Hi Andrii,

Thanks for taking a look.

On Thu, Jul 28, 2022 at 3:40 PM Andrii Nakryiko
[off-list ref] wrote:

On Fri, Jul 22, 2022 at 10:49 AM Yosry Ahmed [off-list ref] wrote:

quoted

[...]

quoted

+
+#define CGROUP_PATH(p, n) {.path = #p"/"#n, .name = #n}
+
+static struct {
+       const char *path, *name;
+       unsigned long long id;
+       int fd;
+} cgroups[] = {
+       CGROUP_PATH(/, test),
+       CGROUP_PATH(/test, child1),
+       CGROUP_PATH(/test, child2),
+       CGROUP_PATH(/test/child1, child1_1),
+       CGROUP_PATH(/test/child1, child1_2),
+       CGROUP_PATH(/test/child2, child2_1),
+       CGROUP_PATH(/test/child2, child2_2),

nit: why are these arguments not explicit string literals?...
CGROUP_PATH("/test/child1", "child1_1") explicitly shows that those
values are used as strings

No particular reason I think. String literals are good. Will fix in v6.

quoted

+};
+
+#define N_CGROUPS ARRAY_SIZE(cgroups)
+#define N_NON_LEAF_CGROUPS 3
+
+int root_cgroup_fd;
+bool mounted_bpffs;
+

static?

Yeah, we were careless about 'static' or 'inline'. I am going to go
over the code and mark functions/vars 'static' properly.

quoted

+static int read_from_file(const char *path, char *buf, size_t size)
+{
+       int fd, len;
+
+       fd = open(path, O_RDONLY);
+       if (fd < 0) {
+               log_err("Open %s", path);
+               return 1;
+       }
+       len = read(fd, buf, size);
+       if (len < 0)
+               log_err("Read %s", path);
+       else
+               buf[len] = 0;
+       close(fd);
+       return len < 0;
+}
+

[...]

quoted

+       /* Also dump stats for root */
+       err = setup_cgroup_iter(obj, root_cgroup_fd, CG_ROOT_NAME);
+       if (!ASSERT_OK(err, "setup_cgroup_iter"))
+               return err;
+
+       /* Attach rstat flusher */
+       link = bpf_program__attach(obj->progs.vmscan_flush);
+       if (!ASSERT_OK_PTR(link, "attach rstat"))
+               return libbpf_get_error(link);

this is dangerous, because ASSERT_OK_PTR might overwrite errno by the
time we get to libbpf_get_error() call. link is NULL and
libbpf_get_error() extracts error from errno. It's best to just return
fixed error code here, or otherwise you'd need to remember err before
ASSERT_OK_PTR() call.

Ack. We can just return a fixed error code here. Thanks.

quoted

+       obj->links.vmscan_flush = link;
+
+       /* Attach tracing programs that will calculate vmscan delays */
+       link = bpf_program__attach(obj->progs.vmscan_start);
+       if (!ASSERT_OK_PTR(obj, "attach raw_tracepoint"))
+               return libbpf_get_error(link);
+       obj->links.vmscan_start = link;
+
+       link = bpf_program__attach(obj->progs.vmscan_end);
+       if (!ASSERT_OK_PTR(obj, "attach raw_tracepoint"))
+               return libbpf_get_error(link);
+       obj->links.vmscan_end = link;
+
+       *skel = obj;
+       return 0;
+}
+
+void destroy_progs(struct cgroup_hierarchical_stats *skel)

static?

Ack.

quoted

+{
+       char path[128];
+       int i;
+
+       for (i = 0; i < N_CGROUPS; i++) {
+               /* Delete files in bpffs that cgroup_iters are pinned in */
+               snprintf(path, 128, "%s%s", BPFFS_VMSCAN,
+                        cgroups[i].name);
+               ASSERT_OK(remove(path), "remove cgroup_iter pin");
+       }
+
+       /* Delete root file in bpffs */
+       snprintf(path, 128, "%s%s", BPFFS_VMSCAN, CG_ROOT_NAME);
+       ASSERT_OK(remove(path), "remove cgroup_iter root pin");
+       cgroup_hierarchical_stats__destroy(skel);
+}
+
+void test_cgroup_hierarchical_stats(void)
+{
+       struct cgroup_hierarchical_stats *skel = NULL;
+
+       if (setup_hierarchy())
+               goto hierarchy_cleanup;
+       if (setup_progs(&skel))
+               goto cleanup;
+       if (induce_vmscan())
+               goto cleanup;
+       check_vmscan_stats();
+cleanup:
+       destroy_progs(skel);
+hierarchy_cleanup:
+       destroy_hierarchy();
+}

diff --git a/tools/testing/selftests/bpf/progs/cgroup_hierarchical_stats.c b/tools/testing/selftests/bpf/progs/cgroup_hierarchical_stats.c
new file mode 100644
index 000000000000..85a65a72482e
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/cgroup_hierarchical_stats.c

@@ -0,0 +1,239 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Functions to manage eBPF programs attached to cgroup subsystems
+ *
+ * Copyright 2022 Google LLC.
+ */
+#include "vmlinux.h"
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_tracing.h>
+
+char _license[] SEC("license") = "GPL";
+
+/*
+ * Start times are stored per-task, not per-cgroup, as multiple tasks in one
+ * cgroup can perform reclain concurrently.

typo: reclaim?

Ack. Will fix.

quoted

+ */
+struct {
+       __uint(type, BPF_MAP_TYPE_TASK_STORAGE);
+       __uint(map_flags, BPF_F_NO_PREALLOC);
+       __type(key, int);
+       __type(value, __u64);
+} vmscan_start_time SEC(".maps");
+

[...]

quoted

+static inline int create_vmscan_percpu_elem(__u64 cg_id, __u64 state)
+{
+       struct vmscan_percpu pcpu_init = {.state = state, .prev = 0};
+       int err;
+
+       err = bpf_map_update_elem(&pcpu_cgroup_vmscan_elapsed, &cg_id,
+                                 &pcpu_init, BPF_NOEXIST);
+       if (err) {
+               bpf_printk("failed to create pcpu entry for cgroup %llu: %d\n"
+                          , cg_id, err);
+               return 1;
+       }
+       return 0;
+}
+
+static inline int create_vmscan_elem(__u64 cg_id, __u64 state, __u64 pending)

all those inlines above are not necessary, they don't have to be
actually inlined, right?

No. They don't have to. Will fix this.

quoted

+{
+       struct vmscan init = {.state = state, .pending = pending};
+       int err;
+
+       err = bpf_map_update_elem(&cgroup_vmscan_elapsed, &cg_id,
+                                 &init, BPF_NOEXIST);
+       if (err) {
+               bpf_printk("failed to create entry for cgroup %llu: %d\n"
+                          , cg_id, err);
+               return 1;
+       }
+       return 0;
+}
+
+SEC("tp_btf/mm_vmscan_memcg_reclaim_begin")
+int BPF_PROG(vmscan_start, int order, gfp_t gfp_flags)
+{
+       struct task_struct *task = bpf_get_current_task_btf();
+       __u64 *start_time_ptr;
+
+       start_time_ptr = bpf_task_storage_get(&vmscan_start_time, task, 0,
+                                         BPF_LOCAL_STORAGE_GET_F_CREATE);
+       if (!start_time_ptr) {
+               bpf_printk("error retrieving storage\n");

does user-space part read these trace_printk messages? If not, let's
remove them from the test

No. I will remove them in v6.

quoted

+               return 0;
+       }
+
+       *start_time_ptr = bpf_ktime_get_ns();
+       return 0;
+}
+

[...]