hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/ICA1GK -------------------------------- Create a bpf iter target for the 'stat' interface, to which the bpf prog can attach. Signed-off-by: GONG Ruiqi <gongruiqi1@huawei.com> --- fs/proc/stat.c | 164 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 164 insertions(+) diff --git a/fs/proc/stat.c b/fs/proc/stat.c index 1d8699285e02..4757e5b1be38 100644 --- a/fs/proc/stat.c +++ b/fs/proc/stat.c @@ -18,6 +18,9 @@ #include <linux/bpf.h> #include <linux/btf.h> #include <linux/btf_ids.h> +#include <linux/pid_namespace.h> +#include <linux/cgroup.h> +#include <linux/cpuset.h> #endif #ifndef arch_irq_stat_cpu @@ -253,4 +256,165 @@ static int __init bpf_proc_stat_kfunc_init(void) &bpf_proc_stat_kfunc_set); } late_initcall(bpf_proc_stat_kfunc_init); + +struct stat_sum_data { + u64 user, nice, system, idle, iowait, irq, softirq, steal; + u64 guest, guest_nice; + u64 sum; + u64 sum_softirq; + unsigned int per_softirq_sums[NR_SOFTIRQS]; +}; + +struct stat_seq_priv { + cpumask_t allowed_mask; + struct cpuacct *cpuacct; + bool sum_printed; + struct task_struct *task; + struct seq_file seqf_pcpu; +}; + +static int seq_file_setup(struct seq_file *seq) +{ + seq->size = PAGE_SIZE << 3; + seq->buf = kvzalloc(seq->size, GFP_KERNEL); + if (!seq->buf) + return -ENOMEM; + return 0; +} + +static void seq_file_destroy(struct seq_file *seq) +{ + if (seq->buf) + kvfree(seq->buf); +} + +static void *bpf_c_start(struct seq_file *m, loff_t *pos) +{ + struct stat_seq_priv *priv = m->private; + struct task_struct *reaper = get_current_level1_reaper(); + + priv->task = reaper ?: current; + task_effective_cpumask(priv->task, &priv->allowed_mask); + priv->cpuacct = task_cpuacct(priv->task); + if (seq_file_setup(&priv->seqf_pcpu)) + return NULL; + + /* + * DO NOT use cpumask_first() here: sys_read may start from somewhere in + * the middle of the file, and *pos may contain a value from the last + * read. + */ + *pos = cpumask_next(*pos - 1, &priv->allowed_mask); + if ((*pos) < nr_cpu_ids) + // avoid 0, which will be treated as NULL + return (void *)(unsigned long)((*pos) + 1); + return NULL; +} + +static void *bpf_c_next(struct seq_file *m, void *v, loff_t *pos) +{ + struct stat_seq_priv *priv = m->private; + + *pos = cpumask_next(*pos, &priv->allowed_mask); + + if ((*pos) == nr_cpu_ids) { + if (!priv->sum_printed) + priv->sum_printed = true; + else { + ++*pos; // just to silence "did not updated position index" msg + return NULL; + } + } + + // avoid 0, which will be treated as NULL + return (void *)(unsigned long)((*pos) + 1); +} + +struct bpf_iter__stat { + __bpf_md_ptr(struct bpf_iter_meta *, meta); + u64 cpuid __aligned(8); + __bpf_md_ptr(struct cpuacct *, cpuacct); + u64 arch_irq_stat_cpu __aligned(8); + u64 arch_irq_stat __aligned(8); + bool print_all __aligned(8); + __bpf_md_ptr(struct seq_file *, seqf_pcpu); +}; + +static int bpf_show_stat(struct seq_file *m, void *v) +{ + struct stat_seq_priv *priv = m->private; + struct bpf_iter__stat ctx; + struct bpf_iter_meta meta; + struct bpf_prog *prog; + u64 cpuid = (unsigned long)v - 1; // decode '+ 1' + + meta.seq = m; + prog = bpf_iter_get_info(&meta, false); + if (!prog) + return show_stat(m, v); + + ctx.meta = &meta; + + ctx.cpuid = cpuid; + ctx.cpuacct = priv->cpuacct; + if (cpuid != nr_cpu_ids) + ctx.arch_irq_stat_cpu = arch_irq_stat_cpu(cpuid); + else + ctx.arch_irq_stat = arch_irq_stat(); + ctx.print_all = (cpuid == nr_cpu_ids); + ctx.seqf_pcpu = &priv->seqf_pcpu; + + return bpf_iter_run_prog(prog, &ctx); +} + +static void bpf_c_stop(struct seq_file *m, void *v) +{ + struct stat_seq_priv *priv = m->private; + + if (priv->task != current) + put_task_struct(priv->task); + seq_file_destroy(&priv->seqf_pcpu); +} + +const struct seq_operations bpf_stat_ops = { + .start = bpf_c_start, + .next = bpf_c_next, + .stop = bpf_c_stop, + .show = bpf_show_stat, +}; + +DEFINE_BPF_ITER_FUNC(stat, struct bpf_iter_meta *meta, + u64 cpuid, + struct cpuacct *cpuacct, + u64 arch_irq_stat_cpu, + u64 arch_irq_stat, + bool print_all, + struct seq_file *seqf_pcpu) + +BTF_ID_LIST(btf_stat_id) +BTF_ID(struct, cpuacct) + +static const struct bpf_iter_seq_info stat_seq_info = { + .seq_ops = &bpf_stat_ops, + .init_seq_private = NULL, + .fini_seq_private = NULL, + .seq_priv_size = sizeof(struct stat_seq_priv), +}; + +static struct bpf_iter_reg stat_reg_info = { + .target = "stat", + .ctx_arg_info_size = 1, + .ctx_arg_info = { + { offsetof(struct bpf_iter__stat, cpuacct), + PTR_TO_BTF_ID, }, + }, + .seq_info = &stat_seq_info, +}; + +static int __init stat_iter_init(void) +{ + stat_reg_info.ctx_arg_info[0].btf_id = btf_stat_id[0]; + return bpf_iter_reg_target(&stat_reg_info); +} +late_initcall(stat_iter_init); #endif /* CONFIG_BPF_RVI */ -- 2.25.1