#include "util/target.h"
#include "util/callchain.h"
#include "util/lock-contention.h"
+#include "util/bpf_skel/lock_data.h"
#include <subcmd/pager.h>
#include <subcmd/parse-options.h>
static int stack_skip = CONTENTION_STACK_SKIP;
static int print_nr_entries = INT_MAX / 2;
-static enum {
- LOCK_AGGR_ADDR,
- LOCK_AGGR_TASK,
- LOCK_AGGR_CALLER,
-} aggr_mode = LOCK_AGGR_ADDR;
+static enum lock_aggr_mode aggr_mode = LOCK_AGGR_ADDR;
static struct thread_stat *thread_stat_find(u32 tid)
{
.map_nr_entries = bpf_map_entries,
.max_stack = max_stack_depth,
.stack_skip = stack_skip,
+ .aggr_mode = show_thread_stats ? LOCK_AGGR_TASK : LOCK_AGGR_CALLER,
};
session = perf_session__new(use_bpf ? NULL : &data, &eops);
if (select_key(true))
goto out_delete;
- if (show_thread_stats)
- aggr_mode = LOCK_AGGR_TASK;
- else
- aggr_mode = LOCK_AGGR_CALLER;
-
if (use_bpf) {
lock_contention_start();
if (argc)
#include "util/map.h"
#include "util/symbol.h"
#include "util/target.h"
+#include "util/thread.h"
#include "util/thread_map.h"
#include "util/lock-contention.h"
#include <linux/zalloc.h>
}
bpf_map__set_value_size(skel->maps.stacks, con->max_stack * sizeof(u64));
- bpf_map__set_max_entries(skel->maps.stacks, con->map_nr_entries);
bpf_map__set_max_entries(skel->maps.lock_stat, con->map_nr_entries);
bpf_map__set_max_entries(skel->maps.tstamp, con->map_nr_entries);
+ if (con->aggr_mode == LOCK_AGGR_TASK) {
+ bpf_map__set_max_entries(skel->maps.task_data, con->map_nr_entries);
+ bpf_map__set_max_entries(skel->maps.stacks, 1);
+ } else {
+ bpf_map__set_max_entries(skel->maps.task_data, 1);
+ bpf_map__set_max_entries(skel->maps.stacks, con->map_nr_entries);
+ }
+
if (target__has_cpu(target))
ncpus = perf_cpu_map__nr(evlist->core.user_requested_cpus);
if (target__has_task(target))
bpf_map_update_elem(fd, &pid, &val, BPF_ANY);
}
+ /* these don't work well if in the rodata section */
skel->bss->stack_skip = con->stack_skip;
+ skel->bss->aggr_mode = con->aggr_mode;
lock_contention_bpf__attach(skel);
return 0;
int lock_contention_read(struct lock_contention *con)
{
- int fd, stack, err = 0;
+ int fd, stack, task_fd, err = 0;
struct contention_key *prev_key, key;
struct contention_data data = {};
struct lock_stat *st = NULL;
fd = bpf_map__fd(skel->maps.lock_stat);
stack = bpf_map__fd(skel->maps.stacks);
+ task_fd = bpf_map__fd(skel->maps.task_data);
con->lost = skel->bss->lost;
if (stack_trace == NULL)
return -1;
+ if (con->aggr_mode == LOCK_AGGR_TASK) {
+ struct thread *idle = __machine__findnew_thread(machine,
+ /*pid=*/0,
+ /*tid=*/0);
+ thread__set_comm(idle, "swapper", /*timestamp=*/0);
+ }
+
prev_key = NULL;
while (!bpf_map_get_next_key(fd, prev_key, &key)) {
struct map *kmap;
st->flags = data.flags;
+ if (con->aggr_mode == LOCK_AGGR_TASK) {
+ struct contention_task_data task;
+ struct thread *t;
+
+ st->addr = key.stack_or_task_id;
+
+ /* do not update idle comm which contains CPU number */
+ if (st->addr) {
+ bpf_map_lookup_elem(task_fd, &key, &task);
+ t = __machine__findnew_thread(machine, /*pid=*/-1,
+ key.stack_or_task_id);
+ thread__set_comm(t, task.comm, /*timestamp=*/0);
+ }
+ goto next;
+ }
+
bpf_map_lookup_elem(stack, &key, stack_trace);
/* skip lock internal functions */
if (st->callstack == NULL)
break;
}
-
+next:
hlist_add_head(&st->hash_entry, con->result);
prev_key = &key;
__uint(max_entries, MAX_ENTRIES);
} lock_stat SEC(".maps");
+struct {
+ __uint(type, BPF_MAP_TYPE_HASH);
+ __uint(key_size, sizeof(__u32));
+ __uint(value_size, sizeof(struct contention_task_data));
+ __uint(max_entries, MAX_ENTRIES);
+} task_data SEC(".maps");
+
struct {
__uint(type, BPF_MAP_TYPE_HASH);
__uint(key_size, sizeof(__u32));
int has_task;
int stack_skip;
+/* determine the key of lock stat */
+int aggr_mode;
+
/* error stat */
int lost;
return 1;
}
+static inline void update_task_data(__u32 pid)
+{
+ struct contention_task_data *p;
+
+ p = bpf_map_lookup_elem(&task_data, &pid);
+ if (p == NULL) {
+ struct contention_task_data data;
+
+ bpf_get_current_comm(data.comm, sizeof(data.comm));
+ bpf_map_update_elem(&task_data, &pid, &data, BPF_NOEXIST);
+ }
+}
+
SEC("tp_btf/contention_begin")
int contention_begin(u64 *ctx)
{
pelem->timestamp = bpf_ktime_get_ns();
pelem->lock = (__u64)ctx[0];
pelem->flags = (__u32)ctx[1];
- pelem->stack_id = bpf_get_stackid(ctx, &stacks, BPF_F_FAST_STACK_CMP | stack_skip);
- if (pelem->stack_id < 0)
- lost++;
+ if (aggr_mode == LOCK_AGGR_CALLER) {
+ pelem->stack_id = bpf_get_stackid(ctx, &stacks,
+ BPF_F_FAST_STACK_CMP | stack_skip);
+ if (pelem->stack_id < 0)
+ lost++;
+ }
+
return 0;
}
duration = bpf_ktime_get_ns() - pelem->timestamp;
- key.stack_or_task_id = pelem->stack_id;
+ if (aggr_mode == LOCK_AGGR_CALLER) {
+ key.stack_or_task_id = pelem->stack_id;
+ } else {
+ key.stack_or_task_id = pid;
+ update_task_data(pid);
+ }
+
data = bpf_map_lookup_elem(&lock_stat, &key);
if (!data) {
struct contention_data first = {