, &new_mems, cpus_updated, mems_updated); else - hotplug_update_tasks_legacy(cs, &new_cpus, &new_mems, + cpuset1_hotplug_update_tasks(cs, &new_cpus, &new_mems, cpus_updated, mems_updated); unlock: @@ -4558,8 +3904,7 @@ unlock: } /** - * cpuset_hotplug_workfn - handle CPU/memory hotunplug for a cpuset - * @work: unused + * cpuset_handle_hotplug - handle CPU/memory hot{,un}plug for a cpuset * * This function is called after either CPU or memory configuration has * changed and updates cpuset accordingly. The top_cpuset is always @@ -4573,8 +3918,10 @@ unlock: * * Note that CPU offlining during suspend is ignored. We don't modify * cpusets across suspend/resume cycles at all. + * + * CPU / memory hotplug is handled synchronously. */ -static void cpuset_hotplug_workfn(struct work_struct *work) +static void cpuset_handle_hotplug(void) { static cpumask_t new_cpus; static nodemask_t new_mems; @@ -4585,6 +3932,7 @@ static void cpuset_hotplug_workfn(struct work_struct *work) if (on_dfl && !alloc_cpumasks(NULL, &tmp)) ptmp = &tmp; + lockdep_assert_cpus_held(); mutex_lock(&cpuset_mutex); /* fetch the available cpus/mems and find out which changed how */ @@ -4600,15 +3948,9 @@ static void cpuset_hotplug_workfn(struct work_struct *work) !cpumask_empty(subpartitions_cpus); mems_updated = !nodes_equal(top_cpuset.effective_mems, new_mems); - /* - * In the rare case that hotplug removes all the cpus in - * subpartitions_cpus, we assumed that cpus are updated. - */ - if (!cpus_updated && top_cpuset.nr_subparts) - cpus_updated = true; - /* For v1, synchronize cpus_allowed to cpu_active_mask */ if (cpus_updated) { + cpuset_force_rebuild(); spin_lock_irq(&callback_lock); if (!on_dfl) cpumask_copy(top_cpuset.cpus_allowed, &new_cpus); @@ -4639,7 +3981,7 @@ static void cpuset_hotplug_workfn(struct work_struct *work) top_cpuset.mems_allowed = new_mems; top_cpuset.effective_mems = new_mems; spin_unlock_irq(&callback_lock); - update_tasks_nodemask(&top_cpuset); + cpuset_update_tasks_nodemask(&top_cpuset); } mutex_unlock(&cpuset_mutex); @@ -4663,11 +4005,9 @@ static void cpuset_hotplug_workfn(struct work_struct *work) rcu_read_unlock(); } - /* rebuild sched domains if cpus_allowed has changed */ - if (cpus_updated || force_rebuild) { - force_rebuild = false; - rebuild_sched_domains(); - } + /* rebuild sched domains if necessary */ + if (force_sd_rebuild) + rebuild_sched_domains_cpuslocked(); free_cpumasks(NULL, ptmp); } @@ -4679,12 +4019,7 @@ void cpuset_update_active_cpus(void) * inside cgroup synchronization. Bounce actual hotplug processing * to a work item to avoid reverse locking order. */ - schedule_work(&cpuset_hotplug_work); -} - -void cpuset_wait_for_hotplug(void) -{ - flush_work(&cpuset_hotplug_work); + cpuset_handle_hotplug(); } /* @@ -4695,7 +4030,7 @@ void cpuset_wait_for_hotplug(void) static int cpuset_track_online_nodes(struct notifier_block *self, unsigned long action, void *arg) { - schedule_work(&cpuset_hotplug_work); + cpuset_handle_hotplug(); return NOTIFY_OK; } @@ -4729,7 +4064,7 @@ void __init cpuset_init_smp(void) * * Description: Returns the cpumask_var_t cpus_allowed of the cpuset * attached to the specified @tsk. Guaranteed to return some non-empty - * subset of cpu_online_mask, even if this means going outside the + * subset of cpu_active_mask, even if this means going outside the * tasks cpuset, except when the task is in the top cpuset. **/ @@ -4743,7 +4078,7 @@ void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask) cs = task_cs(tsk); if (cs != &top_cpuset) - guarantee_online_cpus(tsk, pmask); + guarantee_active_cpus(tsk, pmask); /* * Tasks in the top cpuset won't get update to their cpumasks * when a hotplug online/offline event happens. So we include all @@ -4757,7 +4092,7 @@ void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask) * allowable online cpu left, we fall back to all possible cpus. */ cpumask_andnot(pmask, possible_mask, subpartitions_cpus); - if (!cpumask_intersects(pmask, cpu_online_mask)) + if (!cpumask_intersects(pmask, cpu_active_mask)) cpumask_copy(pmask, possible_mask); } @@ -4867,7 +4202,7 @@ static struct cpuset *nearest_hardwall_ancestor(struct cpuset *cs) } /* - * cpuset_node_allowed - Can we allocate on a memory node? + * cpuset_current_node_allowed - Can current task allocate on a memory node? * @node: is this an allowed node? * @gfp_mask: memory allocation flags * @@ -4906,7 +4241,7 @@ static struct cpuset *nearest_hardwall_ancestor(struct cpuset *cs) * GFP_KERNEL - any node in enclosing hardwalled cpuset ok * GFP_USER - only nodes in current tasks mems allowed ok. */ -bool cpuset_node_allowed(int node, gfp_t gfp_mask) +bool cpuset_current_node_allowed(int node, gfp_t gfp_mask) { struct cpuset *cs; /* current cpuset ancestors */ bool allowed; /* is allocation in zone z allowed? */ @@ -4940,6 +4275,42 @@ bool cpuset_node_allowed(int node, gfp_t gfp_mask) return allowed; } +bool cpuset_node_allowed(struct cgroup *cgroup, int nid) +{ + struct cgroup_subsys_state *css; + struct cpuset *cs; + bool allowed; + + /* + * In v1, mem_cgroup and cpuset are unlikely in the same hierarchy + * and mems_allowed is likely to be empty even if we could get to it, + * so return true to avoid taking a global lock on the empty check. + */ + if (!cpuset_v2()) + return true; + + css = cgroup_get_e_css(cgroup, &cpuset_cgrp_subsys); + if (!css) + return true; + + /* + * Normally, accessing effective_mems would require the cpuset_mutex + * or callback_lock - but node_isset is atomic and the reference + * taken via cgroup_get_e_css is sufficient to protect css. + * + * Since this interface is intended for use by migration paths, we + * relax locking here to avoid taking global locks - while accepting + * there may be rare scenarios where the result may be innaccurate. + * + * Reclaim and migration are subject to these same race conditions, and + * cannot make strong isolation guarantees, so this is acceptable. + */ + cs = container_of(css, struct cpuset, css); + allowed = node_isset(nid, cs->effective_mems); + css_put(css); + return allowed; +} + /** * cpuset_spread_node() - On which node to begin search for a page * @rotor: round robin rotor @@ -4984,19 +4355,6 @@ int cpuset_mem_spread_node(void) } /** - * cpuset_slab_spread_node() - On which node to begin search for a slab page - */ -int cpuset_slab_spread_node(void) -{ - if (current->cpuset_slab_spread_rotor == NUMA_NO_NODE) - current->cpuset_slab_spread_rotor = - node_random(¤t->mems_allowed); - - return cpuset_spread_node(¤t->cpuset_slab_spread_rotor); -} -EXPORT_SYMBOL_GPL(cpuset_mem_spread_node); - -/** * cpuset_mems_allowed_intersects - Does @tsk1's mems_allowed intersect @tsk2's? * @tsk1: pointer to task_struct of some task. * @tsk2: pointer to task_struct of some other task. @@ -5034,79 +4392,6 @@ void cpuset_print_current_mems_allowed(void) rcu_read_unlock(); } -/* - * Collection of memory_pressure is suppressed unless - * this flag is enabled by writing "1" to the special - * cpuset file 'memory_pressure_enabled' in the root cpuset. - */ - -int cpuset_memory_pressure_enabled __read_mostly; - -/* - * __cpuset_memory_pressure_bump - keep stats of per-cpuset reclaims. - * - * Keep a running average of the rate of synchronous (direct) - * page reclaim efforts initiated by tasks in each cpuset. - * - * This represents the rate at which some task in the cpuset - * ran low on memory on all nodes it was allowed to use, and - * had to enter the kernels page reclaim code in an effort to - * create more free memory by tossing clean pages or swapping - * or writing dirty pages. - * - * Display to user space in the per-cpuset read-only file - * "memory_pressure". Value displayed is an integer - * representing the recent rate of entry into the synchronous - * (direct) page reclaim by any task attached to the cpuset. - */ - -void __cpuset_memory_pressure_bump(void) -{ - rcu_read_lock(); - fmeter_markevent(&task_cs(current)->fmeter); - rcu_read_unlock(); -} - -#ifdef CONFIG_PROC_PID_CPUSET -/* - * proc_cpuset_show() - * - Print tasks cpuset path into seq_file. - * - Used for /proc/<pid>/cpuset. - * - No need to task_lock(tsk) on this tsk->cpuset reference, as it - * doesn't really matter if tsk->cpuset changes after we read it, - * and we take cpuset_mutex, keeping cpuset_attach() from changing it - * anyway. - */ -int proc_cpuset_show(struct seq_file *m, struct pid_namespace *ns, - struct pid *pid, struct task_struct *tsk) -{ - char *buf; - struct cgroup_subsys_state *css; - int retval; - - retval = -ENOMEM; - buf = kmalloc(PATH_MAX, GFP_KERNEL); - if (!buf) - goto out; - - css = task_get_css(tsk, cpuset_cgrp_id); - retval = cgroup_path_ns(css->cgroup, buf, PATH_MAX, - current->nsproxy->cgroup_ns); - css_put(css); - if (retval == -E2BIG) - retval = -ENAMETOOLONG; - if (retval < 0) - goto out_free; - seq_puts(m, buf); - seq_putc(m, '\n'); - retval = 0; -out_free: - kfree(buf); -out: - return retval; -} -#endif /* CONFIG_PROC_PID_CPUSET */ - /* Display task mems_allowed in /proc/<pid>/status file. */ void cpuset_task_status_allowed(struct seq_file *m, struct task_struct *task) { diff --git a/kernel/cgroup/dmem.c b/kernel/cgroup/dmem.c new file mode 100644 index 000000000000..10b63433f057 --- /dev/null +++ b/kernel/cgroup/dmem.c @@ -0,0 +1,829 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright 2023-2024 Intel Corporation (Maarten Lankhorst <dev@lankhorst.se>) + * Copyright 2024 Red Hat (Maxime Ripard <mripard@kernel.org>) + * Partially based on the rdma and misc controllers, which bear the following copyrights: + * + * Copyright 2020 Google LLC + * Copyright (C) 2016 Parav Pandit <pandit.parav@gmail.com> + */ + +#include <linux/cgroup.h> +#include <linux/cgroup_dmem.h> +#include <linux/list.h> +#include <linux/mutex.h> +#include <linux/page_counter.h> +#include <linux/parser.h> +#include <linux/slab.h> + +struct dmem_cgroup_region { + /** + * @ref: References keeping the region alive. + * Keeps the region reference alive after a succesful RCU lookup. + */ + struct kref ref; + + /** @rcu: RCU head for freeing */ + struct rcu_head rcu; + + /** + * @region_node: Linked into &dmem_cgroup_regions list. + * Protected by RCU and global spinlock. + */ + struct list_head region_node; + + /** + * @pools: List of pools linked to this region. + * Protected by global spinlock only + */ + struct list_head pools; + + /** @size: Size of region, in bytes */ + u64 size; + + /** @name: Name describing the node, set by dmem_cgroup_register_region */ + char *name; + + /** + * @unregistered: Whether the region is unregistered by its caller. + * No new pools should be added to the region afterwards. + */ + bool unregistered; +}; + +struct dmemcg_state { + struct cgroup_subsys_state css; + + struct list_head pools; +}; + +struct dmem_cgroup_pool_state { + struct dmem_cgroup_region *region; + struct dmemcg_state *cs; + + /* css node, RCU protected against region teardown */ + struct list_head css_node; + + /* dev node, no RCU protection required */ + struct list_head region_node; + + struct rcu_head rcu; + + struct page_counter cnt; + + bool inited; +}; + +/* + * 3 operations require locking protection: + * - Registering and unregistering region to/from list, requires global lock. + * - Adding a dmem_cgroup_pool_state to a CSS, removing when CSS is freed. + * - Adding a dmem_cgroup_pool_state to a region list. + * + * Since for the most common operations RCU provides enough protection, I + * do not think more granular locking makes sense. Most protection is offered + * by RCU and the lockless operating page_counter. + */ +static DEFINE_SPINLOCK(dmemcg_lock); +static LIST_HEAD(dmem_cgroup_regions); + +static inline struct dmemcg_state * +css_to_dmemcs(struct cgroup_subsys_state *css) +{ + return container_of(css, struct dmemcg_state, css); +} + +static inline struct dmemcg_state *get_current_dmemcs(void) +{ + return css_to_dmemcs(task_get_css(current, dmem_cgrp_id)); +} + +static struct dmemcg_state *parent_dmemcs(struct dmemcg_state *cg) +{ + return cg->css.parent ? css_to_dmemcs(cg->css.parent) : NULL; +} + +static void free_cg_pool(struct dmem_cgroup_pool_state *pool) +{ + list_del(&pool->region_node); + kfree(pool); +} + +static void +set_resource_min(struct dmem_cgroup_pool_state *pool, u64 val) +{ + page_counter_set_min(&pool->cnt, val); +} + +static void +set_resource_low(struct dmem_cgroup_pool_state *pool, u64 val) +{ + page_counter_set_low(&pool->cnt, val); +} + +static void +set_resource_max(struct dmem_cgroup_pool_state *pool, u64 val) +{ + page_counter_set_max(&pool->cnt, val); +} + +static u64 get_resource_low(struct dmem_cgroup_pool_state *pool) +{ + return pool ? READ_ONCE(pool->cnt.low) : 0; +} + +static u64 get_resource_min(struct dmem_cgroup_pool_state *pool) +{ + return pool ? READ_ONCE(pool->cnt.min) : 0; +} + +static u64 get_resource_max(struct dmem_cgroup_pool_state *pool) +{ + return pool ? READ_ONCE(pool->cnt.max) : PAGE_COUNTER_MAX; +} + +static u64 get_resource_current(struct dmem_cgroup_pool_state *pool) +{ + return pool ? page_counter_read(&pool->cnt) : 0; +} + +static void reset_all_resource_limits(struct dmem_cgroup_pool_state *rpool) +{ + set_resource_min(rpool, 0); + set_resource_low(rpool, 0); + set_resource_max(rpool, PAGE_COUNTER_MAX); +} + +static void dmemcs_offline(struct cgroup_subsys_state *css) +{ + struct dmemcg_state *dmemcs = css_to_dmemcs(css); + struct dmem_cgroup_pool_state *pool; + + rcu_read_lock(); + list_for_each_entry_rcu(pool, &dmemcs->pools, css_node) + reset_all_resource_limits(pool); + rcu_read_unlock(); +} + +static void dmemcs_free(struct cgroup_subsys_state *css) +{ + struct dmemcg_state *dmemcs = css_to_dmemcs(css); + struct dmem_cgroup_pool_state *pool, *next; + + spin_lock(&dmemcg_lock); + list_for_each_entry_safe(pool, next, &dmemcs->pools, css_node) { + /* + *The pool is dead and all references are 0, + * no need for RCU protection with list_del_rcu or freeing. + */ + list_del(&pool->css_node); + free_cg_pool(pool); + } + spin_unlock(&dmemcg_lock); + + kfree(dmemcs); +} + +static struct cgroup_subsys_state * +dmemcs_alloc(struct cgroup_subsys_state *parent_css) +{ + struct dmemcg_state *dmemcs = kzalloc(sizeof(*dmemcs), GFP_KERNEL); + if (!dmemcs) + return ERR_PTR(-ENOMEM); + + INIT_LIST_HEAD(&dmemcs->pools); + return &dmemcs->css; +} + +static struct dmem_cgroup_pool_state * +find_cg_pool_locked(struct dmemcg_state *dmemcs, struct dmem_cgroup_region *region) +{ + struct dmem_cgroup_pool_state *pool; + + list_for_each_entry_rcu(pool, &dmemcs->pools, css_node, spin_is_locked(&dmemcg_lock)) + if (pool->region == region) + return pool; + + return NULL; +} + +static struct dmem_cgroup_pool_state *pool_parent(struct dmem_cgroup_pool_state *pool) +{ + if (!pool->cnt.parent) + return NULL; + + return container_of(pool->cnt.parent, typeof(*pool), cnt); +} + +static void +dmem_cgroup_calculate_protection(struct dmem_cgroup_pool_state *limit_pool, + struct dmem_cgroup_pool_state *test_pool) +{ + struct page_counter *climit; + struct cgroup_subsys_state *css; + struct dmemcg_state *dmemcg_iter; + struct dmem_cgroup_pool_state *pool, *found_pool; + + climit = &limit_pool->cnt; + + rcu_read_lock(); + + css_for_each_descendant_pre(css, &limit_pool->cs->css) { + dmemcg_iter = container_of(css, struct dmemcg_state, css); + found_pool = NULL; + + list_for_each_entry_rcu(pool, &dmemcg_iter->pools, css_node) { + if (pool->region == limit_pool->region) { + found_pool = pool; + break; + } + } + if (!found_pool) + continue; + + page_counter_calculate_protection( + climit, &found_pool->cnt, true); + + if (found_pool == test_pool) + break; + } + rcu_read_unlock(); +} + +/** + * dmem_cgroup_state_evict_valuable() - Check if we should evict from test_pool + * @limit_pool: The pool for which we hit limits + * @test_pool: The pool for which to test + * @ignore_low: Whether we have to respect low watermarks. + * @ret_hit_low: Pointer to whether it makes sense to consider low watermark. + * + * This function returns true if we can evict from @test_pool, false if not. + * When returning false and @ignore_low is false, @ret_hit_low may + * be set to true to indicate this function can be retried with @ignore_low + * set to true. + * + * Return: bool + */ +bool dmem_cgroup_state_evict_valuable(struct dmem_cgroup_pool_state *limit_pool, + struct dmem_cgroup_pool_state *test_pool, + bool ignore_low, bool *ret_hit_low) +{ + struct dmem_cgroup_pool_state *pool = test_pool; + struct page_counter *ctest; + u64 used, min, low; + + /* Can always evict from current pool, despite limits */ + if (limit_pool == test_pool) + return true; + + if (limit_pool) { + if (!parent_dmemcs(limit_pool->cs)) + return true; + + for (pool = test_pool; pool && limit_pool != pool; pool = pool_parent(pool)) + {} + + if (!pool) + return false; + } else { + /* + * If there is no cgroup limiting memory usage, use the root + * cgroup instead for limit calculations. + */ + for (limit_pool = test_pool; pool_parent(limit_pool); limit_pool = pool_parent(limit_pool)) + {} + } + + ctest = &test_pool->cnt; + + dmem_cgroup_calculate_protection(limit_pool, test_pool); + + used = page_counter_read(ctest); + min = READ_ONCE(ctest->emin); + + if (used <= min) + return false; + + if (!ignore_low) { + low = READ_ONCE(ctest->elow); + if (used > low) + return true; + + *ret_hit_low = true; + return false; + } + return true; +} +EXPORT_SYMBOL_GPL(dmem_cgroup_state_evict_valuable); + +static struct dmem_cgroup_pool_state * +alloc_pool_single(struct dmemcg_state *dmemcs, struct dmem_cgroup_region *region, + struct dmem_cgroup_pool_state **allocpool) +{ + struct dmemcg_state *parent = parent_dmemcs(dmemcs); + struct dmem_cgroup_pool_state *pool, *ppool = NULL; + + if (!*allocpool) { + pool = kzalloc(sizeof(*pool), GFP_NOWAIT); + if (!pool) + return ERR_PTR(-ENOMEM); + } else { + pool = *allocpool; + *allocpool = NULL; + } + + pool->region = region; + pool->cs = dmemcs; + + if (parent) + ppool = find_cg_pool_locked(parent, region); + + page_counter_init(&pool->cnt, + ppool ? &ppool->cnt : NULL, true); + reset_all_resource_limits(pool); + + list_add_tail_rcu(&pool->css_node, &dmemcs->pools); + list_add_tail(&pool->region_node, ®ion->pools); + + if (!parent) + pool->inited = true; + else + pool->inited = ppool ? ppool->inited : false; + return pool; +} + +static struct dmem_cgroup_pool_state * +get_cg_pool_locked(struct dmemcg_state *dmemcs, struct dmem_cgroup_region *region, + struct dmem_cgroup_pool_state **allocpool) +{ + struct dmem_cgroup_pool_state *pool, *ppool, *retpool; + struct dmemcg_state *p, *pp; + + /* + * Recursively create pool, we may not initialize yet on + * recursion, this is done as a separate step. + */ + for (p = dmemcs; p; p = parent_dmemcs(p)) { + pool = find_cg_pool_locked(p, region); + if (!pool) + pool = alloc_pool_single(p, region, allocpool); + + if (IS_ERR(pool)) + return pool; + + if (p == dmemcs && pool->inited) + return pool; + + if (pool->inited) + break; + } + + retpool = pool = find_cg_pool_locked(dmemcs, region); + for (p = dmemcs, pp = parent_dmemcs(dmemcs); pp; p = pp, pp = parent_dmemcs(p)) { + if (pool->inited) + break; + + /* ppool was created if it didn't exist by above loop. */ + ppool = find_cg_pool_locked(pp, region); + + /* Fix up parent links, mark as inited. */ + pool->cnt.parent = &ppool->cnt; + pool->inited = true; + + pool = ppool; + } + + return retpool; +} + +static void dmemcg_free_rcu(struct rcu_head *rcu) +{ + struct dmem_cgroup_region *region = container_of(rcu, typeof(*region), rcu); + struct dmem_cgroup_pool_state *pool, *next; + + list_for_each_entry_safe(pool, next, ®ion->pools, region_node) + free_cg_pool(pool); + kfree(region->name); + kfree(region); +} + +static void dmemcg_free_region(struct kref *ref) +{ + struct dmem_cgroup_region *cgregion = container_of(ref, typeof(*cgregion), ref); + + call_rcu(&cgregion->rcu, dmemcg_free_rcu); +} + +/** + * dmem_cgroup_unregister_region() - Unregister a previously registered region. + * @region: The region to unregister. + * + * This function undoes dmem_cgroup_register_region. + */ +void dmem_cgroup_unregister_region(struct dmem_cgroup_region *region) +{ + struct list_head *entry; + + if (!region) + return; + + spin_lock(&dmemcg_lock); + + /* Remove from global region list */ + list_del_rcu(®ion->region_node); + + list_for_each_rcu(entry, ®ion->pools) { + struct dmem_cgroup_pool_state *pool = + container_of(entry, typeof(*pool), region_node); + + list_del_rcu(&pool->css_node); + } + + /* + * Ensure any RCU based lookups fail. Additionally, + * no new pools should be added to the dead region + * by get_cg_pool_unlocked. + */ + region->unregistered = true; + spin_unlock(&dmemcg_lock); + + kref_put(®ion->ref, dmemcg_free_region); +} +EXPORT_SYMBOL_GPL(dmem_cgroup_unregister_region); + +/** + * dmem_cgroup_register_region() - Register a regions for dev cgroup. + * @size: Size of region to register, in bytes. + * @fmt: Region parameters to register + * + * This function registers a node in the dmem cgroup with the + * name given. After calling this function, the region can be + * used for allocations. + * + * Return: NULL or a struct on success, PTR_ERR on failure. + */ +struct dmem_cgroup_region *dmem_cgroup_register_region(u64 size, const char *fmt, ...) +{ + struct dmem_cgroup_region *ret; + char *region_name; + va_list ap; + + if (!size) + return NULL; + + va_start(ap, fmt); + region_name = kvasprintf(GFP_KERNEL, fmt, ap); + va_end(ap); + if (!region_name) + return ERR_PTR(-ENOMEM); + + ret = kzalloc(sizeof(*ret), GFP_KERNEL); + if (!ret) { + kfree(region_name); + return ERR_PTR(-ENOMEM); + } + + INIT_LIST_HEAD(&ret->pools); + ret->name = region_name; + ret->size = size; + kref_init(&ret->ref); + + spin_lock(&dmemcg_lock); + list_add_tail_rcu(&ret->region_node, &dmem_cgroup_regions); + spin_unlock(&dmemcg_lock); + + return ret; +} +EXPORT_SYMBOL_GPL(dmem_cgroup_register_region); + +static struct dmem_cgroup_region *dmemcg_get_region_by_name(const char *name) +{ + struct dmem_cgroup_region *region; + + list_for_each_entry_rcu(region, &dmem_cgroup_regions, region_node, spin_is_locked(&dmemcg_lock)) + if (!strcmp(name, region->name) && + kref_get_unless_zero(®ion->ref)) + return region; + + return NULL; +} + +/** + * dmem_cgroup_pool_state_put() - Drop a reference to a dmem_cgroup_pool_state + * @pool: &dmem_cgroup_pool_state + * + * Called to drop a reference to the limiting pool returned by + * dmem_cgroup_try_charge(). + */ +void dmem_cgroup_pool_state_put(struct dmem_cgroup_pool_state *pool) +{ + if (pool) + css_put(&pool->cs->css); +} +EXPORT_SYMBOL_GPL(dmem_cgroup_pool_state_put); + +static struct dmem_cgroup_pool_state * +get_cg_pool_unlocked(struct dmemcg_state *cg, struct dmem_cgroup_region *region) +{ + struct dmem_cgroup_pool_state *pool, *allocpool = NULL; + + /* fastpath lookup? */ + rcu_read_lock(); + pool = find_cg_pool_locked(cg, region); + if (pool && !READ_ONCE(pool->inited)) + pool = NULL; + rcu_read_unlock(); + + while (!pool) { + spin_lock(&dmemcg_lock); + if (!region->unregistered) + pool = get_cg_pool_locked(cg, region, &allocpool); + else + pool = ERR_PTR(-ENODEV); + spin_unlock(&dmemcg_lock); + + if (pool == ERR_PTR(-ENOMEM)) { + pool = NULL; + if (WARN_ON(allocpool)) + continue; + + allocpool = kzalloc(sizeof(*allocpool), GFP_KERNEL); + if (allocpool) { + pool = NULL; + continue; + } + } + } + + kfree(allocpool); + return pool; +} + +/** + * dmem_cgroup_uncharge() - Uncharge a pool. + * @pool: Pool to uncharge. + * @size: Size to uncharge. + * + * Undoes the effects of dmem_cgroup_try_charge. + * Must be called with the returned pool as argument, + * and same @index and @size. + */ +void dmem_cgroup_uncharge(struct dmem_cgroup_pool_state *pool, u64 size) +{ + if (!pool) + return; + + page_counter_uncharge(&pool->cnt, size); + css_put(&pool->cs->css); +} +EXPORT_SYMBOL_GPL(dmem_cgroup_uncharge); + +/** + * dmem_cgroup_try_charge() - Try charging a new allocation to a region. + * @region: dmem region to charge + * @size: Size (in bytes) to charge. + * @ret_pool: On succesfull allocation, the pool that is charged. + * @ret_limit_pool: On a failed allocation, the limiting pool. + * + * This function charges the @region region for a size of @size bytes. + * + * If the function succeeds, @ret_pool is set, which must be passed to + * dmem_cgroup_uncharge() when undoing the allocation. + * + * When this function fails with -EAGAIN and @ret_limit_pool is non-null, it + * will be set to the pool for which the limit is hit. This can be used for + * eviction as argument to dmem_cgroup_evict_valuable(). This reference must be freed + * with @dmem_cgroup_pool_state_put(). + * + * Return: 0 on success, -EAGAIN on hitting a limit, or a negative errno on failure. + */ +int dmem_cgroup_try_charge(struct dmem_cgroup_region *region, u64 size, + struct dmem_cgroup_pool_state **ret_pool, + struct dmem_cgroup_pool_state **ret_limit_pool) +{ + struct dmemcg_state *cg; + struct dmem_cgroup_pool_state *pool; + struct page_counter *fail; + int ret; + + *ret_pool = NULL; + if (ret_limit_pool) + *ret_limit_pool = NULL; + + /* + * hold on to css, as cgroup can be removed but resource + * accounting happens on css. + */ + cg = get_current_dmemcs(); + + pool = get_cg_pool_unlocked(cg, region); + if (IS_ERR(pool)) { + ret = PTR_ERR(pool); + goto err; + } + + if (!page_counter_try_charge(&pool->cnt, size, &fail)) { + if (ret_limit_pool) { + *ret_limit_pool = container_of(fail, struct dmem_cgroup_pool_state, cnt); + css_get(&(*ret_limit_pool)->cs->css); + } + ret = -EAGAIN; + goto err; + } + + /* On success, reference from get_current_dmemcs is transferred to *ret_pool */ + *ret_pool = pool; + return 0; + +err: + css_put(&cg->css); + return ret; +} +EXPORT_SYMBOL_GPL(dmem_cgroup_try_charge); + +static int dmem_cgroup_region_capacity_show(struct seq_file *sf, void *v) +{ + struct dmem_cgroup_region *region; + + rcu_read_lock(); + list_for_each_entry_rcu(region, &dmem_cgroup_regions, region_node) { + seq_puts(sf, region->name); + seq_printf(sf, " %llu\n", region->size); + } + rcu_read_unlock(); + return 0; +} + +static int dmemcg_parse_limit(char *options, struct dmem_cgroup_region *region, + u64 *new_limit) +{ + char *end; + + if (!strcmp(options, "max")) { + *new_limit = PAGE_COUNTER_MAX; + return 0; + } + + *new_limit = memparse(options, &end); + if (*end != '\0') + return -EINVAL; + + return 0; +} + +static ssize_t dmemcg_limit_write(struct kernfs_open_file *of, + char *buf, size_t nbytes, loff_t off, + void (*apply)(struct dmem_cgroup_pool_state *, u64)) +{ + struct dmemcg_state *dmemcs = css_to_dmemcs(of_css(of)); + int err = 0; + + while (buf && !err) { + struct dmem_cgroup_pool_state *pool = NULL; + char *options, *region_name; + struct dmem_cgroup_region *region; + u64 new_limit; + + options = buf; + buf = strchr(buf, '\n'); + if (buf) + *buf++ = '\0'; + + options = strstrip(options); + + /* eat empty lines */ + if (!options[0]) + continue; + + region_name = strsep(&options, " \t"); + if (!region_name[0]) + continue; + + rcu_read_lock(); + region = dmemcg_get_region_by_name(region_name); + rcu_read_unlock(); + + if (!region) + return -EINVAL; + + err = dmemcg_parse_limit(options, region, &new_limit); + if (err < 0) + goto out_put; + + pool = get_cg_pool_unlocked(dmemcs, region); + if (IS_ERR(pool)) { + err = PTR_ERR(pool); + goto out_put; + } + + /* And commit */ + apply(pool, new_limit); + +out_put: + kref_put(®ion->ref, dmemcg_free_region); + } + + + return err ?: nbytes; +} + +static int dmemcg_limit_show(struct seq_file *sf, void *v, + u64 (*fn)(struct dmem_cgroup_pool_state *)) +{ + struct dmemcg_state *dmemcs = css_to_dmemcs(seq_css(sf)); + struct dmem_cgroup_region *region; + + rcu_read_lock(); + list_for_each_entry_rcu(region, &dmem_cgroup_regions, region_node) { + struct dmem_cgroup_pool_state *pool = find_cg_pool_locked(dmemcs, region); + u64 val; + + seq_puts(sf, region->name); + + val = fn(pool); + if (val < PAGE_COUNTER_MAX) + seq_printf(sf, " %lld\n", val); + else + seq_puts(sf, " max\n"); + } + rcu_read_unlock(); + + return 0; +} + +static int dmem_cgroup_region_current_show(struct seq_file *sf, void *v) +{ + return dmemcg_limit_show(sf, v, get_resource_current); +} + +static int dmem_cgroup_region_min_show(struct seq_file *sf, void *v) +{ + return dmemcg_limit_show(sf, v, get_resource_min); +} + +static ssize_t dmem_cgroup_region_min_write(struct kernfs_open_file *of, + char *buf, size_t nbytes, loff_t off) +{ + return dmemcg_limit_write(of, buf, nbytes, off, set_resource_min); +} + +static int dmem_cgroup_region_low_show(struct seq_file *sf, void *v) +{ + return dmemcg_limit_show(sf, v, get_resource_low); +} + +static ssize_t dmem_cgroup_region_low_write(struct kernfs_open_file *of, + char *buf, size_t nbytes, loff_t off) +{ + return dmemcg_limit_write(of, buf, nbytes, off, set_resource_low); +} + +static int dmem_cgroup_region_max_show(struct seq_file *sf, void *v) +{ + return dmemcg_limit_show(sf, v, get_resource_max); +} + +static ssize_t dmem_cgroup_region_max_write(struct kernfs_open_file *of, + char *buf, size_t nbytes, loff_t off) +{ + return dmemcg_limit_write(of, buf, nbytes, off, set_resource_max); +} + +static struct cftype files[] = { + { + .name = "capacity", + .seq_show = dmem_cgroup_region_capacity_show, + .flags = CFTYPE_ONLY_ON_ROOT, + }, + { + .name = "current", + .seq_show = dmem_cgroup_region_current_show, + }, + { + .name = "min", + .write = dmem_cgroup_region_min_write, + .seq_show = dmem_cgroup_region_min_show, + .flags = CFTYPE_NOT_ON_ROOT, + }, + { + .name = "low", + .write = dmem_cgroup_region_low_write, + .seq_show = dmem_cgroup_region_low_show, + .flags = CFTYPE_NOT_ON_ROOT, + }, + { + .name = "max", + .write = dmem_cgroup_region_max_write, + .seq_show = dmem_cgroup_region_max_show, + .flags = CFTYPE_NOT_ON_ROOT, + }, + { } /* Zero entry terminates. */ +}; + +struct cgroup_subsys dmem_cgrp_subsys = { + .css_alloc = dmemcs_alloc, + .css_free = dmemcs_free, + .css_offline = dmemcs_offline, + .legacy_cftypes = files, + .dfl_cftypes = files, +}; diff --git a/kernel/cgroup/freezer.c b/kernel/cgroup/freezer.c index 617861a54793..bf1690a167dd 100644 --- a/kernel/cgroup/freezer.c +++ b/kernel/cgroup/freezer.c @@ -9,6 +9,28 @@ #include <trace/events/cgroup.h> /* + * Update CGRP_FROZEN of cgroup.flag + * Return true if flags is updated; false if flags has no change + */ +static bool cgroup_update_frozen_flag(struct cgroup *cgrp, bool frozen) +{ + lockdep_assert_held(&css_set_lock); + + /* Already there? */ + if (test_bit(CGRP_FROZEN, &cgrp->flags) == frozen) + return false; + + if (frozen) + set_bit(CGRP_FROZEN, &cgrp->flags); + else + clear_bit(CGRP_FROZEN, &cgrp->flags); + + cgroup_file_notify(&cgrp->events_file); + TRACE_CGROUP_PATH(notify_frozen, cgrp, frozen); + return true; +} + +/* * Propagate the cgroup frozen state upwards by the cgroup tree. */ static void cgroup_propagate_frozen(struct cgroup *cgrp, bool frozen) @@ -24,24 +46,16 @@ static void cgroup_propagate_frozen(struct cgroup *cgrp, bool frozen) while ((cgrp = cgroup_parent(cgrp))) { if (frozen) { cgrp->freezer.nr_frozen_descendants += desc; - if (!test_bit(CGRP_FROZEN, &cgrp->flags) && - test_bit(CGRP_FREEZE, &cgrp->flags) && - cgrp->freezer.nr_frozen_descendants == - cgrp->nr_descendants) { - set_bit(CGRP_FROZEN, &cgrp->flags); - cgroup_file_notify(&cgrp->events_file); - TRACE_CGROUP_PATH(notify_frozen, cgrp, 1); - desc++; - } + if (!test_bit(CGRP_FREEZE, &cgrp->flags) || + (cgrp->freezer.nr_frozen_descendants != + cgrp->nr_descendants)) + continue; } else { cgrp->freezer.nr_frozen_descendants -= desc; - if (test_bit(CGRP_FROZEN, &cgrp->flags)) { - clear_bit(CGRP_FROZEN, &cgrp->flags); - cgroup_file_notify(&cgrp->events_file); - TRACE_CGROUP_PATH(notify_frozen, cgrp, 0); - desc++; - } } + + if (cgroup_update_frozen_flag(cgrp, frozen)) + desc++; } } @@ -53,8 +67,6 @@ void cgroup_update_frozen(struct cgroup *cgrp) { bool frozen; - lockdep_assert_held(&css_set_lock); - /* * If the cgroup has to be frozen (CGRP_FREEZE bit set), * and all tasks are frozen and/or stopped, let's consider @@ -63,24 +75,9 @@ void cgroup_update_frozen(struct cgroup *cgrp) frozen = test_bit(CGRP_FREEZE, &cgrp->flags) && cgrp->freezer.nr_frozen_tasks == __cgroup_task_count(cgrp); - if (frozen) { - /* Already there? */ - if (test_bit(CGRP_FROZEN, &cgrp->flags)) - return; - - set_bit(CGRP_FROZEN, &cgrp->flags); - } else { - /* Already there? */ - if (!test_bit(CGRP_FROZEN, &cgrp->flags)) - return; - - clear_bit(CGRP_FROZEN, &cgrp->flags); - } - cgroup_file_notify(&cgrp->events_file); - TRACE_CGROUP_PATH(notify_frozen, cgrp, frozen); - - /* Update the state of ancestor cgroups. */ - cgroup_propagate_frozen(cgrp, frozen); + /* If flags is updated, update the state of ancestor cgroups. */ + if (cgroup_update_frozen_flag(cgrp, frozen)) + cgroup_propagate_frozen(cgrp, frozen); } /* @@ -260,8 +257,10 @@ void cgroup_freezer_migrate_task(struct task_struct *task, void cgroup_freeze(struct cgroup *cgrp, bool freeze) { struct cgroup_subsys_state *css; + struct cgroup *parent; struct cgroup *dsct; bool applied = false; + bool old_e; lockdep_assert_held(&cgroup_mutex); @@ -282,22 +281,18 @@ void cgroup_freeze(struct cgroup *cgrp, bool freeze) if (cgroup_is_dead(dsct)) continue; - if (freeze) { - dsct->freezer.e_freeze++; - /* - * Already frozen because of ancestor's settings? - */ - if (dsct->freezer.e_freeze > 1) - continue; - } else { - dsct->freezer.e_freeze--; - /* - * Still frozen because of ancestor's settings? - */ - if (dsct->freezer.e_freeze > 0) - continue; - - WARN_ON_ONCE(dsct->freezer.e_freeze < 0); + /* + * e_freeze is affected by parent's e_freeze and dst's freeze. + * If old e_freeze eq new e_freeze, no change, its children + * will not be affected. So do nothing and skip the subtree + */ + old_e = dsct->freezer.e_freeze; + parent = cgroup_parent(dsct); + dsct->freezer.e_freeze = (dsct->freezer.freeze || + parent->freezer.e_freeze); + if (dsct->freezer.e_freeze == old_e) { + css = css_rightmost_descendant(css); + continue; } /* diff --git a/kernel/cgroup/legacy_freezer.c b/kernel/cgroup/legacy_freezer.c index 66d1708042a7..039d1eb2f215 100644 --- a/kernel/cgroup/legacy_freezer.c +++ b/kernel/cgroup/legacy_freezer.c @@ -106,8 +106,7 @@ freezer_css_alloc(struct cgroup_subsys_state *parent_css) * @css: css being created * * We're committing to creation of @css. Mark it online and inherit - * parent's freezing state while holding both parent's and our - * freezer->lock. + * parent's freezing state while holding cpus read lock and freezer_mutex. */ static int freezer_css_online(struct cgroup_subsys_state *css) { @@ -133,7 +132,7 @@ static int freezer_css_online(struct cgroup_subsys_state *css) * freezer_css_offline - initiate destruction of a freezer css * @css: css being destroyed * - * @css is going away. Mark it dead and decrement system_freezing_count if + * @css is going away. Mark it dead and decrement freezer_active if * it was holding one. */ static void freezer_css_offline(struct cgroup_subsys_state *css) @@ -431,9 +430,11 @@ static ssize_t freezer_write(struct kernfs_open_file *of, if (strcmp(buf, freezer_state_strs(0)) == 0) freeze = false; - else if (strcmp(buf, freezer_state_strs(CGROUP_FROZEN)) == 0) + else if (strcmp(buf, freezer_state_strs(CGROUP_FROZEN)) == 0) { + pr_info_once("Freezing with imperfect legacy cgroup freezer. " + "See cgroup.freeze of cgroup v2\n"); freeze = true; - else + } else return -EINVAL; freezer_change_state(css_freezer(of_css(of)), freeze); diff --git a/kernel/cgroup/misc.c b/kernel/cgroup/misc.c index 79a3717a5803..6a01d91ea4cb 100644 --- a/kernel/cgroup/misc.c +++ b/kernel/cgroup/misc.c @@ -24,6 +24,10 @@ static const char *const misc_res_name[] = { /* AMD SEV-ES ASIDs resource */ "sev_es", #endif +#ifdef CONFIG_INTEL_TDX_HOST + /* Intel TDX HKIDs resource */ + "tdx", +#endif }; /* Root misc cgroup */ @@ -68,22 +72,6 @@ static inline bool valid_type(enum misc_res_type type) } /** - * misc_cg_res_total_usage() - Get the current total usage of the resource. - * @type: misc res type. - * - * Context: Any context. - * Return: Current total usage of the resource. - */ -u64 misc_cg_res_total_usage(enum misc_res_type type) -{ - if (valid_type(type)) - return atomic64_read(&root_cg.res[type].usage); - - return 0; -} -EXPORT_SYMBOL_GPL(misc_cg_res_total_usage); - -/** * misc_cg_set_capacity() - Set the capacity of the misc cgroup res. * @type: Type of the misc res. * @capacity: Supported capacity of the misc res on the host. @@ -121,6 +109,30 @@ static void misc_cg_cancel_charge(enum misc_res_type type, struct misc_cg *cg, misc_res_name[type]); } +static void misc_cg_update_watermark(struct misc_res *res, u64 new_usage) +{ + u64 old; + + while (true) { + old = atomic64_read(&res->watermark); + if (new_usage <= old) + break; + if (atomic64_cmpxchg(&res->watermark, old, new_usage) == old) + break; + } +} + +static void misc_cg_event(enum misc_res_type type, struct misc_cg *cg) +{ + atomic64_inc(&cg->res[type].events_local); + cgroup_file_notify(&cg->events_local_file); + + for (; parent_misc(cg); cg = parent_misc(cg)) { + atomic64_inc(&cg->res[type].events); + cgroup_file_notify(&cg->events_file); + } +} + /** * misc_cg_try_charge() - Try charging the misc cgroup. * @type: Misc res type to charge. @@ -159,14 +171,12 @@ int misc_cg_try_charge(enum misc_res_type type, struct misc_cg *cg, u64 amount) ret = -EBUSY; goto err_charge; } + misc_cg_update_watermark(res, new_usage); } return 0; err_charge: - for (j = i; j; j = parent_misc(j)) { - atomic64_inc(&j->res[type].events); - cgroup_file_notify(&j->events_file); - } + misc_cg_event(type, i); for (j = cg; j != i; j = parent_misc(j)) misc_cg_cancel_charge(type, j, amount); @@ -308,6 +318,29 @@ static int misc_cg_current_show(struct seq_file *sf, void *v) } /** + * misc_cg_peak_show() - Show the peak usage of the misc cgroup. + * @sf: Interface file + * @v: Arguments passed + * + * Context: Any context. + * Return: 0 to denote successful print. + */ +static int misc_cg_peak_show(struct seq_file *sf, void *v) +{ + int i; + u64 watermark; + struct misc_cg *cg = css_misc(seq_css(sf)); + + for (i = 0; i < MISC_CG_RES_TYPES; i++) { + watermark = atomic64_read(&cg->res[i].watermark); + if (READ_ONCE(misc_res_capacity[i]) || watermark) + seq_printf(sf, "%s %llu\n", misc_res_name[i], watermark); + } + + return 0; +} + +/** * misc_cg_capacity_show() - Show the total capacity of misc res on the host. * @sf: Interface file * @v: Arguments passed @@ -331,20 +364,33 @@ static int misc_cg_capacity_show(struct seq_file *sf, void *v) return 0; } -static int misc_events_show(struct seq_file *sf, void *v) +static int __misc_events_show(struct seq_file *sf, bool local) { struct misc_cg *cg = css_misc(seq_css(sf)); u64 events; int i; for (i = 0; i < MISC_CG_RES_TYPES; i++) { - events = atomic64_read(&cg->res[i].events); + if (local) + events = atomic64_read(&cg->res[i].events_local); + else + events = atomic64_read(&cg->res[i].events); if (READ_ONCE(misc_res_capacity[i]) || events) seq_printf(sf, "%s.max %llu\n", misc_res_name[i], events); } return 0; } +static int misc_events_show(struct seq_file *sf, void *v) +{ + return __misc_events_show(sf, false); +} + +static int misc_events_local_show(struct seq_file *sf, void *v) +{ + return __misc_events_show(sf, true); +} + /* Misc cgroup interface files */ static struct cftype misc_cg_files[] = { { @@ -358,6 +404,10 @@ static struct cftype misc_cg_files[] = { .seq_show = misc_cg_current_show, }, { + .name = "peak", + .seq_show = misc_cg_peak_show, + }, + { .name = "capacity", .seq_show = misc_cg_capacity_show, .flags = CFTYPE_ONLY_ON_ROOT, @@ -368,6 +418,12 @@ static struct cftype misc_cg_files[] = { .file_offset = offsetof(struct misc_cg, events_file), .seq_show = misc_events_show, }, + { + .name = "events.local", + .flags = CFTYPE_NOT_ON_ROOT, + .file_offset = offsetof(struct misc_cg, events_local_file), + .seq_show = misc_events_local_show, + }, {} }; diff --git a/kernel/cgroup/pids.c b/kernel/cgroup/pids.c index 7695e60bcb40..8f61114c36dd 100644 --- a/kernel/cgroup/pids.c +++ b/kernel/cgroup/pids.c @@ -38,6 +38,14 @@ #define PIDS_MAX (PID_MAX_LIMIT + 1ULL) #define PIDS_MAX_STR "max" +enum pidcg_event { + /* Fork failed in subtree because this pids_cgroup limit was hit. */ + PIDCG_MAX, + /* Fork failed in this pids_cgroup because ancestor limit was hit. */ + PIDCG_FORKFAIL, + NR_PIDCG_EVENTS, +}; + struct pids_cgroup { struct cgroup_subsys_state css; @@ -49,11 +57,12 @@ struct pids_cgroup { atomic64_t limit; int64_t watermark; - /* Handle for "pids.events" */ + /* Handles for pids.events[.local] */ struct cgroup_file events_file; + struct cgroup_file events_local_file; - /* Number of times fork failed because limit was hit. */ - atomic64_t events_limit; + atomic64_t events[NR_PIDCG_EVENTS]; + atomic64_t events_local[NR_PIDCG_EVENTS]; }; static struct pids_cgroup *css_pids(struct cgroup_subsys_state *css) @@ -75,9 +84,7 @@ pids_css_alloc(struct cgroup_subsys_state *parent) if (!pids) return ERR_PTR(-ENOMEM); - atomic64_set(&pids->counter, 0); atomic64_set(&pids->limit, PIDS_MAX); - atomic64_set(&pids->events_limit, 0); return &pids->css; } @@ -150,12 +157,13 @@ static void pids_charge(struct pids_cgroup *pids, int num) * pids_try_charge - hierarchically try to charge the pid count * @pids: the pid cgroup state * @num: the number of pids to charge + * @fail: storage of pid cgroup causing the fail * * This function follows the set limit. It will fail if the charge would cause * the new value to exceed the hierarchical limit. Returns 0 if the charge * succeeded, otherwise -EAGAIN. */ -static int pids_try_charge(struct pids_cgroup *pids, int num) +static int pids_try_charge(struct pids_cgroup *pids, int num, struct pids_cgroup **fail) { struct pids_cgroup *p, *q; @@ -168,9 +176,10 @@ static int pids_try_charge(struct pids_cgroup *pids, int num) * p->limit is %PIDS_MAX then we know that this test will never * fail. */ - if (new > limit) + if (new > limit) { + *fail = p; goto revert; - + } /* * Not technically accurate if we go over limit somewhere up * the hierarchy, but that's tolerable for the watermark. @@ -231,44 +240,54 @@ static void pids_cancel_attach(struct cgroup_taskset *tset) } } +static void pids_event(struct pids_cgroup *pids_forking, + struct pids_cgroup *pids_over_limit) +{ + struct pids_cgroup *p = pids_forking; + + /* Only log the first time limit is hit. */ + if (atomic64_inc_return(&p->events_local[PIDCG_FORKFAIL]) == 1) { + pr_info("cgroup: fork rejected by pids controller in "); + pr_cont_cgroup_path(p->css.cgroup); + pr_cont("\n"); + } + if (!cgroup_subsys_on_dfl(pids_cgrp_subsys) || + cgrp_dfl_root.flags & CGRP_ROOT_PIDS_LOCAL_EVENTS) { + cgroup_file_notify(&p->events_local_file); + return; + } + + atomic64_inc(&pids_over_limit->events_local[PIDCG_MAX]); + cgroup_file_notify(&pids_over_limit->events_local_file); + + for (p = pids_over_limit; parent_pids(p); p = parent_pids(p)) { + atomic64_inc(&p->events[PIDCG_MAX]); + cgroup_file_notify(&p->events_file); + } +} + /* * task_css_check(true) in pids_can_fork() and pids_cancel_fork() relies * on cgroup_threadgroup_change_begin() held by the copy_process(). */ static int pids_can_fork(struct task_struct *task, struct css_set *cset) { - struct cgroup_subsys_state *css; - struct pids_cgroup *pids; + struct pids_cgroup *pids, *pids_over_limit; int err; - if (cset) - css = cset->subsys[pids_cgrp_id]; - else - css = task_css_check(current, pids_cgrp_id, true); - pids = css_pids(css); - err = pids_try_charge(pids, 1); - if (err) { - /* Only log the first time events_limit is incremented. */ - if (atomic64_inc_return(&pids->events_limit) == 1) { - pr_info("cgroup: fork rejected by pids controller in "); - pr_cont_cgroup_path(css->cgroup); - pr_cont("\n"); - } - cgroup_file_notify(&pids->events_file); - } + pids = css_pids(cset->subsys[pids_cgrp_id]); + err = pids_try_charge(pids, 1, &pids_over_limit); + if (err) + pids_event(pids, pids_over_limit); + return err; } static void pids_cancel_fork(struct task_struct *task, struct css_set *cset) { - struct cgroup_subsys_state *css; struct pids_cgroup *pids; - if (cset) - css = cset->subsys[pids_cgrp_id]; - else - css = task_css_check(current, pids_cgrp_id, true); - pids = css_pids(css); + pids = css_pids(cset->subsys[pids_cgrp_id]); pids_uncharge(pids, 1); } @@ -339,11 +358,32 @@ static s64 pids_peak_read(struct cgroup_subsys_state *css, return READ_ONCE(pids->watermark); } -static int pids_events_show(struct seq_file *sf, void *v) +static int __pids_events_show(struct seq_file *sf, bool local) { struct pids_cgroup *pids = css_pids(seq_css(sf)); + enum pidcg_event pe = PIDCG_MAX; + atomic64_t *events; + + if (!cgroup_subsys_on_dfl(pids_cgrp_subsys) || + cgrp_dfl_root.flags & CGRP_ROOT_PIDS_LOCAL_EVENTS) { + pe = PIDCG_FORKFAIL; + local = true; + } + events = local ? pids->events_local : pids->events; - seq_printf(sf, "max %lld\n", (s64)atomic64_read(&pids->events_limit)); + seq_printf(sf, "max %lld\n", (s64)atomic64_read(&events[pe])); + return 0; +} + +static int pids_events_show(struct seq_file *sf, void *v) +{ + __pids_events_show(sf, false); + return 0; +} + +static int pids_events_local_show(struct seq_file *sf, void *v) +{ + __pids_events_show(sf, true); return 0; } @@ -370,9 +410,42 @@ static struct cftype pids_files[] = { .file_offset = offsetof(struct pids_cgroup, events_file), .flags = CFTYPE_NOT_ON_ROOT, }, + { + .name = "events.local", + .seq_show = pids_events_local_show, + .file_offset = offsetof(struct pids_cgroup, events_local_file), + .flags = CFTYPE_NOT_ON_ROOT, + }, + { } /* terminate */ +}; + +static struct cftype pids_files_legacy[] = { + { + .name = "max", + .write = pids_max_write, + .seq_show = pids_max_show, + .flags = CFTYPE_NOT_ON_ROOT, + }, + { + .name = "current", + .read_s64 = pids_current_read, + .flags = CFTYPE_NOT_ON_ROOT, + }, + { + .name = "peak", + .flags = CFTYPE_NOT_ON_ROOT, + .read_s64 = pids_peak_read, + }, + { + .name = "events", + .seq_show = pids_events_show, + .file_offset = offsetof(struct pids_cgroup, events_file), + .flags = CFTYPE_NOT_ON_ROOT, + }, { } /* terminate */ }; + struct cgroup_subsys pids_cgrp_subsys = { .css_alloc = pids_css_alloc, .css_free = pids_css_free, @@ -381,7 +454,7 @@ struct cgroup_subsys pids_cgrp_subsys = { .can_fork = pids_can_fork, .cancel_fork = pids_cancel_fork, .release = pids_release, - .legacy_cftypes = pids_files, + .legacy_cftypes = pids_files_legacy, .dfl_cftypes = pids_files, .threaded = true, }; diff --git a/kernel/cgroup/rstat.c b/kernel/cgroup/rstat.c index 07e2284bb499..ce4752ab9e09 100644 --- a/kernel/cgroup/rstat.c +++ b/kernel/cgroup/rstat.c @@ -7,48 +7,150 @@ #include <linux/btf.h> #include <linux/btf_ids.h> -static DEFINE_SPINLOCK(cgroup_rstat_lock); -static DEFINE_PER_CPU(raw_spinlock_t, cgroup_rstat_cpu_lock); +#include <trace/events/cgroup.h> + +static DEFINE_SPINLOCK(rstat_base_lock); +static DEFINE_PER_CPU(raw_spinlock_t, rstat_base_cpu_lock); static void cgroup_base_stat_flush(struct cgroup *cgrp, int cpu); -static struct cgroup_rstat_cpu *cgroup_rstat_cpu(struct cgroup *cgrp, int cpu) +/* + * Determines whether a given css can participate in rstat. + * css's that are cgroup::self use rstat for base stats. + * Other css's associated with a subsystem use rstat only when + * they define the ss->css_rstat_flush callback. + */ +static inline bool css_uses_rstat(struct cgroup_subsys_state *css) +{ + return css_is_self(css) || css->ss->css_rstat_flush != NULL; +} + +static struct css_rstat_cpu *css_rstat_cpu( + struct cgroup_subsys_state *css, int cpu) +{ + return per_cpu_ptr(css->rstat_cpu, cpu); +} + +static struct cgroup_rstat_base_cpu *cgroup_rstat_base_cpu( + struct cgroup *cgrp, int cpu) +{ + return per_cpu_ptr(cgrp->rstat_base_cpu, cpu); +} + +static spinlock_t *ss_rstat_lock(struct cgroup_subsys *ss) +{ + if (ss) + return &ss->rstat_ss_lock; + + return &rstat_base_lock; +} + +static raw_spinlock_t *ss_rstat_cpu_lock(struct cgroup_subsys *ss, int cpu) +{ + if (ss) + return per_cpu_ptr(ss->rstat_ss_cpu_lock, cpu); + + return per_cpu_ptr(&rstat_base_cpu_lock, cpu); +} + +/* + * Helper functions for rstat per CPU locks. + * + * This makes it easier to diagnose locking issues and contention in + * production environments. The parameter @fast_path determine the + * tracepoints being added, allowing us to diagnose "flush" related + * operations without handling high-frequency fast-path "update" events. + */ +static __always_inline +unsigned long _css_rstat_cpu_lock(struct cgroup_subsys_state *css, int cpu, + const bool fast_path) { - return per_cpu_ptr(cgrp->rstat_cpu, cpu); + struct cgroup *cgrp = css->cgroup; + raw_spinlock_t *cpu_lock; + unsigned long flags; + bool contended; + + /* + * The _irqsave() is needed because the locks used for flushing are + * spinlock_t which is a sleeping lock on PREEMPT_RT. Acquiring this lock + * with the _irq() suffix only disables interrupts on a non-PREEMPT_RT + * kernel. The raw_spinlock_t below disables interrupts on both + * configurations. The _irqsave() ensures that interrupts are always + * disabled and later restored. + */ + cpu_lock = ss_rstat_cpu_lock(css->ss, cpu); + contended = !raw_spin_trylock_irqsave(cpu_lock, flags); + if (contended) { + if (fast_path) + trace_cgroup_rstat_cpu_lock_contended_fastpath(cgrp, cpu, contended); + else + trace_cgroup_rstat_cpu_lock_contended(cgrp, cpu, contended); + + raw_spin_lock_irqsave(cpu_lock, flags); + } + + if (fast_path) + trace_cgroup_rstat_cpu_locked_fastpath(cgrp, cpu, contended); + else + trace_cgroup_rstat_cpu_locked(cgrp, cpu, contended); + + return flags; +} + +static __always_inline +void _css_rstat_cpu_unlock(struct cgroup_subsys_state *css, int cpu, + unsigned long flags, const bool fast_path) +{ + struct cgroup *cgrp = css->cgroup; + raw_spinlock_t *cpu_lock; + + if (fast_path) + trace_cgroup_rstat_cpu_unlock_fastpath(cgrp, cpu, false); + else + trace_cgroup_rstat_cpu_unlock(cgrp, cpu, false); + + cpu_lock = ss_rstat_cpu_lock(css->ss, cpu); + raw_spin_unlock_irqrestore(cpu_lock, flags); } /** - * cgroup_rstat_updated - keep track of updated rstat_cpu - * @cgrp: target cgroup + * css_rstat_updated - keep track of updated rstat_cpu + * @css: target cgroup subsystem state * @cpu: cpu on which rstat_cpu was updated * - * @cgrp's rstat_cpu on @cpu was updated. Put it on the parent's matching - * rstat_cpu->updated_children list. See the comment on top of - * cgroup_rstat_cpu definition for details. + * @css's rstat_cpu on @cpu was updated. Put it on the parent's matching + * rstat_cpu->updated_children list. See the comment on top of + * css_rstat_cpu definition for details. */ -__bpf_kfunc void cgroup_rstat_updated(struct cgroup *cgrp, int cpu) +__bpf_kfunc void css_rstat_updated(struct cgroup_subsys_state *css, int cpu) { - raw_spinlock_t *cpu_lock = per_cpu_ptr(&cgroup_rstat_cpu_lock, cpu); unsigned long flags; /* + * Since bpf programs can call this function, prevent access to + * uninitialized rstat pointers. + */ + if (!css_uses_rstat(css)) + return; + + /* * Speculative already-on-list test. This may race leading to * temporary inaccuracies, which is fine. * * Because @parent's updated_children is terminated with @parent - * instead of NULL, we can tell whether @cgrp is on the list by + * instead of NULL, we can tell whether @css is on the list by * testing the next pointer for NULL. */ - if (data_race(cgroup_rstat_cpu(cgrp, cpu)->updated_next)) + if (data_race(css_rstat_cpu(css, cpu)->updated_next)) return; - raw_spin_lock_irqsave(cpu_lock, flags); + flags = _css_rstat_cpu_lock(css, cpu, true); - /* put @cgrp and all ancestors on the corresponding updated lists */ + /* put @css and all ancestors on the corresponding updated lists */ while (true) { - struct cgroup_rstat_cpu *rstatc = cgroup_rstat_cpu(cgrp, cpu); - struct cgroup *parent = cgroup_parent(cgrp); - struct cgroup_rstat_cpu *prstatc; + struct css_rstat_cpu *rstatc = css_rstat_cpu(css, cpu); + struct cgroup_subsys_state *parent = css->parent; + struct css_rstat_cpu *prstatc; /* * Both additions and removals are bottom-up. If a cgroup @@ -59,53 +161,78 @@ __bpf_kfunc void cgroup_rstat_updated(struct cgroup *cgrp, int cpu) /* Root has no parent to link it to, but mark it busy */ if (!parent) { - rstatc->updated_next = cgrp; + rstatc->updated_next = css; break; } - prstatc = cgroup_rstat_cpu(parent, cpu); + prstatc = css_rstat_cpu(parent, cpu); rstatc->updated_next = prstatc->updated_children; - prstatc->updated_children = cgrp; + prstatc->updated_children = css; - cgrp = parent; + css = parent; } - raw_spin_unlock_irqrestore(cpu_lock, flags); + _css_rstat_cpu_unlock(css, cpu, flags, true); } /** - * cgroup_rstat_push_children - push children cgroups into the given list + * css_rstat_push_children - push children css's into the given list * @head: current head of the list (= subtree root) * @child: first child of the root * @cpu: target cpu - * Return: A new singly linked list of cgroups to be flush + * Return: A new singly linked list of css's to be flushed * - * Iteratively traverse down the cgroup_rstat_cpu updated tree level by + * Iteratively traverse down the css_rstat_cpu updated tree level by * level and push all the parents first before their next level children - * into a singly linked list built from the tail backward like "pushing" - * cgroups into a stack. The root is pushed by the caller. + * into a singly linked list via the rstat_flush_next pointer built from the + * tail backward like "pushing" css's into a stack. The root is pushed by + * the caller. */ -static struct cgroup *cgroup_rstat_push_children(struct cgroup *head, - struct cgroup *child, int cpu) +static struct cgroup_subsys_state *css_rstat_push_children( + struct cgroup_subsys_state *head, + struct cgroup_subsys_state *child, int cpu) { - struct cgroup *chead = child; /* Head of child cgroup level */ - struct cgroup *ghead = NULL; /* Head of grandchild cgroup level */ - struct cgroup *parent, *grandchild; - struct cgroup_rstat_cpu *crstatc; + struct cgroup_subsys_state *cnext = child; /* Next head of child css level */ + struct cgroup_subsys_state *ghead = NULL; /* Head of grandchild css level */ + struct cgroup_subsys_state *parent, *grandchild; + struct css_rstat_cpu *crstatc; child->rstat_flush_next = NULL; + /* + * The subsystem rstat lock must be held for the whole duration from + * here as the rstat_flush_next list is being constructed to when + * it is consumed later in css_rstat_flush(). + */ + lockdep_assert_held(ss_rstat_lock(head->ss)); + + /* + * Notation: -> updated_next pointer + * => rstat_flush_next pointer + * + * Assuming the following sample updated_children lists: + * P: C1 -> C2 -> P + * C1: G11 -> G12 -> C1 + * C2: G21 -> G22 -> C2 + * + * After 1st iteration: + * head => C2 => C1 => NULL + * ghead => G21 => G11 => NULL + * + * After 2nd iteration: + * head => G12 => G11 => G22 => G21 => C2 => C1 => NULL + */ next_level: - while (chead) { - child = chead; - chead = child->rstat_flush_next; - parent = cgroup_parent(child); + while (cnext) { + child = cnext; + cnext = child->rstat_flush_next; + parent = child->parent; - /* updated_next is parent cgroup terminated */ + /* updated_next is parent cgroup terminated if !NULL */ while (child != parent) { child->rstat_flush_next = head; head = child; - crstatc = cgroup_rstat_cpu(child, cpu); + crstatc = css_rstat_cpu(child, cpu); grandchild = crstatc->updated_children; if (grandchild != child) { /* Push the grand child to the next level */ @@ -119,7 +246,7 @@ next_level: } if (ghead) { - chead = ghead; + cnext = ghead; ghead = NULL; goto next_level; } @@ -127,39 +254,31 @@ next_level: } /** - * cgroup_rstat_updated_list - return a list of updated cgroups to be flushed - * @root: root of the cgroup subtree to traverse + * css_rstat_updated_list - build a list of updated css's to be flushed + * @root: root of the css subtree to traverse * @cpu: target cpu - * Return: A singly linked list of cgroups to be flushed + * Return: A singly linked list of css's to be flushed * * Walks the updated rstat_cpu tree on @cpu from @root. During traversal, - * each returned cgroup is unlinked from the updated tree. + * each returned css is unlinked from the updated tree. * * The only ordering guarantee is that, for a parent and a child pair * covered by a given traversal, the child is before its parent in * the list. * * Note that updated_children is self terminated and points to a list of - * child cgroups if not empty. Whereas updated_next is like a sibling link - * within the children list and terminated by the parent cgroup. An exception - * here is the cgroup root whose updated_next can be self terminated. + * child css's if not empty. Whereas updated_next is like a sibling link + * within the children list and terminated by the parent css. An exception + * here is the css root whose updated_next can be self terminated. */ -static struct cgroup *cgroup_rstat_updated_list(struct cgroup *root, int cpu) +static struct cgroup_subsys_state *css_rstat_updated_list( + struct cgroup_subsys_state *root, int cpu) { - raw_spinlock_t *cpu_lock = per_cpu_ptr(&cgroup_rstat_cpu_lock, cpu); - struct cgroup_rstat_cpu *rstatc = cgroup_rstat_cpu(root, cpu); - struct cgroup *head = NULL, *parent, *child; + struct css_rstat_cpu *rstatc = css_rstat_cpu(root, cpu); + struct cgroup_subsys_state *head = NULL, *parent, *child; unsigned long flags; - /* - * The _irqsave() is needed because cgroup_rstat_lock is - * spinlock_t which is a sleeping lock on PREEMPT_RT. Acquiring - * this lock with the _irq() suffix only disables interrupts on - * a non-PREEMPT_RT kernel. The raw_spinlock_t below disables - * interrupts on both configurations. The _irqsave() ensures - * that interrupts are always disabled and later restored. - */ - raw_spin_lock_irqsave(cpu_lock, flags); + flags = _css_rstat_cpu_lock(root, cpu, false); /* Return NULL if this subtree is not on-list */ if (!rstatc->updated_next) @@ -169,17 +288,17 @@ static struct cgroup *cgroup_rstat_updated_list(struct cgroup *root, int cpu) * Unlink @root from its parent. As the updated_children list is * singly linked, we have to walk it to find the removal point. */ - parent = cgroup_parent(root); + parent = root->parent; if (parent) { - struct cgroup_rstat_cpu *prstatc; - struct cgroup **nextp; + struct css_rstat_cpu *prstatc; + struct cgroup_subsys_state **nextp; - prstatc = cgroup_rstat_cpu(parent, cpu); + prstatc = css_rstat_cpu(parent, cpu); nextp = &prstatc->updated_children; while (*nextp != root) { - struct cgroup_rstat_cpu *nrstatc; + struct css_rstat_cpu *nrstatc; - nrstatc = cgroup_rstat_cpu(*nextp, cpu); + nrstatc = css_rstat_cpu(*nextp, cpu); WARN_ON_ONCE(*nextp == parent); nextp = &nrstatc->updated_next; } @@ -194,16 +313,16 @@ static struct cgroup *cgroup_rstat_updated_list(struct cgroup *root, int cpu) child = rstatc->updated_children; rstatc->updated_children = root; if (child != root) - head = cgroup_rstat_push_children(head, child, cpu); + head = css_rstat_push_children(head, child, cpu); unlock_ret: - raw_spin_unlock_irqrestore(cpu_lock, flags); + _css_rstat_cpu_unlock(root, cpu, flags, false); return head; } /* * A hook for bpf stat collectors to attach to and flush their stats. - * Together with providing bpf kfuncs for cgroup_rstat_updated() and - * cgroup_rstat_flush(), this enables a complete workflow where bpf progs that + * Together with providing bpf kfuncs for css_rstat_updated() and + * css_rstat_flush(), this enables a complete workflow where bpf progs that * collect cgroup stats can integrate with rstat for efficient flushing. * * A static noinline declaration here could cause the compiler to optimize away @@ -222,135 +341,195 @@ __weak noinline void bpf_rstat_flush(struct cgroup *cgrp, __bpf_hook_end(); -/* see cgroup_rstat_flush() */ -static void cgroup_rstat_flush_locked(struct cgroup *cgrp) - __releases(&cgroup_rstat_lock) __acquires(&cgroup_rstat_lock) +/* + * Helper functions for locking. + * + * This makes it easier to diagnose locking issues and contention in + * production environments. The parameter @cpu_in_loop indicate lock + * was released and re-taken when collection data from the CPUs. The + * value -1 is used when obtaining the main lock else this is the CPU + * number processed last. + */ +static inline void __css_rstat_lock(struct cgroup_subsys_state *css, + int cpu_in_loop) + __acquires(ss_rstat_lock(css->ss)) { - int cpu; - - lockdep_assert_held(&cgroup_rstat_lock); - - for_each_possible_cpu(cpu) { - struct cgroup *pos = cgroup_rstat_updated_list(cgrp, cpu); - - for (; pos; pos = pos->rstat_flush_next) { - struct cgroup_subsys_state *css; - - cgroup_base_stat_flush(pos, cpu); - bpf_rstat_flush(pos, cgroup_parent(pos), cpu); + struct cgroup *cgrp = css->cgroup; + spinlock_t *lock; + bool contended; + + lock = ss_rstat_lock(css->ss); + contended = !spin_trylock_irq(lock); + if (contended) { + trace_cgroup_rstat_lock_contended(cgrp, cpu_in_loop, contended); + spin_lock_irq(lock); + } + trace_cgroup_rstat_locked(cgrp, cpu_in_loop, contended); +} - rcu_read_lock(); - list_for_each_entry_rcu(css, &pos->rstat_css_list, - rstat_css_node) - css->ss->css_rstat_flush(css, cpu); - rcu_read_unlock(); - } +static inline void __css_rstat_unlock(struct cgroup_subsys_state *css, + int cpu_in_loop) + __releases(ss_rstat_lock(css->ss)) +{ + struct cgroup *cgrp = css->cgroup; + spinlock_t *lock; - /* play nice and yield if necessary */ - if (need_resched() || spin_needbreak(&cgroup_rstat_lock)) { - spin_unlock_irq(&cgroup_rstat_lock); - if (!cond_resched()) - cpu_relax(); - spin_lock_irq(&cgroup_rstat_lock); - } - } + lock = ss_rstat_lock(css->ss); + trace_cgroup_rstat_unlock(cgrp, cpu_in_loop, false); + spin_unlock_irq(lock); } /** - * cgroup_rstat_flush - flush stats in @cgrp's subtree - * @cgrp: target cgroup + * css_rstat_flush - flush stats in @css's rstat subtree + * @css: target cgroup subsystem state * - * Collect all per-cpu stats in @cgrp's subtree into the global counters - * and propagate them upwards. After this function returns, all cgroups in - * the subtree have up-to-date ->stat. + * Collect all per-cpu stats in @css's subtree into the global counters + * and propagate them upwards. After this function returns, all rstat + * nodes in the subtree have up-to-date ->stat. * - * This also gets all cgroups in the subtree including @cgrp off the + * This also gets all rstat nodes in the subtree including @css off the * ->updated_children lists. * * This function may block. */ -__bpf_kfunc void cgroup_rstat_flush(struct cgroup *cgrp) +__bpf_kfunc void css_rstat_flush(struct cgroup_subsys_state *css) { - might_sleep(); + int cpu; + bool is_self = css_is_self(css); - spin_lock_irq(&cgroup_rstat_lock); - cgroup_rstat_flush_locked(cgrp); - spin_unlock_irq(&cgroup_rstat_lock); -} + /* + * Since bpf programs can call this function, prevent access to + * uninitialized rstat pointers. + */ + if (!css_uses_rstat(css)) + return; -/** - * cgroup_rstat_flush_hold - flush stats in @cgrp's subtree and hold - * @cgrp: target cgroup - * - * Flush stats in @cgrp's subtree and prevent further flushes. Must be - * paired with cgroup_rstat_flush_release(). - * - * This function may block. - */ -void cgroup_rstat_flush_hold(struct cgroup *cgrp) - __acquires(&cgroup_rstat_lock) -{ might_sleep(); - spin_lock_irq(&cgroup_rstat_lock); - cgroup_rstat_flush_locked(cgrp); -} + for_each_possible_cpu(cpu) { + struct cgroup_subsys_state *pos; -/** - * cgroup_rstat_flush_release - release cgroup_rstat_flush_hold() - */ -void cgroup_rstat_flush_release(void) - __releases(&cgroup_rstat_lock) -{ - spin_unlock_irq(&cgroup_rstat_lock); + /* Reacquire for each CPU to avoid disabling IRQs too long */ + __css_rstat_lock(css, cpu); + pos = css_rstat_updated_list(css, cpu); + for (; pos; pos = pos->rstat_flush_next) { + if (is_self) { + cgroup_base_stat_flush(pos->cgroup, cpu); + bpf_rstat_flush(pos->cgroup, + cgroup_parent(pos->cgroup), cpu); + } else + pos->ss->css_rstat_flush(pos, cpu); + } + __css_rstat_unlock(css, cpu); + if (!cond_resched()) + cpu_relax(); + } } -int cgroup_rstat_init(struct cgroup *cgrp) +int css_rstat_init(struct cgroup_subsys_state *css) { + struct cgroup *cgrp = css->cgroup; int cpu; + bool is_self = css_is_self(css); + + if (is_self) { + /* the root cgrp has rstat_base_cpu preallocated */ + if (!cgrp->rstat_base_cpu) { + cgrp->rstat_base_cpu = alloc_percpu(struct cgroup_rstat_base_cpu); + if (!cgrp->rstat_base_cpu) + return -ENOMEM; + } + } else if (css->ss->css_rstat_flush == NULL) + return 0; + + /* the root cgrp's self css has rstat_cpu preallocated */ + if (!css->rstat_cpu) { + css->rstat_cpu = alloc_percpu(struct css_rstat_cpu); + if (!css->rstat_cpu) { + if (is_self) + free_percpu(cgrp->rstat_base_cpu); - /* the root cgrp has rstat_cpu preallocated */ - if (!cgrp->rstat_cpu) { - cgrp->rstat_cpu = alloc_percpu(struct cgroup_rstat_cpu); - if (!cgrp->rstat_cpu) return -ENOMEM; + } } /* ->updated_children list is self terminated */ for_each_possible_cpu(cpu) { - struct cgroup_rstat_cpu *rstatc = cgroup_rstat_cpu(cgrp, cpu); + struct css_rstat_cpu *rstatc = css_rstat_cpu(css, cpu); + + rstatc->updated_children = css; + + if (is_self) { + struct cgroup_rstat_base_cpu *rstatbc; - rstatc->updated_children = cgrp; - u64_stats_init(&rstatc->bsync); + rstatbc = cgroup_rstat_base_cpu(cgrp, cpu); + u64_stats_init(&rstatbc->bsync); + } } return 0; } -void cgroup_rstat_exit(struct cgroup *cgrp) +void css_rstat_exit(struct cgroup_subsys_state *css) { int cpu; - cgroup_rstat_flush(cgrp); + if (!css_uses_rstat(css)) + return; + + css_rstat_flush(css); /* sanity check */ for_each_possible_cpu(cpu) { - struct cgroup_rstat_cpu *rstatc = cgroup_rstat_cpu(cgrp, cpu); + struct css_rstat_cpu *rstatc = css_rstat_cpu(css, cpu); - if (WARN_ON_ONCE(rstatc->updated_children != cgrp) || + if (WARN_ON_ONCE(rstatc->updated_children != css) || WARN_ON_ONCE(rstatc->updated_next)) return; } - free_percpu(cgrp->rstat_cpu); - cgrp->rstat_cpu = NULL; + if (css_is_self(css)) { + struct cgroup *cgrp = css->cgroup; + + free_percpu(cgrp->rstat_base_cpu); + cgrp->rstat_base_cpu = NULL; + } + + free_percpu(css->rstat_cpu); + css->rstat_cpu = NULL; } -void __init cgroup_rstat_boot(void) +/** + * ss_rstat_init - subsystem-specific rstat initialization + * @ss: target subsystem + * + * If @ss is NULL, the static locks associated with the base stats + * are initialized. If @ss is non-NULL, the subsystem-specific locks + * are initialized. + */ +int __init ss_rstat_init(struct cgroup_subsys *ss) { int cpu; +#ifdef CONFIG_SMP + /* + * On uniprocessor machines, arch_spinlock_t is defined as an empty + * struct. Avoid allocating a size of zero by having this block + * excluded in this case. It's acceptable to leave the subsystem locks + * unitialized since the associated lock functions are no-ops in the + * non-smp case. + */ + if (ss) { + ss->rstat_ss_cpu_lock = alloc_percpu(raw_spinlock_t); + if (!ss->rstat_ss_cpu_lock) + return -ENOMEM; + } +#endif + + spin_lock_init(ss_rstat_lock(ss)); for_each_possible_cpu(cpu) - raw_spin_lock_init(per_cpu_ptr(&cgroup_rstat_cpu_lock, cpu)); + raw_spin_lock_init(ss_rstat_cpu_lock(ss, cpu)); + + return 0; } /* @@ -366,6 +545,7 @@ static void cgroup_base_stat_add(struct cgroup_base_stat *dst_bstat, #ifdef CONFIG_SCHED_CORE dst_bstat->forceidle_sum += src_bstat->forceidle_sum; #endif + dst_bstat->ntime += src_bstat->ntime; } static void cgroup_base_stat_sub(struct cgroup_base_stat *dst_bstat, @@ -377,13 +557,14 @@ static void cgroup_base_stat_sub(struct cgroup_base_stat *dst_bstat, #ifdef CONFIG_SCHED_CORE dst_bstat->forceidle_sum -= src_bstat->forceidle_sum; #endif + dst_bstat->ntime -= src_bstat->ntime; } static void cgroup_base_stat_flush(struct cgroup *cgrp, int cpu) { - struct cgroup_rstat_cpu *rstatc = cgroup_rstat_cpu(cgrp, cpu); + struct cgroup_rstat_base_cpu *rstatbc = cgroup_rstat_base_cpu(cgrp, cpu); struct cgroup *parent = cgroup_parent(cgrp); - struct cgroup_rstat_cpu *prstatc; + struct cgroup_rstat_base_cpu *prstatbc; struct cgroup_base_stat delta; unsigned seq; @@ -393,15 +574,15 @@ static void cgroup_base_stat_flush(struct cgroup *cgrp, int cpu) /* fetch the current per-cpu values */ do { - seq = __u64_stats_fetch_begin(&rstatc->bsync); - delta = rstatc->bstat; - } while (__u64_stats_fetch_retry(&rstatc->bsync, seq)); + seq = __u64_stats_fetch_begin(&rstatbc->bsync); + delta = rstatbc->bstat; + } while (__u64_stats_fetch_retry(&rstatbc->bsync, seq)); /* propagate per-cpu delta to cgroup and per-cpu global statistics */ - cgroup_base_stat_sub(&delta, &rstatc->last_bstat); + cgroup_base_stat_sub(&delta, &rstatbc->last_bstat); cgroup_base_stat_add(&cgrp->bstat, &delta); - cgroup_base_stat_add(&rstatc->last_bstat, &delta); - cgroup_base_stat_add(&rstatc->subtree_bstat, &delta); + cgroup_base_stat_add(&rstatbc->last_bstat, &delta); + cgroup_base_stat_add(&rstatbc->subtree_bstat, &delta); /* propagate cgroup and per-cpu global delta to parent (unless that's root) */ if (cgroup_parent(parent)) { @@ -410,71 +591,73 @@ static void cgroup_base_stat_flush(struct cgroup *cgrp, int cpu) cgroup_base_stat_add(&parent->bstat, &delta); cgroup_base_stat_add(&cgrp->last_bstat, &delta); - delta = rstatc->subtree_bstat; - prstatc = cgroup_rstat_cpu(parent, cpu); - cgroup_base_stat_sub(&delta, &rstatc->last_subtree_bstat); - cgroup_base_stat_add(&prstatc->subtree_bstat, &delta); - cgroup_base_stat_add(&rstatc->last_subtree_bstat, &delta); + delta = rstatbc->subtree_bstat; + prstatbc = cgroup_rstat_base_cpu(parent, cpu); + cgroup_base_stat_sub(&delta, &rstatbc->last_subtree_bstat); + cgroup_base_stat_add(&prstatbc->subtree_bstat, &delta); + cgroup_base_stat_add(&rstatbc->last_subtree_bstat, &delta); } } -static struct cgroup_rstat_cpu * +static struct cgroup_rstat_base_cpu * cgroup_base_stat_cputime_account_begin(struct cgroup *cgrp, unsigned long *flags) { - struct cgroup_rstat_cpu *rstatc; + struct cgroup_rstat_base_cpu *rstatbc; - rstatc = get_cpu_ptr(cgrp->rstat_cpu); - *flags = u64_stats_update_begin_irqsave(&rstatc->bsync); - return rstatc; + rstatbc = get_cpu_ptr(cgrp->rstat_base_cpu); + *flags = u64_stats_update_begin_irqsave(&rstatbc->bsync); + return rstatbc; } static void cgroup_base_stat_cputime_account_end(struct cgroup *cgrp, - struct cgroup_rstat_cpu *rstatc, + struct cgroup_rstat_base_cpu *rstatbc, unsigned long flags) { - u64_stats_update_end_irqrestore(&rstatc->bsync, flags); - cgroup_rstat_updated(cgrp, smp_processor_id()); - put_cpu_ptr(rstatc); + u64_stats_update_end_irqrestore(&rstatbc->bsync, flags); + css_rstat_updated(&cgrp->self, smp_processor_id()); + put_cpu_ptr(rstatbc); } void __cgroup_account_cputime(struct cgroup *cgrp, u64 delta_exec) { - struct cgroup_rstat_cpu *rstatc; + struct cgroup_rstat_base_cpu *rstatbc; unsigned long flags; - rstatc = cgroup_base_stat_cputime_account_begin(cgrp, &flags); - rstatc->bstat.cputime.sum_exec_runtime += delta_exec; - cgroup_base_stat_cputime_account_end(cgrp, rstatc, flags); + rstatbc = cgroup_base_stat_cputime_account_begin(cgrp, &flags); + rstatbc->bstat.cputime.sum_exec_runtime += delta_exec; + cgroup_base_stat_cputime_account_end(cgrp, rstatbc, flags); } void __cgroup_account_cputime_field(struct cgroup *cgrp, enum cpu_usage_stat index, u64 delta_exec) { - struct cgroup_rstat_cpu *rstatc; + struct cgroup_rstat_base_cpu *rstatbc; unsigned long flags; - rstatc = cgroup_base_stat_cputime_account_begin(cgrp, &flags); + rstatbc = cgroup_base_stat_cputime_account_begin(cgrp, &flags); switch (index) { - case CPUTIME_USER: case CPUTIME_NICE: - rstatc->bstat.cputime.utime += delta_exec; + rstatbc->bstat.ntime += delta_exec; + fallthrough; + case CPUTIME_USER: + rstatbc->bstat.cputime.utime += delta_exec; break; case CPUTIME_SYSTEM: case CPUTIME_IRQ: case CPUTIME_SOFTIRQ: - rstatc->bstat.cputime.stime += delta_exec; + rstatbc->bstat.cputime.stime += delta_exec; break; #ifdef CONFIG_SCHED_CORE case CPUTIME_FORCEIDLE: - rstatc->bstat.forceidle_sum += delta_exec; + rstatbc->bstat.forceidle_sum += delta_exec; break; #endif default: break; } - cgroup_base_stat_cputime_account_end(cgrp, rstatc, flags); + cgroup_base_stat_cputime_account_end(cgrp, rstatbc, flags); } /* @@ -508,63 +691,62 @@ static void root_cgroup_cputime(struct cgroup_base_stat *bstat) cputime->sum_exec_runtime += user; cputime->sum_exec_runtime += sys; - cputime->sum_exec_runtime += cpustat[CPUTIME_STEAL]; #ifdef CONFIG_SCHED_CORE bstat->forceidle_sum += cpustat[CPUTIME_FORCEIDLE]; #endif + bstat->ntime += cpustat[CPUTIME_NICE]; } } + +static void cgroup_force_idle_show(struct seq_file *seq, struct cgroup_base_stat *bstat) +{ +#ifdef CONFIG_SCHED_CORE + u64 forceidle_time = bstat->forceidle_sum; + + do_div(forceidle_time, NSEC_PER_USEC); + seq_printf(seq, "core_sched.force_idle_usec %llu\n", forceidle_time); +#endif +} + void cgroup_base_stat_cputime_show(struct seq_file *seq) { struct cgroup *cgrp = seq_css(seq)->cgroup; - u64 usage, utime, stime; struct cgroup_base_stat bstat; -#ifdef CONFIG_SCHED_CORE - u64 forceidle_time; -#endif if (cgroup_parent(cgrp)) { - cgroup_rstat_flush_hold(cgrp); - usage = cgrp->bstat.cputime.sum_exec_runtime; + css_rstat_flush(&cgrp->self); + __css_rstat_lock(&cgrp->self, -1); + bstat = cgrp->bstat; cputime_adjust(&cgrp->bstat.cputime, &cgrp->prev_cputime, - &utime, &stime); -#ifdef CONFIG_SCHED_CORE - forceidle_time = cgrp->bstat.forceidle_sum; -#endif - cgroup_rstat_flush_release(); + &bstat.cputime.utime, &bstat.cputime.stime); + __css_rstat_unlock(&cgrp->self, -1); } else { root_cgroup_cputime(&bstat); - usage = bstat.cputime.sum_exec_runtime; - utime = bstat.cputime.utime; - stime = bstat.cputime.stime; -#ifdef CONFIG_SCHED_CORE - forceidle_time = bstat.forceidle_sum; -#endif } - do_div(usage, NSEC_PER_USEC); - do_div(utime, NSEC_PER_USEC); - do_div(stime, NSEC_PER_USEC); -#ifdef CONFIG_SCHED_CORE - do_div(forceidle_time, NSEC_PER_USEC); -#endif + do_div(bstat.cputime.sum_exec_runtime, NSEC_PER_USEC); + do_div(bstat.cputime.utime, NSEC_PER_USEC); + do_div(bstat.cputime.stime, NSEC_PER_USEC); + do_div(bstat.ntime, NSEC_PER_USEC); seq_printf(seq, "usage_usec %llu\n" - "user_usec %llu\n" - "system_usec %llu\n", - usage, utime, stime); - -#ifdef CONFIG_SCHED_CORE - seq_printf(seq, "core_sched.force_idle_usec %llu\n", forceidle_time); -#endif + "user_usec %llu\n" + "system_usec %llu\n" + "nice_usec %llu\n", + bstat.cputime.sum_exec_runtime, + bstat.cputime.utime, + bstat.cputime.stime, + bstat.ntime); + + cgroup_force_idle_show(seq, &bstat); } -/* Add bpf kfuncs for cgroup_rstat_updated() and cgroup_rstat_flush() */ +/* Add bpf kfuncs for css_rstat_updated() and css_rstat_flush() */ BTF_KFUNCS_START(bpf_rstat_kfunc_ids) -BTF_ID_FLAGS(func, cgroup_rstat_updated) -BTF_ID_FLAGS(func, cgroup_rstat_flush, KF_SLEEPABLE) +BTF_ID_FLAGS(func, css_rstat_updated) +BTF_ID_FLAGS(func, css_rstat_flush, KF_SLEEPABLE) BTF_KFUNCS_END(bpf_rstat_kfunc_ids) static const struct btf_kfunc_id_set bpf_rstat_kfunc_set = { |