@@ -212,6 +212,10 @@ CONTENTION OPTIONS
--lock-cgroup::
Show lock contention stat by cgroup. Requires --use-bpf.
+-G::
+--cgroup-filter=<value>::
+ Show lock contention only in the given cgroups (comma separated list).
+
SEE ALSO
--------
@@ -10,6 +10,7 @@
#include "util/thread.h"
#include "util/header.h"
#include "util/target.h"
+#include "util/cgroup.h"
#include "util/callchain.h"
#include "util/lock-contention.h"
#include "util/bpf_skel/lock_data.h"
@@ -1631,6 +1632,9 @@ static void lock_filter_finish(void)
zfree(&filters.syms);
filters.nr_syms = 0;
+
+ zfree(&filters.cgrps);
+ filters.nr_cgrps = 0;
}
static void sort_contention_result(void)
@@ -2488,6 +2492,56 @@ static int parse_output(const struct option *opt __maybe_unused, const char *str
return 0;
}
+static bool add_lock_cgroup(char *name)
+{
+ u64 *tmp;
+ struct cgroup *cgrp;
+
+ cgrp = cgroup__new(name, /*do_open=*/false);
+ if (cgrp == NULL) {
+ pr_err("Failed to create cgroup: %s\n", name);
+ return false;
+ }
+
+ if (read_cgroup_id(cgrp) < 0) {
+ pr_err("Failed to read cgroup id for %s\n", name);
+ cgroup__put(cgrp);
+ return false;
+ }
+
+ tmp = realloc(filters.cgrps, (filters.nr_cgrps + 1) * sizeof(*filters.cgrps));
+ if (tmp == NULL) {
+ pr_err("Memory allocation failure\n");
+ return false;
+ }
+
+ tmp[filters.nr_cgrps++] = cgrp->id;
+ filters.cgrps = tmp;
+ cgroup__put(cgrp);
+ return true;
+}
+
+static int parse_cgroup_filter(const struct option *opt __maybe_unused, const char *str,
+ int unset __maybe_unused)
+{
+ char *s, *tmp, *tok;
+ int ret = 0;
+
+ s = strdup(str);
+ if (s == NULL)
+ return -1;
+
+ for (tok = strtok_r(s, ", ", &tmp); tok; tok = strtok_r(NULL, ", ", &tmp)) {
+ if (!add_lock_cgroup(tok)) {
+ ret = -1;
+ break;
+ }
+ }
+
+ free(s);
+ return ret;
+}
+
int cmd_lock(int argc, const char **argv)
{
const struct option lock_options[] = {
@@ -2562,6 +2616,8 @@ int cmd_lock(int argc, const char **argv)
OPT_STRING_NOEMPTY('x', "field-separator", &symbol_conf.field_sep, "separator",
"print result in CSV format with custom separator"),
OPT_BOOLEAN('g', "lock-cgroup", &show_lock_cgroups, "show lock stats by cgroup"),
+ OPT_CALLBACK('G', "cgroup-filter", NULL, "CGROUPS",
+ "Filter specific cgroups", parse_cgroup_filter),
OPT_PARENT(lock_options)
};
@@ -21,7 +21,7 @@ static struct lock_contention_bpf *skel;
int lock_contention_prepare(struct lock_contention *con)
{
int i, fd;
- int ncpus = 1, ntasks = 1, ntypes = 1, naddrs = 1;
+ int ncpus = 1, ntasks = 1, ntypes = 1, naddrs = 1, ncgrps = 1;
struct evlist *evlist = con->evlist;
struct target *target = con->target;
@@ -51,6 +51,8 @@ int lock_contention_prepare(struct lock_contention *con)
ntasks = perf_thread_map__nr(evlist->core.threads);
if (con->filters->nr_types)
ntypes = con->filters->nr_types;
+ if (con->filters->nr_cgrps)
+ ncgrps = con->filters->nr_cgrps;
/* resolve lock name filters to addr */
if (con->filters->nr_syms) {
@@ -85,6 +87,7 @@ int lock_contention_prepare(struct lock_contention *con)
bpf_map__set_max_entries(skel->maps.task_filter, ntasks);
bpf_map__set_max_entries(skel->maps.type_filter, ntypes);
bpf_map__set_max_entries(skel->maps.addr_filter, naddrs);
+ bpf_map__set_max_entries(skel->maps.cgroup_filter, ncgrps);
if (lock_contention_bpf__load(skel) < 0) {
pr_err("Failed to load lock-contention BPF skeleton\n");
@@ -146,6 +149,16 @@ int lock_contention_prepare(struct lock_contention *con)
bpf_map_update_elem(fd, &con->filters->addrs[i], &val, BPF_ANY);
}
+ if (con->filters->nr_cgrps) {
+ u8 val = 1;
+
+ skel->bss->has_cgroup = 1;
+ fd = bpf_map__fd(skel->maps.cgroup_filter);
+
+ for (i = 0; i < con->filters->nr_cgrps; i++)
+ bpf_map_update_elem(fd, &con->filters->cgrps[i], &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;
@@ -92,6 +92,13 @@ struct {
__uint(max_entries, 1);
} addr_filter SEC(".maps");
+struct {
+ __uint(type, BPF_MAP_TYPE_HASH);
+ __uint(key_size, sizeof(__u64));
+ __uint(value_size, sizeof(__u8));
+ __uint(max_entries, 1);
+} cgroup_filter SEC(".maps");
+
struct rw_semaphore___old {
struct task_struct *owner;
} __attribute__((preserve_access_index));
@@ -114,6 +121,7 @@ int has_cpu;
int has_task;
int has_type;
int has_addr;
+int has_cgroup;
int needs_callstack;
int stack_skip;
int lock_owner;
@@ -194,6 +202,15 @@ static inline int can_record(u64 *ctx)
return 0;
}
+ if (has_cgroup) {
+ __u8 *ok;
+ __u64 cgrp = get_current_cgroup_id();
+
+ ok = bpf_map_lookup_elem(&cgroup_filter, &cgrp);
+ if (!ok)
+ return 0;
+ }
+
return 1;
}
@@ -114,7 +114,7 @@ static struct cgroup *evlist__find_cgroup(struct evlist *evlist, const char *str
return NULL;
}
-static struct cgroup *cgroup__new(const char *name, bool do_open)
+struct cgroup *cgroup__new(const char *name, bool do_open)
{
struct cgroup *cgroup = zalloc(sizeof(*cgroup));
@@ -26,6 +26,7 @@ void cgroup__put(struct cgroup *cgroup);
struct evlist;
struct rblist;
+struct cgroup *cgroup__new(const char *name, bool do_open);
struct cgroup *evlist__findnew_cgroup(struct evlist *evlist, const char *name);
int evlist__expand_cgroup(struct evlist *evlist, const char *cgroups,
struct rblist *metric_events, bool open_cgroup);
@@ -9,9 +9,11 @@ struct lock_filter {
int nr_types;
int nr_addrs;
int nr_syms;
+ int nr_cgrps;
unsigned int *types;
unsigned long *addrs;
char **syms;
+ u64 *cgrps;
};
struct lock_stat {
The -G/--cgroup-filter is to limit lock contention collection on the tasks in the specific cgroups only. $ sudo ./perf lock con -abt -G /user.slice/.../vte-spawn-52221fb8-b33f-4a52-b5c3-e35d1e6fc0e0.scope \ ./perf bench sched messaging # Running 'sched/messaging' benchmark: # 20 sender and receiver processes per group # 10 groups == 400 processes run Total time: 0.174 [sec] contended total wait max wait avg wait pid comm 4 114.45 us 60.06 us 28.61 us 214847 sched-messaging 2 111.40 us 60.84 us 55.70 us 214848 sched-messaging 2 106.09 us 59.42 us 53.04 us 214837 sched-messaging 1 81.70 us 81.70 us 81.70 us 214709 sched-messaging 68 78.44 us 6.83 us 1.15 us 214633 sched-messaging 69 73.71 us 2.69 us 1.07 us 214632 sched-messaging 4 72.62 us 60.83 us 18.15 us 214850 sched-messaging 2 71.75 us 67.60 us 35.88 us 214840 sched-messaging 2 69.29 us 67.53 us 34.65 us 214804 sched-messaging 2 69.00 us 68.23 us 34.50 us 214826 sched-messaging ... Export cgroup__new() function as it's needed from outside. Signed-off-by: Namhyung Kim <namhyung@kernel.org> --- tools/perf/Documentation/perf-lock.txt | 4 ++ tools/perf/builtin-lock.c | 56 +++++++++++++++++++ tools/perf/util/bpf_lock_contention.c | 15 ++++- .../perf/util/bpf_skel/lock_contention.bpf.c | 17 ++++++ tools/perf/util/cgroup.c | 2 +- tools/perf/util/cgroup.h | 1 + tools/perf/util/lock-contention.h | 2 + 7 files changed, 95 insertions(+), 2 deletions(-)