aboutsummaryrefslogtreecommitdiffstatshomepage
path: root/tools/perf/bench
diff options
context:
space:
mode:
Diffstat (limited to 'tools/perf/bench')
-rw-r--r--tools/perf/bench/Build40
-rw-r--r--tools/perf/bench/bench.h28
-rw-r--r--tools/perf/bench/breakpoint.c262
-rw-r--r--tools/perf/bench/epoll-ctl.c70
-rw-r--r--tools/perf/bench/epoll-wait.c79
-rw-r--r--tools/perf/bench/evlist-open-close.c269
-rw-r--r--tools/perf/bench/find-bit-bench.c139
-rw-r--r--tools/perf/bench/futex-hash.c130
-rw-r--r--tools/perf/bench/futex-lock-pi.c135
-rw-r--r--tools/perf/bench/futex-requeue.c235
-rw-r--r--tools/perf/bench/futex-wake-parallel.c152
-rw-r--r--tools/perf/bench/futex-wake.c138
-rw-r--r--tools/perf/bench/futex.c74
-rw-r--r--tools/perf/bench/futex.h92
-rw-r--r--tools/perf/bench/inject-buildid.c486
-rw-r--r--tools/perf/bench/mem-functions.c21
-rw-r--r--tools/perf/bench/mem-memcpy-x86-64-asm-def.h4
-rw-r--r--tools/perf/bench/mem-memcpy-x86-64-asm.S3
-rw-r--r--tools/perf/bench/mem-memcpy-x86-64-lib.c24
-rw-r--r--tools/perf/bench/mem-memset-x86-64-asm-def.h4
-rw-r--r--tools/perf/bench/mem-memset-x86-64-asm.S3
-rw-r--r--tools/perf/bench/numa.c429
-rw-r--r--tools/perf/bench/pmu-scan.c186
-rw-r--r--tools/perf/bench/sched-messaging.c128
-rw-r--r--tools/perf/bench/sched-pipe.c190
-rw-r--r--tools/perf/bench/sched-seccomp-notify.c178
-rw-r--r--tools/perf/bench/synthesize.c14
-rw-r--r--tools/perf/bench/syscall.c188
-rw-r--r--tools/perf/bench/uprobe.c213
29 files changed, 3267 insertions, 647 deletions
diff --git a/tools/perf/bench/Build b/tools/perf/bench/Build
index 768e408757a0..b558ab98719f 100644
--- a/tools/perf/bench/Build
+++ b/tools/perf/bench/Build
@@ -1,18 +1,26 @@
-perf-y += sched-messaging.o
-perf-y += sched-pipe.o
-perf-y += mem-functions.o
-perf-y += futex-hash.o
-perf-y += futex-wake.o
-perf-y += futex-wake-parallel.o
-perf-y += futex-requeue.o
-perf-y += futex-lock-pi.o
-perf-y += epoll-wait.o
-perf-y += epoll-ctl.o
-perf-y += synthesize.o
-perf-y += kallsyms-parse.o
+perf-bench-y += sched-messaging.o
+perf-bench-y += sched-pipe.o
+perf-bench-y += sched-seccomp-notify.o
+perf-bench-y += syscall.o
+perf-bench-y += mem-functions.o
+perf-bench-y += futex.o
+perf-bench-y += futex-hash.o
+perf-bench-y += futex-wake.o
+perf-bench-y += futex-wake-parallel.o
+perf-bench-y += futex-requeue.o
+perf-bench-y += futex-lock-pi.o
+perf-bench-y += epoll-wait.o
+perf-bench-y += epoll-ctl.o
+perf-bench-y += synthesize.o
+perf-bench-y += kallsyms-parse.o
+perf-bench-y += find-bit-bench.o
+perf-bench-y += inject-buildid.o
+perf-bench-y += evlist-open-close.o
+perf-bench-y += breakpoint.o
+perf-bench-y += pmu-scan.o
+perf-bench-y += uprobe.o
-perf-$(CONFIG_X86_64) += mem-memcpy-x86-64-lib.o
-perf-$(CONFIG_X86_64) += mem-memcpy-x86-64-asm.o
-perf-$(CONFIG_X86_64) += mem-memset-x86-64-asm.o
+perf-bench-$(CONFIG_X86_64) += mem-memcpy-x86-64-asm.o
+perf-bench-$(CONFIG_X86_64) += mem-memset-x86-64-asm.o
-perf-$(CONFIG_NUMA) += numa.o
+perf-bench-$(CONFIG_NUMA) += numa.o
diff --git a/tools/perf/bench/bench.h b/tools/perf/bench/bench.h
index 61cae4966cae..9f736423af53 100644
--- a/tools/perf/bench/bench.h
+++ b/tools/perf/bench/bench.h
@@ -10,31 +10,25 @@ extern struct timeval bench__start, bench__end, bench__runtime;
* The madvise transparent hugepage constants were added in glibc
* 2.13. For compatibility with older versions of glibc, define these
* tokens if they are not already defined.
- *
- * PA-RISC uses different madvise values from other architectures and
- * needs to be special-cased.
*/
-#ifdef __hppa__
-# ifndef MADV_HUGEPAGE
-# define MADV_HUGEPAGE 67
-# endif
-# ifndef MADV_NOHUGEPAGE
-# define MADV_NOHUGEPAGE 68
-# endif
-#else
# ifndef MADV_HUGEPAGE
# define MADV_HUGEPAGE 14
# endif
# ifndef MADV_NOHUGEPAGE
# define MADV_NOHUGEPAGE 15
# endif
-#endif
int bench_numa(int argc, const char **argv);
int bench_sched_messaging(int argc, const char **argv);
int bench_sched_pipe(int argc, const char **argv);
+int bench_sched_seccomp_notify(int argc, const char **argv);
+int bench_syscall_basic(int argc, const char **argv);
+int bench_syscall_getpgid(int argc, const char **argv);
+int bench_syscall_fork(int argc, const char **argv);
+int bench_syscall_execve(int argc, const char **argv);
int bench_mem_memcpy(int argc, const char **argv);
int bench_mem_memset(int argc, const char **argv);
+int bench_mem_find_bit(int argc, const char **argv);
int bench_futex_hash(int argc, const char **argv);
int bench_futex_wake(int argc, const char **argv);
int bench_futex_wake_parallel(int argc, const char **argv);
@@ -45,6 +39,16 @@ int bench_epoll_wait(int argc, const char **argv);
int bench_epoll_ctl(int argc, const char **argv);
int bench_synthesize(int argc, const char **argv);
int bench_kallsyms_parse(int argc, const char **argv);
+int bench_inject_build_id(int argc, const char **argv);
+int bench_evlist_open_close(int argc, const char **argv);
+int bench_breakpoint_thread(int argc, const char **argv);
+int bench_breakpoint_enable(int argc, const char **argv);
+int bench_uprobe_baseline(int argc, const char **argv);
+int bench_uprobe_empty(int argc, const char **argv);
+int bench_uprobe_trace_printk(int argc, const char **argv);
+int bench_uprobe_empty_ret(int argc, const char **argv);
+int bench_uprobe_trace_printk_ret(int argc, const char **argv);
+int bench_pmu_scan(int argc, const char **argv);
#define BENCH_FORMAT_DEFAULT_STR "default"
#define BENCH_FORMAT_DEFAULT 0
diff --git a/tools/perf/bench/breakpoint.c b/tools/perf/bench/breakpoint.c
new file mode 100644
index 000000000000..dfd18f5db97d
--- /dev/null
+++ b/tools/perf/bench/breakpoint.c
@@ -0,0 +1,262 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <subcmd/parse-options.h>
+#include <linux/hw_breakpoint.h>
+#include <linux/perf_event.h>
+#include <linux/time64.h>
+#include <sys/syscall.h>
+#include <sys/ioctl.h>
+#include <sys/time.h>
+#include <pthread.h>
+#include <stddef.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <errno.h>
+#include "bench.h"
+#include "futex.h"
+
+struct {
+ unsigned int nbreakpoints;
+ unsigned int nparallel;
+ unsigned int nthreads;
+} thread_params = {
+ .nbreakpoints = 1,
+ .nparallel = 1,
+ .nthreads = 1,
+};
+
+static const struct option thread_options[] = {
+ OPT_UINTEGER('b', "breakpoints", &thread_params.nbreakpoints,
+ "Specify amount of breakpoints"),
+ OPT_UINTEGER('p', "parallelism", &thread_params.nparallel, "Specify amount of parallelism"),
+ OPT_UINTEGER('t', "threads", &thread_params.nthreads, "Specify amount of threads"),
+ OPT_END()
+};
+
+static const char * const thread_usage[] = {
+ "perf bench breakpoint thread <options>",
+ NULL
+};
+
+struct breakpoint {
+ int fd;
+ char watched;
+};
+
+static int breakpoint_setup(void *addr)
+{
+ struct perf_event_attr attr = { .size = 0, };
+ int fd;
+
+ attr.type = PERF_TYPE_BREAKPOINT;
+ attr.size = sizeof(attr);
+ attr.inherit = 1;
+ attr.exclude_kernel = 1;
+ attr.exclude_hv = 1;
+ attr.bp_addr = (unsigned long)addr;
+ attr.bp_type = HW_BREAKPOINT_RW;
+ attr.bp_len = HW_BREAKPOINT_LEN_1;
+ fd = syscall(SYS_perf_event_open, &attr, 0, -1, -1, 0);
+
+ if (fd < 0)
+ fd = -errno;
+
+ return fd;
+}
+
+static void *passive_thread(void *arg)
+{
+ unsigned int *done = (unsigned int *)arg;
+
+ while (!__atomic_load_n(done, __ATOMIC_RELAXED))
+ futex_wait(done, 0, NULL, 0);
+ return NULL;
+}
+
+static void *active_thread(void *arg)
+{
+ unsigned int *done = (unsigned int *)arg;
+
+ while (!__atomic_load_n(done, __ATOMIC_RELAXED));
+ return NULL;
+}
+
+static void *breakpoint_thread(void *arg)
+{
+ unsigned int i, done;
+ int *repeat = (int *)arg;
+ pthread_t *threads;
+
+ threads = calloc(thread_params.nthreads, sizeof(threads[0]));
+ if (!threads)
+ exit((perror("calloc"), EXIT_FAILURE));
+
+ while (__atomic_fetch_sub(repeat, 1, __ATOMIC_RELAXED) > 0) {
+ done = 0;
+ for (i = 0; i < thread_params.nthreads; i++) {
+ if (pthread_create(&threads[i], NULL, passive_thread, &done))
+ exit((perror("pthread_create"), EXIT_FAILURE));
+ }
+ __atomic_store_n(&done, 1, __ATOMIC_RELAXED);
+ futex_wake(&done, thread_params.nthreads, 0);
+ for (i = 0; i < thread_params.nthreads; i++)
+ pthread_join(threads[i], NULL);
+ }
+ free(threads);
+ return NULL;
+}
+
+// The benchmark creates nbreakpoints inheritable breakpoints,
+// then starts nparallel threads which create and join bench_repeat batches of nthreads threads.
+int bench_breakpoint_thread(int argc, const char **argv)
+{
+ unsigned int i, result_usec;
+ int repeat = bench_repeat;
+ struct breakpoint *breakpoints;
+ pthread_t *parallel;
+ struct timeval start, stop, diff;
+
+ if (parse_options(argc, argv, thread_options, thread_usage, 0)) {
+ usage_with_options(thread_usage, thread_options);
+ exit(EXIT_FAILURE);
+ }
+ breakpoints = calloc(thread_params.nbreakpoints, sizeof(breakpoints[0]));
+ parallel = calloc(thread_params.nparallel, sizeof(parallel[0]));
+ if (!breakpoints || !parallel)
+ exit((perror("calloc"), EXIT_FAILURE));
+
+ for (i = 0; i < thread_params.nbreakpoints; i++) {
+ breakpoints[i].fd = breakpoint_setup(&breakpoints[i].watched);
+
+ if (breakpoints[i].fd < 0) {
+ if (breakpoints[i].fd == -ENODEV) {
+ printf("Skipping perf bench breakpoint thread: No hardware support\n");
+ return 0;
+ }
+ exit((perror("perf_event_open"), EXIT_FAILURE));
+ }
+ }
+ gettimeofday(&start, NULL);
+ for (i = 0; i < thread_params.nparallel; i++) {
+ if (pthread_create(&parallel[i], NULL, breakpoint_thread, &repeat))
+ exit((perror("pthread_create"), EXIT_FAILURE));
+ }
+ for (i = 0; i < thread_params.nparallel; i++)
+ pthread_join(parallel[i], NULL);
+ gettimeofday(&stop, NULL);
+ timersub(&stop, &start, &diff);
+ for (i = 0; i < thread_params.nbreakpoints; i++)
+ close(breakpoints[i].fd);
+ free(parallel);
+ free(breakpoints);
+ switch (bench_format) {
+ case BENCH_FORMAT_DEFAULT:
+ printf("# Created/joined %d threads with %d breakpoints and %d parallelism\n",
+ bench_repeat, thread_params.nbreakpoints, thread_params.nparallel);
+ printf(" %14s: %lu.%03lu [sec]\n\n", "Total time",
+ (long)diff.tv_sec, (long)(diff.tv_usec / USEC_PER_MSEC));
+ result_usec = diff.tv_sec * USEC_PER_SEC + diff.tv_usec;
+ printf(" %14lf usecs/op\n",
+ (double)result_usec / bench_repeat / thread_params.nthreads);
+ printf(" %14lf usecs/op/cpu\n",
+ (double)result_usec / bench_repeat /
+ thread_params.nthreads * thread_params.nparallel);
+ break;
+ case BENCH_FORMAT_SIMPLE:
+ printf("%lu.%03lu\n", (long)diff.tv_sec, (long)(diff.tv_usec / USEC_PER_MSEC));
+ break;
+ default:
+ fprintf(stderr, "Unknown format: %d\n", bench_format);
+ exit(EXIT_FAILURE);
+ }
+ return 0;
+}
+
+struct {
+ unsigned int npassive;
+ unsigned int nactive;
+} enable_params = {
+ .nactive = 0,
+ .npassive = 0,
+};
+
+static const struct option enable_options[] = {
+ OPT_UINTEGER('p', "passive", &enable_params.npassive, "Specify amount of passive threads"),
+ OPT_UINTEGER('a', "active", &enable_params.nactive, "Specify amount of active threads"),
+ OPT_END()
+};
+
+static const char * const enable_usage[] = {
+ "perf bench breakpoint enable <options>",
+ NULL
+};
+
+// The benchmark creates an inheritable breakpoint,
+// then starts npassive threads that block and nactive threads that actively spin
+// and then disables and enables the breakpoint bench_repeat times.
+int bench_breakpoint_enable(int argc, const char **argv)
+{
+ unsigned int i, nthreads, result_usec, done = 0;
+ char watched;
+ int fd;
+ pthread_t *threads;
+ struct timeval start, stop, diff;
+
+ if (parse_options(argc, argv, enable_options, enable_usage, 0)) {
+ usage_with_options(enable_usage, enable_options);
+ exit(EXIT_FAILURE);
+ }
+ fd = breakpoint_setup(&watched);
+
+ if (fd < 0) {
+ if (fd == -ENODEV) {
+ printf("Skipping perf bench breakpoint enable: No hardware support\n");
+ return 0;
+ }
+ exit((perror("perf_event_open"), EXIT_FAILURE));
+ }
+ nthreads = enable_params.npassive + enable_params.nactive;
+ threads = calloc(nthreads, sizeof(threads[0]));
+ if (!threads)
+ exit((perror("calloc"), EXIT_FAILURE));
+
+ for (i = 0; i < nthreads; i++) {
+ if (pthread_create(&threads[i], NULL,
+ i < enable_params.npassive ? passive_thread : active_thread, &done))
+ exit((perror("pthread_create"), EXIT_FAILURE));
+ }
+ usleep(10000); // let the threads block
+ gettimeofday(&start, NULL);
+ for (i = 0; i < bench_repeat; i++) {
+ if (ioctl(fd, PERF_EVENT_IOC_DISABLE, 0))
+ exit((perror("ioctl(PERF_EVENT_IOC_DISABLE)"), EXIT_FAILURE));
+ if (ioctl(fd, PERF_EVENT_IOC_ENABLE, 0))
+ exit((perror("ioctl(PERF_EVENT_IOC_ENABLE)"), EXIT_FAILURE));
+ }
+ gettimeofday(&stop, NULL);
+ timersub(&stop, &start, &diff);
+ __atomic_store_n(&done, 1, __ATOMIC_RELAXED);
+ futex_wake(&done, enable_params.npassive, 0);
+ for (i = 0; i < nthreads; i++)
+ pthread_join(threads[i], NULL);
+ free(threads);
+ close(fd);
+ switch (bench_format) {
+ case BENCH_FORMAT_DEFAULT:
+ printf("# Enabled/disabled breakpoint %d time with %d passive and %d active threads\n",
+ bench_repeat, enable_params.npassive, enable_params.nactive);
+ printf(" %14s: %lu.%03lu [sec]\n\n", "Total time",
+ (long)diff.tv_sec, (long)(diff.tv_usec / USEC_PER_MSEC));
+ result_usec = diff.tv_sec * USEC_PER_SEC + diff.tv_usec;
+ printf(" %14lf usecs/op\n", (double)result_usec / bench_repeat);
+ break;
+ case BENCH_FORMAT_SIMPLE:
+ printf("%lu.%03lu\n", (long)diff.tv_sec, (long)(diff.tv_usec / USEC_PER_MSEC));
+ break;
+ default:
+ fprintf(stderr, "Unknown format: %d\n", bench_format);
+ exit(EXIT_FAILURE);
+ }
+ return 0;
+}
diff --git a/tools/perf/bench/epoll-ctl.c b/tools/perf/bench/epoll-ctl.c
index ca2d591aad8a..d66d852b90e4 100644
--- a/tools/perf/bench/epoll-ctl.c
+++ b/tools/perf/bench/epoll-ctl.c
@@ -21,9 +21,9 @@
#include <sys/resource.h>
#include <sys/epoll.h>
#include <sys/eventfd.h>
-#include <internal/cpumap.h>
#include <perf/cpumap.h>
+#include "../util/mutex.h"
#include "../util/stat.h"
#include <subcmd/parse-options.h>
#include "bench.h"
@@ -59,10 +59,10 @@ static unsigned int nested = 0;
/* amount of fds to monitor, per thread */
static unsigned int nfds = 64;
-static pthread_mutex_t thread_lock;
+static struct mutex thread_lock;
static unsigned int threads_starting;
static struct stats all_stats[EPOLL_NR_OPS];
-static pthread_cond_t thread_parent, thread_worker;
+static struct cond thread_parent, thread_worker;
struct worker {
int tid;
@@ -107,7 +107,7 @@ static void nest_epollfd(void)
printinfo("Nesting level(s): %d\n", nested);
epollfdp = calloc(nested, sizeof(int));
- if (!epollfd)
+ if (!epollfdp)
err(EXIT_FAILURE, "calloc");
for (i = 0; i < nested; i++) {
@@ -175,12 +175,12 @@ static void *workerfn(void *arg)
struct timespec ts = { .tv_sec = 0,
.tv_nsec = 250 };
- pthread_mutex_lock(&thread_lock);
+ mutex_lock(&thread_lock);
threads_starting--;
if (!threads_starting)
- pthread_cond_signal(&thread_parent);
- pthread_cond_wait(&thread_worker, &thread_lock);
- pthread_mutex_unlock(&thread_lock);
+ cond_signal(&thread_parent);
+ cond_wait(&thread_worker, &thread_lock);
+ mutex_unlock(&thread_lock);
/* Let 'em loose */
do {
@@ -223,13 +223,20 @@ static void init_fdmaps(struct worker *w, int pct)
static int do_threads(struct worker *worker, struct perf_cpu_map *cpu)
{
pthread_attr_t thread_attr, *attrp = NULL;
- cpu_set_t cpuset;
+ cpu_set_t *cpuset;
unsigned int i, j;
int ret = 0;
+ int nrcpus;
+ size_t size;
if (!noaffinity)
pthread_attr_init(&thread_attr);
+ nrcpus = cpu__max_cpu().cpu;
+ cpuset = CPU_ALLOC(nrcpus);
+ BUG_ON(!cpuset);
+ size = CPU_ALLOC_SIZE(nrcpus);
+
for (i = 0; i < nthreads; i++) {
struct worker *w = &worker[i];
@@ -253,22 +260,28 @@ static int do_threads(struct worker *worker, struct perf_cpu_map *cpu)
init_fdmaps(w, 50);
if (!noaffinity) {
- CPU_ZERO(&cpuset);
- CPU_SET(cpu->map[i % cpu->nr], &cpuset);
+ CPU_ZERO_S(size, cpuset);
+ CPU_SET_S(perf_cpu_map__cpu(cpu, i % perf_cpu_map__nr(cpu)).cpu,
+ size, cpuset);
- ret = pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset);
- if (ret)
+ ret = pthread_attr_setaffinity_np(&thread_attr, size, cpuset);
+ if (ret) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_attr_setaffinity_np");
+ }
attrp = &thread_attr;
}
ret = pthread_create(&w->thread, attrp, workerfn,
(void *)(struct worker *) w);
- if (ret)
+ if (ret) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_create");
+ }
}
+ CPU_FREE(cpuset);
if (!noaffinity)
pthread_attr_destroy(&thread_attr);
@@ -317,7 +330,7 @@ int bench_epoll_ctl(int argc, const char **argv)
act.sa_sigaction = toggle_done;
sigaction(SIGINT, &act, NULL);
- cpu = perf_cpu_map__new(NULL);
+ cpu = perf_cpu_map__new_online_cpus();
if (!cpu)
goto errmem;
@@ -334,7 +347,7 @@ int bench_epoll_ctl(int argc, const char **argv)
/* default to the number of CPUs */
if (!nthreads)
- nthreads = cpu->nr;
+ nthreads = perf_cpu_map__nr(cpu);
worker = calloc(nthreads, sizeof(*worker));
if (!worker)
@@ -355,9 +368,9 @@ int bench_epoll_ctl(int argc, const char **argv)
for (i = 0; i < EPOLL_NR_OPS; i++)
init_stats(&all_stats[i]);
- pthread_mutex_init(&thread_lock, NULL);
- pthread_cond_init(&thread_parent, NULL);
- pthread_cond_init(&thread_worker, NULL);
+ mutex_init(&thread_lock);
+ cond_init(&thread_parent);
+ cond_init(&thread_worker);
threads_starting = nthreads;
@@ -365,11 +378,11 @@ int bench_epoll_ctl(int argc, const char **argv)
do_threads(worker, cpu);
- pthread_mutex_lock(&thread_lock);
+ mutex_lock(&thread_lock);
while (threads_starting)
- pthread_cond_wait(&thread_parent, &thread_lock);
- pthread_cond_broadcast(&thread_worker);
- pthread_mutex_unlock(&thread_lock);
+ cond_wait(&thread_parent, &thread_lock);
+ cond_broadcast(&thread_worker);
+ mutex_unlock(&thread_lock);
sleep(nsecs);
toggle_done(0, NULL, NULL);
@@ -382,9 +395,9 @@ int bench_epoll_ctl(int argc, const char **argv)
}
/* cleanup & report results */
- pthread_cond_destroy(&thread_parent);
- pthread_cond_destroy(&thread_worker);
- pthread_mutex_destroy(&thread_lock);
+ cond_destroy(&thread_parent);
+ cond_destroy(&thread_worker);
+ mutex_destroy(&thread_lock);
for (i = 0; i < nthreads; i++) {
unsigned long t[EPOLL_NR_OPS];
@@ -408,6 +421,11 @@ int bench_epoll_ctl(int argc, const char **argv)
print_summary();
close(epollfd);
+ perf_cpu_map__put(cpu);
+ for (i = 0; i < nthreads; i++)
+ free(worker[i].fdmap);
+
+ free(worker);
return ret;
errmem:
err(EXIT_FAILURE, "calloc");
diff --git a/tools/perf/bench/epoll-wait.c b/tools/perf/bench/epoll-wait.c
index 75dca9773186..20fe4f72b4af 100644
--- a/tools/perf/bench/epoll-wait.c
+++ b/tools/perf/bench/epoll-wait.c
@@ -17,7 +17,7 @@
* While the second model, enabled via --multiq option, uses multiple
* queueing (which refers to one epoll instance per worker). For example,
* short lived tcp connections in a high throughput httpd server will
- * ditribute the accept()'ing connections across CPUs. In this case each
+ * distribute the accept()'ing connections across CPUs. In this case each
* worker does a limited amount of processing.
*
* [queue A] ---> [worker]
@@ -76,10 +76,10 @@
#include <sys/epoll.h>
#include <sys/eventfd.h>
#include <sys/types.h>
-#include <internal/cpumap.h>
#include <perf/cpumap.h>
#include "../util/stat.h"
+#include "../util/mutex.h"
#include <subcmd/parse-options.h>
#include "bench.h"
@@ -110,10 +110,10 @@ static bool multiq; /* use an epoll instance per thread */
/* amount of fds to monitor, per thread */
static unsigned int nfds = 64;
-static pthread_mutex_t thread_lock;
+static struct mutex thread_lock;
static unsigned int threads_starting;
static struct stats throughput_stats;
-static pthread_cond_t thread_parent, thread_worker;
+static struct cond thread_parent, thread_worker;
struct worker {
int tid;
@@ -190,16 +190,16 @@ static void *workerfn(void *arg)
int to = nonblocking? 0 : -1;
int efd = multiq ? w->epollfd : epollfd;
- pthread_mutex_lock(&thread_lock);
+ mutex_lock(&thread_lock);
threads_starting--;
if (!threads_starting)
- pthread_cond_signal(&thread_parent);
- pthread_cond_wait(&thread_worker, &thread_lock);
- pthread_mutex_unlock(&thread_lock);
+ cond_signal(&thread_parent);
+ cond_wait(&thread_worker, &thread_lock);
+ mutex_unlock(&thread_lock);
do {
/*
- * Block undefinitely waiting for the IN event.
+ * Block indefinitely waiting for the IN event.
* In order to stress the epoll_wait(2) syscall,
* call it event per event, instead of a larger
* batch (max)limit.
@@ -292,9 +292,11 @@ static void print_summary(void)
static int do_threads(struct worker *worker, struct perf_cpu_map *cpu)
{
pthread_attr_t thread_attr, *attrp = NULL;
- cpu_set_t cpuset;
+ cpu_set_t *cpuset;
unsigned int i, j;
int ret = 0, events = EPOLLIN;
+ int nrcpus;
+ size_t size;
if (oneshot)
events |= EPOLLONESHOT;
@@ -307,6 +309,11 @@ static int do_threads(struct worker *worker, struct perf_cpu_map *cpu)
if (!noaffinity)
pthread_attr_init(&thread_attr);
+ nrcpus = cpu__max_cpu().cpu;
+ cpuset = CPU_ALLOC(nrcpus);
+ BUG_ON(!cpuset);
+ size = CPU_ALLOC_SIZE(nrcpus);
+
for (i = 0; i < nthreads; i++) {
struct worker *w = &worker[i];
@@ -342,22 +349,28 @@ static int do_threads(struct worker *worker, struct perf_cpu_map *cpu)
}
if (!noaffinity) {
- CPU_ZERO(&cpuset);
- CPU_SET(cpu->map[i % cpu->nr], &cpuset);
+ CPU_ZERO_S(size, cpuset);
+ CPU_SET_S(perf_cpu_map__cpu(cpu, i % perf_cpu_map__nr(cpu)).cpu,
+ size, cpuset);
- ret = pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset);
- if (ret)
+ ret = pthread_attr_setaffinity_np(&thread_attr, size, cpuset);
+ if (ret) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_attr_setaffinity_np");
+ }
attrp = &thread_attr;
}
ret = pthread_create(&w->thread, attrp, workerfn,
(void *)(struct worker *) w);
- if (ret)
+ if (ret) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_create");
+ }
}
+ CPU_FREE(cpuset);
if (!noaffinity)
pthread_attr_destroy(&thread_attr);
@@ -407,7 +420,12 @@ static int cmpworker(const void *p1, const void *p2)
struct worker *w1 = (struct worker *) p1;
struct worker *w2 = (struct worker *) p2;
- return w1->tid > w2->tid;
+
+ if (w1->tid > w2->tid)
+ return 1;
+ if (w1->tid < w2->tid)
+ return -1;
+ return 0;
}
int bench_epoll_wait(int argc, const char **argv)
@@ -431,7 +449,7 @@ int bench_epoll_wait(int argc, const char **argv)
act.sa_sigaction = toggle_done;
sigaction(SIGINT, &act, NULL);
- cpu = perf_cpu_map__new(NULL);
+ cpu = perf_cpu_map__new_online_cpus();
if (!cpu)
goto errmem;
@@ -453,7 +471,7 @@ int bench_epoll_wait(int argc, const char **argv)
/* default to the number of CPUs and leave one for the writer pthread */
if (!nthreads)
- nthreads = cpu->nr - 1;
+ nthreads = perf_cpu_map__nr(cpu) - 1;
worker = calloc(nthreads, sizeof(*worker));
if (!worker) {
@@ -473,9 +491,9 @@ int bench_epoll_wait(int argc, const char **argv)
getpid(), nthreads, oneshot ? " (EPOLLONESHOT semantics)": "", nfds, nsecs);
init_stats(&throughput_stats);
- pthread_mutex_init(&thread_lock, NULL);
- pthread_cond_init(&thread_parent, NULL);
- pthread_cond_init(&thread_worker, NULL);
+ mutex_init(&thread_lock);
+ cond_init(&thread_parent);
+ cond_init(&thread_worker);
threads_starting = nthreads;
@@ -483,11 +501,11 @@ int bench_epoll_wait(int argc, const char **argv)
do_threads(worker, cpu);
- pthread_mutex_lock(&thread_lock);
+ mutex_lock(&thread_lock);
while (threads_starting)
- pthread_cond_wait(&thread_parent, &thread_lock);
- pthread_cond_broadcast(&thread_worker);
- pthread_mutex_unlock(&thread_lock);
+ cond_wait(&thread_parent, &thread_lock);
+ cond_broadcast(&thread_worker);
+ mutex_unlock(&thread_lock);
/*
* At this point the workers should be blocked waiting for read events
@@ -510,9 +528,9 @@ int bench_epoll_wait(int argc, const char **argv)
err(EXIT_FAILURE, "pthread_join");
/* cleanup & report results */
- pthread_cond_destroy(&thread_parent);
- pthread_cond_destroy(&thread_worker);
- pthread_mutex_destroy(&thread_lock);
+ cond_destroy(&thread_parent);
+ cond_destroy(&thread_worker);
+ mutex_destroy(&thread_lock);
/* sort the array back before reporting */
if (randomize)
@@ -536,6 +554,11 @@ int bench_epoll_wait(int argc, const char **argv)
print_summary();
close(epollfd);
+ perf_cpu_map__put(cpu);
+ for (i = 0; i < nthreads; i++)
+ free(worker[i].fdmap);
+
+ free(worker);
return ret;
errmem:
err(EXIT_FAILURE, "calloc");
diff --git a/tools/perf/bench/evlist-open-close.c b/tools/perf/bench/evlist-open-close.c
new file mode 100644
index 000000000000..79cedcf94a39
--- /dev/null
+++ b/tools/perf/bench/evlist-open-close.c
@@ -0,0 +1,269 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <inttypes.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <limits.h>
+#include "bench.h"
+#include "../util/debug.h"
+#include "../util/stat.h"
+#include "../util/evlist.h"
+#include "../util/evsel.h"
+#include "../util/strbuf.h"
+#include "../util/record.h"
+#include "../util/parse-events.h"
+#include "internal/threadmap.h"
+#include "internal/cpumap.h"
+#include <linux/perf_event.h>
+#include <linux/kernel.h>
+#include <linux/time64.h>
+#include <linux/string.h>
+#include <subcmd/parse-options.h>
+
+#define MMAP_FLUSH_DEFAULT 1
+
+static int iterations = 100;
+static int nr_events = 1;
+static const char *event_string = "dummy";
+
+static inline u64 timeval2usec(struct timeval *tv)
+{
+ return tv->tv_sec * USEC_PER_SEC + tv->tv_usec;
+}
+
+static struct record_opts opts = {
+ .sample_time = true,
+ .mmap_pages = UINT_MAX,
+ .user_freq = UINT_MAX,
+ .user_interval = ULLONG_MAX,
+ .freq = 4000,
+ .target = {
+ .uses_mmap = true,
+ .default_per_cpu = true,
+ },
+ .mmap_flush = MMAP_FLUSH_DEFAULT,
+ .nr_threads_synthesize = 1,
+ .ctl_fd = -1,
+ .ctl_fd_ack = -1,
+};
+
+static int evlist__count_evsel_fds(struct evlist *evlist)
+{
+ struct evsel *evsel;
+ int cnt = 0;
+
+ evlist__for_each_entry(evlist, evsel)
+ cnt += evsel->core.threads->nr * perf_cpu_map__nr(evsel->core.cpus);
+
+ return cnt;
+}
+
+static struct evlist *bench__create_evlist(char *evstr)
+{
+ struct parse_events_error err;
+ struct evlist *evlist = evlist__new();
+ int ret;
+
+ if (!evlist) {
+ pr_err("Not enough memory to create evlist\n");
+ return NULL;
+ }
+
+ parse_events_error__init(&err);
+ ret = parse_events(evlist, evstr, &err);
+ if (ret) {
+ parse_events_error__print(&err, evstr);
+ parse_events_error__exit(&err);
+ pr_err("Run 'perf list' for a list of valid events\n");
+ ret = 1;
+ goto out_delete_evlist;
+ }
+ parse_events_error__exit(&err);
+ ret = evlist__create_maps(evlist, &opts.target);
+ if (ret < 0) {
+ pr_err("Not enough memory to create thread/cpu maps\n");
+ goto out_delete_evlist;
+ }
+
+ evlist__config(evlist, &opts, NULL);
+
+ return evlist;
+
+out_delete_evlist:
+ evlist__delete(evlist);
+ return NULL;
+}
+
+static int bench__do_evlist_open_close(struct evlist *evlist)
+{
+ char sbuf[STRERR_BUFSIZE];
+ int err = evlist__open(evlist);
+
+ if (err < 0) {
+ pr_err("evlist__open: %s\n", str_error_r(errno, sbuf, sizeof(sbuf)));
+ return err;
+ }
+
+ err = evlist__mmap(evlist, opts.mmap_pages);
+ if (err < 0) {
+ pr_err("evlist__mmap: %s\n", str_error_r(errno, sbuf, sizeof(sbuf)));
+ return err;
+ }
+
+ evlist__enable(evlist);
+ evlist__disable(evlist);
+ evlist__munmap(evlist);
+ evlist__close(evlist);
+
+ return 0;
+}
+
+static int bench_evlist_open_close__run(char *evstr)
+{
+ // used to print statistics only
+ struct evlist *evlist = bench__create_evlist(evstr);
+ double time_average, time_stddev;
+ struct timeval start, end, diff;
+ struct stats time_stats;
+ u64 runtime_us;
+ int i, err;
+
+ if (!evlist)
+ return -ENOMEM;
+
+ init_stats(&time_stats);
+
+ printf(" Number of cpus:\t%d\n", perf_cpu_map__nr(evlist->core.user_requested_cpus));
+ printf(" Number of threads:\t%d\n", evlist->core.threads->nr);
+ printf(" Number of events:\t%d (%d fds)\n",
+ evlist->core.nr_entries, evlist__count_evsel_fds(evlist));
+ printf(" Number of iterations:\t%d\n", iterations);
+
+ evlist__delete(evlist);
+
+ for (i = 0; i < iterations; i++) {
+ pr_debug("Started iteration %d\n", i);
+ evlist = bench__create_evlist(evstr);
+ if (!evlist)
+ return -ENOMEM;
+
+ gettimeofday(&start, NULL);
+ err = bench__do_evlist_open_close(evlist);
+ if (err) {
+ evlist__delete(evlist);
+ return err;
+ }
+
+ gettimeofday(&end, NULL);
+ timersub(&end, &start, &diff);
+ runtime_us = timeval2usec(&diff);
+ update_stats(&time_stats, runtime_us);
+
+ evlist__delete(evlist);
+ pr_debug("Iteration %d took:\t%" PRIu64 "us\n", i, runtime_us);
+ }
+
+ time_average = avg_stats(&time_stats);
+ time_stddev = stddev_stats(&time_stats);
+ printf(" Average open-close took: %.3f usec (+- %.3f usec)\n", time_average, time_stddev);
+
+ return 0;
+}
+
+static char *bench__repeat_event_string(const char *evstr, int n)
+{
+ char sbuf[STRERR_BUFSIZE];
+ struct strbuf buf;
+ int i, str_size = strlen(evstr),
+ final_size = str_size * n + n,
+ err = strbuf_init(&buf, final_size);
+
+ if (err) {
+ pr_err("strbuf_init: %s\n", str_error_r(err, sbuf, sizeof(sbuf)));
+ goto out_error;
+ }
+
+ for (i = 0; i < n; i++) {
+ err = strbuf_add(&buf, evstr, str_size);
+ if (err) {
+ pr_err("strbuf_add: %s\n", str_error_r(err, sbuf, sizeof(sbuf)));
+ goto out_error;
+ }
+
+ err = strbuf_addch(&buf, i == n-1 ? '\0' : ',');
+ if (err) {
+ pr_err("strbuf_addch: %s\n", str_error_r(err, sbuf, sizeof(sbuf)));
+ goto out_error;
+ }
+ }
+
+ return strbuf_detach(&buf, NULL);
+
+out_error:
+ strbuf_release(&buf);
+ return NULL;
+}
+
+
+int bench_evlist_open_close(int argc, const char **argv)
+{
+ const struct option options[] = {
+ OPT_STRING('e', "event", &event_string, "event",
+ "event selector. use 'perf list' to list available events"),
+ OPT_INTEGER('n', "nr-events", &nr_events,
+ "number of dummy events to create (default 1). If used with -e, it clones those events n times (1 = no change)"),
+ OPT_INTEGER('i', "iterations", &iterations,
+ "Number of iterations used to compute average (default=100)"),
+ OPT_BOOLEAN('a', "all-cpus", &opts.target.system_wide,
+ "system-wide collection from all CPUs"),
+ OPT_STRING('C', "cpu", &opts.target.cpu_list, "cpu",
+ "list of cpus where to open events"),
+ OPT_STRING('p', "pid", &opts.target.pid, "pid",
+ "record events on existing process id"),
+ OPT_STRING('t', "tid", &opts.target.tid, "tid",
+ "record events on existing thread id"),
+ OPT_STRING('u', "uid", &opts.target.uid_str, "user", "user to profile"),
+ OPT_BOOLEAN(0, "per-thread", &opts.target.per_thread, "use per-thread mmaps"),
+ OPT_END()
+ };
+ const char *const bench_usage[] = {
+ "perf bench internals evlist-open-close <options>",
+ NULL
+ };
+ char *evstr, errbuf[BUFSIZ];
+ int err;
+
+ argc = parse_options(argc, argv, options, bench_usage, 0);
+ if (argc) {
+ usage_with_options(bench_usage, options);
+ exit(EXIT_FAILURE);
+ }
+
+ err = target__validate(&opts.target);
+ if (err) {
+ target__strerror(&opts.target, err, errbuf, sizeof(errbuf));
+ pr_err("%s\n", errbuf);
+ goto out;
+ }
+
+ err = target__parse_uid(&opts.target);
+ if (err) {
+ target__strerror(&opts.target, err, errbuf, sizeof(errbuf));
+ pr_err("%s", errbuf);
+ goto out;
+ }
+
+ /* Enable ignoring missing threads when -u/-p option is defined. */
+ opts.ignore_missing_thread = opts.target.uid != UINT_MAX || opts.target.pid;
+
+ evstr = bench__repeat_event_string(event_string, nr_events);
+ if (!evstr) {
+ err = -ENOMEM;
+ goto out;
+ }
+
+ err = bench_evlist_open_close__run(evstr);
+
+ free(evstr);
+out:
+ return err;
+}
diff --git a/tools/perf/bench/find-bit-bench.c b/tools/perf/bench/find-bit-bench.c
new file mode 100644
index 000000000000..7e25b0e413f6
--- /dev/null
+++ b/tools/perf/bench/find-bit-bench.c
@@ -0,0 +1,139 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Benchmark find_next_bit and related bit operations.
+ *
+ * Copyright 2020 Google LLC.
+ */
+#include <stdlib.h>
+#include "bench.h"
+#include "../util/stat.h"
+#include <linux/bitmap.h>
+#include <linux/bitops.h>
+#include <linux/time64.h>
+#include <subcmd/parse-options.h>
+
+static unsigned int outer_iterations = 5;
+static unsigned int inner_iterations = 100000;
+
+static const struct option options[] = {
+ OPT_UINTEGER('i', "outer-iterations", &outer_iterations,
+ "Number of outer iterations used"),
+ OPT_UINTEGER('j', "inner-iterations", &inner_iterations,
+ "Number of inner iterations used"),
+ OPT_END()
+};
+
+static const char *const bench_usage[] = {
+ "perf bench mem find_bit <options>",
+ NULL
+};
+
+static unsigned int accumulator;
+static unsigned int use_of_val;
+
+static noinline void workload(int val)
+{
+ use_of_val += val;
+ accumulator++;
+}
+
+#if (defined(__i386__) || defined(__x86_64__)) && defined(__GCC_ASM_FLAG_OUTPUTS__)
+static bool asm_test_bit(long nr, const unsigned long *addr)
+{
+ bool oldbit;
+
+ asm volatile("bt %2,%1"
+ : "=@ccc" (oldbit)
+ : "m" (*(unsigned long *)addr), "Ir" (nr) : "memory");
+
+ return oldbit;
+}
+#else
+#define asm_test_bit test_bit
+#endif
+
+static int do_for_each_set_bit(unsigned int num_bits)
+{
+ unsigned long *to_test = bitmap_zalloc(num_bits);
+ struct timeval start, end, diff;
+ u64 runtime_us;
+ struct stats fb_time_stats, tb_time_stats;
+ double time_average, time_stddev;
+ unsigned int bit, i, j;
+ unsigned int set_bits, skip;
+
+ init_stats(&fb_time_stats);
+ init_stats(&tb_time_stats);
+
+ for (set_bits = 1; set_bits <= num_bits; set_bits <<= 1) {
+ bitmap_zero(to_test, num_bits);
+ skip = num_bits / set_bits;
+ for (i = 0; i < num_bits; i += skip)
+ __set_bit(i, to_test);
+
+ for (i = 0; i < outer_iterations; i++) {
+#ifndef NDEBUG
+ unsigned int old = accumulator;
+#endif
+
+ gettimeofday(&start, NULL);
+ for (j = 0; j < inner_iterations; j++) {
+ for_each_set_bit(bit, to_test, num_bits)
+ workload(bit);
+ }
+ gettimeofday(&end, NULL);
+ assert(old + (inner_iterations * set_bits) == accumulator);
+ timersub(&end, &start, &diff);
+ runtime_us = diff.tv_sec * USEC_PER_SEC + diff.tv_usec;
+ update_stats(&fb_time_stats, runtime_us);
+
+#ifndef NDEBUG
+ old = accumulator;
+#endif
+ gettimeofday(&start, NULL);
+ for (j = 0; j < inner_iterations; j++) {
+ for (bit = 0; bit < num_bits; bit++) {
+ if (asm_test_bit(bit, to_test))
+ workload(bit);
+ }
+ }
+ gettimeofday(&end, NULL);
+ assert(old + (inner_iterations * set_bits) == accumulator);
+ timersub(&end, &start, &diff);
+ runtime_us = diff.tv_sec * USEC_PER_SEC + diff.tv_usec;
+ update_stats(&tb_time_stats, runtime_us);
+ }
+
+ printf("%d operations %d bits set of %d bits\n",
+ inner_iterations, set_bits, num_bits);
+ time_average = avg_stats(&fb_time_stats);
+ time_stddev = stddev_stats(&fb_time_stats);
+ printf(" Average for_each_set_bit took: %.3f usec (+- %.3f usec)\n",
+ time_average, time_stddev);
+ time_average = avg_stats(&tb_time_stats);
+ time_stddev = stddev_stats(&tb_time_stats);
+ printf(" Average test_bit loop took: %.3f usec (+- %.3f usec)\n",
+ time_average, time_stddev);
+
+ if (use_of_val == accumulator) /* Try to avoid compiler tricks. */
+ printf("\n");
+ }
+ bitmap_free(to_test);
+ return 0;
+}
+
+int bench_mem_find_bit(int argc, const char **argv)
+{
+ int err = 0, i;
+
+ argc = parse_options(argc, argv, options, bench_usage, 0);
+ if (argc) {
+ usage_with_options(bench_usage, options);
+ exit(EXIT_FAILURE);
+ }
+
+ for (i = 1; i <= 2048; i <<= 1)
+ do_for_each_set_bit(i);
+
+ return err;
+}
diff --git a/tools/perf/bench/futex-hash.c b/tools/perf/bench/futex-hash.c
index 915bf3da7ce2..d2d6d7f3ea33 100644
--- a/tools/perf/bench/futex-hash.c
+++ b/tools/perf/bench/futex-hash.c
@@ -20,9 +20,11 @@
#include <linux/kernel.h>
#include <linux/zalloc.h>
#include <sys/time.h>
-#include <internal/cpumap.h>
+#include <sys/mman.h>
+#include <sys/prctl.h>
#include <perf/cpumap.h>
+#include "../util/mutex.h"
#include "../util/stat.h"
#include <subcmd/parse-options.h>
#include "bench.h"
@@ -30,18 +32,14 @@
#include <err.h>
-static unsigned int nthreads = 0;
-static unsigned int nsecs = 10;
-/* amount of futexes per thread */
-static unsigned int nfutexes = 1024;
-static bool fshared = false, done = false, silent = false;
+static bool done = false;
static int futex_flag = 0;
struct timeval bench__start, bench__end, bench__runtime;
-static pthread_mutex_t thread_lock;
+static struct mutex thread_lock;
static unsigned int threads_starting;
static struct stats throughput_stats;
-static pthread_cond_t thread_parent, thread_worker;
+static struct cond thread_parent, thread_worker;
struct worker {
int tid;
@@ -50,12 +48,21 @@ struct worker {
unsigned long ops;
};
+static struct bench_futex_parameters params = {
+ .nfutexes = 1024,
+ .runtime = 10,
+ .nbuckets = -1,
+};
+
static const struct option options[] = {
- OPT_UINTEGER('t', "threads", &nthreads, "Specify amount of threads"),
- OPT_UINTEGER('r', "runtime", &nsecs, "Specify runtime (in seconds)"),
- OPT_UINTEGER('f', "futexes", &nfutexes, "Specify amount of futexes per threads"),
- OPT_BOOLEAN( 's', "silent", &silent, "Silent mode: do not display data/details"),
- OPT_BOOLEAN( 'S', "shared", &fshared, "Use shared futexes instead of private ones"),
+ OPT_INTEGER( 'b', "buckets", &params.nbuckets, "Specify amount of hash buckets"),
+ OPT_BOOLEAN( 'I', "immutable", &params.buckets_immutable, "Make the hash buckets immutable"),
+ OPT_UINTEGER('t', "threads", &params.nthreads, "Specify amount of threads"),
+ OPT_UINTEGER('r', "runtime", &params.runtime, "Specify runtime (in seconds)"),
+ OPT_UINTEGER('f', "futexes", &params.nfutexes, "Specify amount of futexes per threads"),
+ OPT_BOOLEAN( 's', "silent", &params.silent, "Silent mode: do not display data/details"),
+ OPT_BOOLEAN( 'S', "shared", &params.fshared, "Use shared futexes instead of private ones"),
+ OPT_BOOLEAN( 'm', "mlockall", &params.mlockall, "Lock all current and future memory"),
OPT_END()
};
@@ -71,15 +78,15 @@ static void *workerfn(void *arg)
unsigned int i;
unsigned long ops = w->ops; /* avoid cacheline bouncing */
- pthread_mutex_lock(&thread_lock);
+ mutex_lock(&thread_lock);
threads_starting--;
if (!threads_starting)
- pthread_cond_signal(&thread_parent);
- pthread_cond_wait(&thread_worker, &thread_lock);
- pthread_mutex_unlock(&thread_lock);
+ cond_signal(&thread_parent);
+ cond_wait(&thread_worker, &thread_lock);
+ mutex_unlock(&thread_lock);
do {
- for (i = 0; i < nfutexes; i++, ops++) {
+ for (i = 0; i < params.nfutexes; i++, ops++) {
/*
* We want the futex calls to fail in order to stress
* the hashing of uaddr and not measure other steps,
@@ -87,7 +94,7 @@ static void *workerfn(void *arg)
* the critical region protected by hb->lock.
*/
ret = futex_wait(&w->futex[i], 1234, NULL, futex_flag);
- if (!silent &&
+ if (!params.silent &&
(!ret || errno != EAGAIN || errno != EWOULDBLOCK))
warn("Non-expected futex return call");
}
@@ -113,19 +120,22 @@ static void print_summary(void)
double stddev = stddev_stats(&throughput_stats);
printf("%sAveraged %ld operations/sec (+- %.2f%%), total secs = %d\n",
- !silent ? "\n" : "", avg, rel_stddev_stats(stddev, avg),
+ !params.silent ? "\n" : "", avg, rel_stddev_stats(stddev, avg),
(int)bench__runtime.tv_sec);
+ futex_print_nbuckets(&params);
}
int bench_futex_hash(int argc, const char **argv)
{
int ret = 0;
- cpu_set_t cpuset;
+ cpu_set_t *cpuset;
struct sigaction act;
unsigned int i;
pthread_attr_t thread_attr;
struct worker *worker = NULL;
struct perf_cpu_map *cpu;
+ int nrcpus;
+ size_t size;
argc = parse_options(argc, argv, options, bench_futex_hash_usage, 0);
if (argc) {
@@ -133,7 +143,7 @@ int bench_futex_hash(int argc, const char **argv)
exit(EXIT_FAILURE);
}
- cpu = perf_cpu_map__new(NULL);
+ cpu = perf_cpu_map__new_online_cpus();
if (!cpu)
goto errmem;
@@ -142,80 +152,96 @@ int bench_futex_hash(int argc, const char **argv)
act.sa_sigaction = toggle_done;
sigaction(SIGINT, &act, NULL);
- if (!nthreads) /* default to the number of CPUs */
- nthreads = cpu->nr;
+ if (params.mlockall) {
+ if (mlockall(MCL_CURRENT | MCL_FUTURE))
+ err(EXIT_FAILURE, "mlockall");
+ }
+
+ if (!params.nthreads) /* default to the number of CPUs */
+ params.nthreads = perf_cpu_map__nr(cpu);
- worker = calloc(nthreads, sizeof(*worker));
+ worker = calloc(params.nthreads, sizeof(*worker));
if (!worker)
goto errmem;
- if (!fshared)
+ if (!params.fshared)
futex_flag = FUTEX_PRIVATE_FLAG;
+ futex_set_nbuckets_param(&params);
printf("Run summary [PID %d]: %d threads, each operating on %d [%s] futexes for %d secs.\n\n",
- getpid(), nthreads, nfutexes, fshared ? "shared":"private", nsecs);
+ getpid(), params.nthreads, params.nfutexes, params.fshared ? "shared":"private", params.runtime);
init_stats(&throughput_stats);
- pthread_mutex_init(&thread_lock, NULL);
- pthread_cond_init(&thread_parent, NULL);
- pthread_cond_init(&thread_worker, NULL);
+ mutex_init(&thread_lock);
+ cond_init(&thread_parent);
+ cond_init(&thread_worker);
- threads_starting = nthreads;
+ threads_starting = params.nthreads;
pthread_attr_init(&thread_attr);
gettimeofday(&bench__start, NULL);
- for (i = 0; i < nthreads; i++) {
+
+ nrcpus = cpu__max_cpu().cpu;
+ cpuset = CPU_ALLOC(nrcpus);
+ BUG_ON(!cpuset);
+ size = CPU_ALLOC_SIZE(nrcpus);
+
+ for (i = 0; i < params.nthreads; i++) {
worker[i].tid = i;
- worker[i].futex = calloc(nfutexes, sizeof(*worker[i].futex));
+ worker[i].futex = calloc(params.nfutexes, sizeof(*worker[i].futex));
if (!worker[i].futex)
goto errmem;
- CPU_ZERO(&cpuset);
- CPU_SET(cpu->map[i % cpu->nr], &cpuset);
+ CPU_ZERO_S(size, cpuset);
- ret = pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset);
- if (ret)
+ CPU_SET_S(perf_cpu_map__cpu(cpu, i % perf_cpu_map__nr(cpu)).cpu, size, cpuset);
+ ret = pthread_attr_setaffinity_np(&thread_attr, size, cpuset);
+ if (ret) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_attr_setaffinity_np");
-
+ }
ret = pthread_create(&worker[i].thread, &thread_attr, workerfn,
(void *)(struct worker *) &worker[i]);
- if (ret)
+ if (ret) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_create");
+ }
}
+ CPU_FREE(cpuset);
pthread_attr_destroy(&thread_attr);
- pthread_mutex_lock(&thread_lock);
+ mutex_lock(&thread_lock);
while (threads_starting)
- pthread_cond_wait(&thread_parent, &thread_lock);
- pthread_cond_broadcast(&thread_worker);
- pthread_mutex_unlock(&thread_lock);
+ cond_wait(&thread_parent, &thread_lock);
+ cond_broadcast(&thread_worker);
+ mutex_unlock(&thread_lock);
- sleep(nsecs);
+ sleep(params.runtime);
toggle_done(0, NULL, NULL);
- for (i = 0; i < nthreads; i++) {
+ for (i = 0; i < params.nthreads; i++) {
ret = pthread_join(worker[i].thread, NULL);
if (ret)
err(EXIT_FAILURE, "pthread_join");
}
/* cleanup & report results */
- pthread_cond_destroy(&thread_parent);
- pthread_cond_destroy(&thread_worker);
- pthread_mutex_destroy(&thread_lock);
+ cond_destroy(&thread_parent);
+ cond_destroy(&thread_worker);
+ mutex_destroy(&thread_lock);
- for (i = 0; i < nthreads; i++) {
+ for (i = 0; i < params.nthreads; i++) {
unsigned long t = bench__runtime.tv_sec > 0 ?
worker[i].ops / bench__runtime.tv_sec : 0;
update_stats(&throughput_stats, t);
- if (!silent) {
- if (nfutexes == 1)
+ if (!params.silent) {
+ if (params.nfutexes == 1)
printf("[thread %2d] futex: %p [ %ld ops/sec ]\n",
worker[i].tid, &worker[i].futex[0], t);
else
printf("[thread %2d] futexes: %p ... %p [ %ld ops/sec ]\n",
worker[i].tid, &worker[i].futex[0],
- &worker[i].futex[nfutexes-1], t);
+ &worker[i].futex[params.nfutexes-1], t);
}
zfree(&worker[i].futex);
diff --git a/tools/perf/bench/futex-lock-pi.c b/tools/perf/bench/futex-lock-pi.c
index bb25d8beb3b8..5144a158512c 100644
--- a/tools/perf/bench/futex-lock-pi.c
+++ b/tools/perf/bench/futex-lock-pi.c
@@ -8,13 +8,13 @@
#include <pthread.h>
#include <signal.h>
+#include "../util/mutex.h"
#include "../util/stat.h"
#include <subcmd/parse-options.h>
#include <linux/compiler.h>
#include <linux/kernel.h>
#include <linux/zalloc.h>
#include <errno.h>
-#include <internal/cpumap.h>
#include <perf/cpumap.h>
#include "bench.h"
#include "futex.h"
@@ -22,6 +22,7 @@
#include <err.h>
#include <stdlib.h>
#include <sys/time.h>
+#include <sys/mman.h>
struct worker {
int tid;
@@ -32,22 +33,27 @@ struct worker {
static u_int32_t global_futex = 0;
static struct worker *worker;
-static unsigned int nsecs = 10;
-static bool silent = false, multi = false;
-static bool done = false, fshared = false;
-static unsigned int nthreads = 0;
+static bool done = false;
static int futex_flag = 0;
-static pthread_mutex_t thread_lock;
+static struct mutex thread_lock;
static unsigned int threads_starting;
static struct stats throughput_stats;
-static pthread_cond_t thread_parent, thread_worker;
+static struct cond thread_parent, thread_worker;
+
+static struct bench_futex_parameters params = {
+ .nbuckets = -1,
+ .runtime = 10,
+};
static const struct option options[] = {
- OPT_UINTEGER('t', "threads", &nthreads, "Specify amount of threads"),
- OPT_UINTEGER('r', "runtime", &nsecs, "Specify runtime (in seconds)"),
- OPT_BOOLEAN( 'M', "multi", &multi, "Use multiple futexes"),
- OPT_BOOLEAN( 's', "silent", &silent, "Silent mode: do not display data/details"),
- OPT_BOOLEAN( 'S', "shared", &fshared, "Use shared futexes instead of private ones"),
+ OPT_INTEGER( 'b', "buckets", &params.nbuckets, "Specify amount of hash buckets"),
+ OPT_BOOLEAN( 'I', "immutable", &params.buckets_immutable, "Make the hash buckets immutable"),
+ OPT_UINTEGER('t', "threads", &params.nthreads, "Specify amount of threads"),
+ OPT_UINTEGER('r', "runtime", &params.runtime, "Specify runtime (in seconds)"),
+ OPT_BOOLEAN( 'M', "multi", &params.multi, "Use multiple futexes"),
+ OPT_BOOLEAN( 's', "silent", &params.silent, "Silent mode: do not display data/details"),
+ OPT_BOOLEAN( 'S', "shared", &params.fshared, "Use shared futexes instead of private ones"),
+ OPT_BOOLEAN( 'm', "mlockall", &params.mlockall, "Lock all current and future memory"),
OPT_END()
};
@@ -62,8 +68,9 @@ static void print_summary(void)
double stddev = stddev_stats(&throughput_stats);
printf("%sAveraged %ld operations/sec (+- %.2f%%), total secs = %d\n",
- !silent ? "\n" : "", avg, rel_stddev_stats(stddev, avg),
+ !params.silent ? "\n" : "", avg, rel_stddev_stats(stddev, avg),
(int)bench__runtime.tv_sec);
+ futex_print_nbuckets(&params);
}
static void toggle_done(int sig __maybe_unused,
@@ -81,12 +88,12 @@ static void *workerfn(void *arg)
struct worker *w = (struct worker *) arg;
unsigned long ops = w->ops;
- pthread_mutex_lock(&thread_lock);
+ mutex_lock(&thread_lock);
threads_starting--;
if (!threads_starting)
- pthread_cond_signal(&thread_parent);
- pthread_cond_wait(&thread_worker, &thread_lock);
- pthread_mutex_unlock(&thread_lock);
+ cond_signal(&thread_parent);
+ cond_wait(&thread_worker, &thread_lock);
+ mutex_unlock(&thread_lock);
do {
int ret;
@@ -94,7 +101,7 @@ static void *workerfn(void *arg)
ret = futex_lock_pi(w->futex, NULL, futex_flag);
if (ret) { /* handle lock acquisition */
- if (!silent)
+ if (!params.silent)
warn("thread %d: Could not lock pi-lock for %p (%d)",
w->tid, w->futex, ret);
if (done)
@@ -105,7 +112,7 @@ static void *workerfn(void *arg)
usleep(1);
ret = futex_unlock_pi(w->futex, futex_flag);
- if (ret && !silent)
+ if (ret && !params.silent)
warn("thread %d: Could not unlock pi-lock for %p (%d)",
w->tid, w->futex, ret);
ops++; /* account for thread's share of work */
@@ -115,33 +122,47 @@ static void *workerfn(void *arg)
return NULL;
}
-static void create_threads(struct worker *w, pthread_attr_t thread_attr,
- struct perf_cpu_map *cpu)
+static void create_threads(struct worker *w, struct perf_cpu_map *cpu)
{
- cpu_set_t cpuset;
+ cpu_set_t *cpuset;
unsigned int i;
+ int nrcpus = cpu__max_cpu().cpu;
+ size_t size;
+
+ threads_starting = params.nthreads;
+
+ cpuset = CPU_ALLOC(nrcpus);
+ BUG_ON(!cpuset);
+ size = CPU_ALLOC_SIZE(nrcpus);
- threads_starting = nthreads;
+ for (i = 0; i < params.nthreads; i++) {
+ pthread_attr_t thread_attr;
- for (i = 0; i < nthreads; i++) {
+ pthread_attr_init(&thread_attr);
worker[i].tid = i;
- if (multi) {
+ if (params.multi) {
worker[i].futex = calloc(1, sizeof(u_int32_t));
if (!worker[i].futex)
err(EXIT_FAILURE, "calloc");
} else
worker[i].futex = &global_futex;
- CPU_ZERO(&cpuset);
- CPU_SET(cpu->map[i % cpu->nr], &cpuset);
+ CPU_ZERO_S(size, cpuset);
+ CPU_SET_S(perf_cpu_map__cpu(cpu, i % perf_cpu_map__nr(cpu)).cpu, size, cpuset);
- if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset))
+ if (pthread_attr_setaffinity_np(&thread_attr, size, cpuset)) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_attr_setaffinity_np");
+ }
- if (pthread_create(&w[i].thread, &thread_attr, workerfn, &worker[i]))
+ if (pthread_create(&w[i].thread, &thread_attr, workerfn, &worker[i])) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_create");
+ }
+ pthread_attr_destroy(&thread_attr);
}
+ CPU_FREE(cpuset);
}
int bench_futex_lock_pi(int argc, const char **argv)
@@ -149,14 +170,13 @@ int bench_futex_lock_pi(int argc, const char **argv)
int ret = 0;
unsigned int i;
struct sigaction act;
- pthread_attr_t thread_attr;
struct perf_cpu_map *cpu;
argc = parse_options(argc, argv, options, bench_futex_lock_pi_usage, 0);
if (argc)
goto err;
- cpu = perf_cpu_map__new(NULL);
+ cpu = perf_cpu_map__new_online_cpus();
if (!cpu)
err(EXIT_FAILURE, "calloc");
@@ -165,67 +185,72 @@ int bench_futex_lock_pi(int argc, const char **argv)
act.sa_sigaction = toggle_done;
sigaction(SIGINT, &act, NULL);
- if (!nthreads)
- nthreads = cpu->nr;
+ if (params.mlockall) {
+ if (mlockall(MCL_CURRENT | MCL_FUTURE))
+ err(EXIT_FAILURE, "mlockall");
+ }
+
+ if (!params.nthreads)
+ params.nthreads = perf_cpu_map__nr(cpu);
- worker = calloc(nthreads, sizeof(*worker));
+ worker = calloc(params.nthreads, sizeof(*worker));
if (!worker)
err(EXIT_FAILURE, "calloc");
- if (!fshared)
+ if (!params.fshared)
futex_flag = FUTEX_PRIVATE_FLAG;
printf("Run summary [PID %d]: %d threads doing pi lock/unlock pairing for %d secs.\n\n",
- getpid(), nthreads, nsecs);
+ getpid(), params.nthreads, params.runtime);
init_stats(&throughput_stats);
- pthread_mutex_init(&thread_lock, NULL);
- pthread_cond_init(&thread_parent, NULL);
- pthread_cond_init(&thread_worker, NULL);
+ mutex_init(&thread_lock);
+ cond_init(&thread_parent);
+ cond_init(&thread_worker);
+ futex_set_nbuckets_param(&params);
- threads_starting = nthreads;
- pthread_attr_init(&thread_attr);
+ threads_starting = params.nthreads;
gettimeofday(&bench__start, NULL);
- create_threads(worker, thread_attr, cpu);
- pthread_attr_destroy(&thread_attr);
+ create_threads(worker, cpu);
- pthread_mutex_lock(&thread_lock);
+ mutex_lock(&thread_lock);
while (threads_starting)
- pthread_cond_wait(&thread_parent, &thread_lock);
- pthread_cond_broadcast(&thread_worker);
- pthread_mutex_unlock(&thread_lock);
+ cond_wait(&thread_parent, &thread_lock);
+ cond_broadcast(&thread_worker);
+ mutex_unlock(&thread_lock);
- sleep(nsecs);
+ sleep(params.runtime);
toggle_done(0, NULL, NULL);
- for (i = 0; i < nthreads; i++) {
+ for (i = 0; i < params.nthreads; i++) {
ret = pthread_join(worker[i].thread, NULL);
if (ret)
err(EXIT_FAILURE, "pthread_join");
}
/* cleanup & report results */
- pthread_cond_destroy(&thread_parent);
- pthread_cond_destroy(&thread_worker);
- pthread_mutex_destroy(&thread_lock);
+ cond_destroy(&thread_parent);
+ cond_destroy(&thread_worker);
+ mutex_destroy(&thread_lock);
- for (i = 0; i < nthreads; i++) {
+ for (i = 0; i < params.nthreads; i++) {
unsigned long t = bench__runtime.tv_sec > 0 ?
worker[i].ops / bench__runtime.tv_sec : 0;
update_stats(&throughput_stats, t);
- if (!silent)
+ if (!params.silent)
printf("[thread %3d] futex: %p [ %ld ops/sec ]\n",
worker[i].tid, worker[i].futex, t);
- if (multi)
+ if (params.multi)
zfree(&worker[i].futex);
}
print_summary();
free(worker);
+ perf_cpu_map__put(cpu);
return ret;
err:
usage_with_options(bench_futex_lock_pi_usage, options);
diff --git a/tools/perf/bench/futex-requeue.c b/tools/perf/bench/futex-requeue.c
index 7a15c2e61022..a2f91ee1950b 100644
--- a/tools/perf/bench/futex-requeue.c
+++ b/tools/perf/bench/futex-requeue.c
@@ -6,7 +6,8 @@
* on futex2, N at a time.
*
* This program is particularly useful to measure the latency of nthread
- * requeues without waking up any tasks -- thus mimicking a regular futex_wait.
+ * requeues without waking up any tasks (in the non-pi case) -- thus
+ * mimicking a regular futex_wait.
*/
/* For the CLR_() macros */
@@ -14,13 +15,13 @@
#include <pthread.h>
#include <signal.h>
+#include "../util/mutex.h"
#include "../util/stat.h"
#include <subcmd/parse-options.h>
#include <linux/compiler.h>
#include <linux/kernel.h>
#include <linux/time64.h>
#include <errno.h>
-#include <internal/cpumap.h>
#include <perf/cpumap.h>
#include "bench.h"
#include "futex.h"
@@ -28,28 +29,38 @@
#include <err.h>
#include <stdlib.h>
#include <sys/time.h>
+#include <sys/mman.h>
static u_int32_t futex1 = 0, futex2 = 0;
-/*
- * How many tasks to requeue at a time.
- * Default to 1 in order to make the kernel work more.
- */
-static unsigned int nrequeue = 1;
-
static pthread_t *worker;
-static bool done = false, silent = false, fshared = false;
-static pthread_mutex_t thread_lock;
-static pthread_cond_t thread_parent, thread_worker;
+static bool done = false;
+static struct mutex thread_lock;
+static struct cond thread_parent, thread_worker;
static struct stats requeuetime_stats, requeued_stats;
-static unsigned int threads_starting, nthreads = 0;
+static unsigned int threads_starting;
static int futex_flag = 0;
+static struct bench_futex_parameters params = {
+ .nbuckets = -1,
+ /*
+ * How many tasks to requeue at a time.
+ * Default to 1 in order to make the kernel work more.
+ */
+ .nrequeue = 1,
+};
+
static const struct option options[] = {
- OPT_UINTEGER('t', "threads", &nthreads, "Specify amount of threads"),
- OPT_UINTEGER('q', "nrequeue", &nrequeue, "Specify amount of threads to requeue at once"),
- OPT_BOOLEAN( 's', "silent", &silent, "Silent mode: do not display data/details"),
- OPT_BOOLEAN( 'S', "shared", &fshared, "Use shared futexes instead of private ones"),
+ OPT_INTEGER( 'b', "buckets", &params.nbuckets, "Specify amount of hash buckets"),
+ OPT_BOOLEAN( 'I', "immutable", &params.buckets_immutable, "Make the hash buckets immutable"),
+ OPT_UINTEGER('t', "threads", &params.nthreads, "Specify amount of threads"),
+ OPT_UINTEGER('q', "nrequeue", &params.nrequeue, "Specify amount of threads to requeue at once"),
+ OPT_BOOLEAN( 's', "silent", &params.silent, "Silent mode: do not display data/details"),
+ OPT_BOOLEAN( 'S', "shared", &params.fshared, "Use shared futexes instead of private ones"),
+ OPT_BOOLEAN( 'm', "mlockall", &params.mlockall, "Lock all current and future memory"),
+ OPT_BOOLEAN( 'B', "broadcast", &params.broadcast, "Requeue all threads at once"),
+ OPT_BOOLEAN( 'p', "pi", &params.pi, "Use PI-aware variants of FUTEX_CMP_REQUEUE"),
+
OPT_END()
};
@@ -66,43 +77,87 @@ static void print_summary(void)
printf("Requeued %d of %d threads in %.4f ms (+-%.2f%%)\n",
requeued_avg,
- nthreads,
+ params.nthreads,
requeuetime_avg / USEC_PER_MSEC,
rel_stddev_stats(requeuetime_stddev, requeuetime_avg));
+ futex_print_nbuckets(&params);
}
static void *workerfn(void *arg __maybe_unused)
{
- pthread_mutex_lock(&thread_lock);
+ int ret;
+
+ mutex_lock(&thread_lock);
threads_starting--;
if (!threads_starting)
- pthread_cond_signal(&thread_parent);
- pthread_cond_wait(&thread_worker, &thread_lock);
- pthread_mutex_unlock(&thread_lock);
+ cond_signal(&thread_parent);
+ cond_wait(&thread_worker, &thread_lock);
+ mutex_unlock(&thread_lock);
+
+ while (1) {
+ if (!params.pi) {
+ ret = futex_wait(&futex1, 0, NULL, futex_flag);
+ if (!ret)
+ break;
+
+ if (ret && errno != EAGAIN) {
+ if (!params.silent)
+ warnx("futex_wait");
+ break;
+ }
+ } else {
+ ret = futex_wait_requeue_pi(&futex1, 0, &futex2,
+ NULL, futex_flag);
+ if (!ret) {
+ /* got the lock at futex2 */
+ futex_unlock_pi(&futex2, futex_flag);
+ break;
+ }
+
+ if (ret && errno != EAGAIN) {
+ if (!params.silent)
+ warnx("futex_wait_requeue_pi");
+ break;
+ }
+ }
+ }
- futex_wait(&futex1, 0, NULL, futex_flag);
return NULL;
}
-static void block_threads(pthread_t *w,
- pthread_attr_t thread_attr, struct perf_cpu_map *cpu)
+static void block_threads(pthread_t *w, struct perf_cpu_map *cpu)
{
- cpu_set_t cpuset;
+ cpu_set_t *cpuset;
unsigned int i;
+ int nrcpus = cpu__max_cpu().cpu;
+ size_t size;
+
+ threads_starting = params.nthreads;
- threads_starting = nthreads;
+ cpuset = CPU_ALLOC(nrcpus);
+ BUG_ON(!cpuset);
+ size = CPU_ALLOC_SIZE(nrcpus);
/* create and block all threads */
- for (i = 0; i < nthreads; i++) {
- CPU_ZERO(&cpuset);
- CPU_SET(cpu->map[i % cpu->nr], &cpuset);
+ for (i = 0; i < params.nthreads; i++) {
+ pthread_attr_t thread_attr;
- if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset))
+ pthread_attr_init(&thread_attr);
+ CPU_ZERO_S(size, cpuset);
+ CPU_SET_S(perf_cpu_map__cpu(cpu, i % perf_cpu_map__nr(cpu)).cpu, size, cpuset);
+
+ if (pthread_attr_setaffinity_np(&thread_attr, size, cpuset)) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_attr_setaffinity_np");
+ }
- if (pthread_create(&w[i], &thread_attr, workerfn, NULL))
+ if (pthread_create(&w[i], &thread_attr, workerfn, NULL)) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_create");
+ }
+ pthread_attr_destroy(&thread_attr);
}
+ CPU_FREE(cpuset);
}
static void toggle_done(int sig __maybe_unused,
@@ -117,14 +172,13 @@ int bench_futex_requeue(int argc, const char **argv)
int ret = 0;
unsigned int i, j;
struct sigaction act;
- pthread_attr_t thread_attr;
struct perf_cpu_map *cpu;
argc = parse_options(argc, argv, options, bench_futex_requeue_usage, 0);
if (argc)
goto err;
- cpu = perf_cpu_map__new(NULL);
+ cpu = perf_cpu_map__new_online_cpus();
if (!cpu)
err(EXIT_FAILURE, "cpu_map__new");
@@ -133,55 +187,83 @@ int bench_futex_requeue(int argc, const char **argv)
act.sa_sigaction = toggle_done;
sigaction(SIGINT, &act, NULL);
- if (!nthreads)
- nthreads = cpu->nr;
+ if (params.mlockall) {
+ if (mlockall(MCL_CURRENT | MCL_FUTURE))
+ err(EXIT_FAILURE, "mlockall");
+ }
+
+ if (!params.nthreads)
+ params.nthreads = perf_cpu_map__nr(cpu);
- worker = calloc(nthreads, sizeof(*worker));
+ worker = calloc(params.nthreads, sizeof(*worker));
if (!worker)
err(EXIT_FAILURE, "calloc");
- if (!fshared)
+ if (!params.fshared)
futex_flag = FUTEX_PRIVATE_FLAG;
- if (nrequeue > nthreads)
- nrequeue = nthreads;
+ if (params.nrequeue > params.nthreads)
+ params.nrequeue = params.nthreads;
- printf("Run summary [PID %d]: Requeuing %d threads (from [%s] %p to %p), "
- "%d at a time.\n\n", getpid(), nthreads,
- fshared ? "shared":"private", &futex1, &futex2, nrequeue);
+ if (params.broadcast)
+ params.nrequeue = params.nthreads;
+
+ futex_set_nbuckets_param(&params);
+
+ printf("Run summary [PID %d]: Requeuing %d threads (from [%s] %p to %s%p), "
+ "%d at a time.\n\n", getpid(), params.nthreads,
+ params.fshared ? "shared":"private", &futex1,
+ params.pi ? "PI ": "", &futex2, params.nrequeue);
init_stats(&requeued_stats);
init_stats(&requeuetime_stats);
- pthread_attr_init(&thread_attr);
- pthread_mutex_init(&thread_lock, NULL);
- pthread_cond_init(&thread_parent, NULL);
- pthread_cond_init(&thread_worker, NULL);
+ mutex_init(&thread_lock);
+ cond_init(&thread_parent);
+ cond_init(&thread_worker);
for (j = 0; j < bench_repeat && !done; j++) {
- unsigned int nrequeued = 0;
+ unsigned int nrequeued = 0, wakeups = 0;
struct timeval start, end, runtime;
/* create, launch & block all threads */
- block_threads(worker, thread_attr, cpu);
+ block_threads(worker, cpu);
/* make sure all threads are already blocked */
- pthread_mutex_lock(&thread_lock);
+ mutex_lock(&thread_lock);
while (threads_starting)
- pthread_cond_wait(&thread_parent, &thread_lock);
- pthread_cond_broadcast(&thread_worker);
- pthread_mutex_unlock(&thread_lock);
+ cond_wait(&thread_parent, &thread_lock);
+ cond_broadcast(&thread_worker);
+ mutex_unlock(&thread_lock);
usleep(100000);
/* Ok, all threads are patiently blocked, start requeueing */
gettimeofday(&start, NULL);
- while (nrequeued < nthreads) {
+ while (nrequeued < params.nthreads) {
+ int r;
+
/*
- * Do not wakeup any tasks blocked on futex1, allowing
- * us to really measure futex_wait functionality.
+ * For the regular non-pi case, do not wakeup any tasks
+ * blocked on futex1, allowing us to really measure
+ * futex_wait functionality. For the PI case the first
+ * waiter is always awoken.
*/
- nrequeued += futex_cmp_requeue(&futex1, 0, &futex2, 0,
- nrequeue, futex_flag);
+ if (!params.pi) {
+ r = futex_cmp_requeue(&futex1, 0, &futex2, 0,
+ params.nrequeue,
+ futex_flag);
+ } else {
+ r = futex_cmp_requeue_pi(&futex1, 0, &futex2,
+ params.nrequeue,
+ futex_flag);
+ wakeups++; /* assume no error */
+ }
+
+ if (r < 0)
+ err(EXIT_FAILURE, "couldn't requeue from %p to %p",
+ &futex1, &futex2);
+
+ nrequeued += r;
}
gettimeofday(&end, NULL);
@@ -190,17 +272,32 @@ int bench_futex_requeue(int argc, const char **argv)
update_stats(&requeued_stats, nrequeued);
update_stats(&requeuetime_stats, runtime.tv_usec);
- if (!silent) {
- printf("[Run %d]: Requeued %d of %d threads in %.4f ms\n",
- j + 1, nrequeued, nthreads, runtime.tv_usec / (double)USEC_PER_MSEC);
+ if (!params.silent) {
+ if (!params.pi)
+ printf("[Run %d]: Requeued %d of %d threads in "
+ "%.4f ms\n", j + 1, nrequeued,
+ params.nthreads,
+ runtime.tv_usec / (double)USEC_PER_MSEC);
+ else {
+ nrequeued -= wakeups;
+ printf("[Run %d]: Awoke and Requeued (%d+%d) of "
+ "%d threads in %.4f ms\n",
+ j + 1, wakeups, nrequeued,
+ params.nthreads,
+ runtime.tv_usec / (double)USEC_PER_MSEC);
+ }
+
}
- /* everybody should be blocked on futex2, wake'em up */
- nrequeued = futex_wake(&futex2, nrequeued, futex_flag);
- if (nthreads != nrequeued)
- warnx("couldn't wakeup all tasks (%d/%d)", nrequeued, nthreads);
+ if (!params.pi) {
+ /* everybody should be blocked on futex2, wake'em up */
+ nrequeued = futex_wake(&futex2, nrequeued, futex_flag);
+ if (params.nthreads != nrequeued)
+ warnx("couldn't wakeup all tasks (%d/%d)",
+ nrequeued, params.nthreads);
+ }
- for (i = 0; i < nthreads; i++) {
+ for (i = 0; i < params.nthreads; i++) {
ret = pthread_join(worker[i], NULL);
if (ret)
err(EXIT_FAILURE, "pthread_join");
@@ -208,14 +305,14 @@ int bench_futex_requeue(int argc, const char **argv)
}
/* cleanup & report results */
- pthread_cond_destroy(&thread_parent);
- pthread_cond_destroy(&thread_worker);
- pthread_mutex_destroy(&thread_lock);
- pthread_attr_destroy(&thread_attr);
+ cond_destroy(&thread_parent);
+ cond_destroy(&thread_worker);
+ mutex_destroy(&thread_lock);
print_summary();
free(worker);
+ perf_cpu_map__put(cpu);
return ret;
err:
usage_with_options(bench_futex_requeue_usage, options);
diff --git a/tools/perf/bench/futex-wake-parallel.c b/tools/perf/bench/futex-wake-parallel.c
index cd2b81a845ac..ee66482c29fd 100644
--- a/tools/perf/bench/futex-wake-parallel.c
+++ b/tools/perf/bench/futex-wake-parallel.c
@@ -10,6 +10,7 @@
#include "bench.h"
#include <linux/compiler.h>
#include "../util/debug.h"
+#include "../util/mutex.h"
#ifndef HAVE_PTHREAD_BARRIER
int bench_futex_wake_parallel(int argc __maybe_unused, const char **argv __maybe_unused)
@@ -29,12 +30,12 @@ int bench_futex_wake_parallel(int argc __maybe_unused, const char **argv __maybe
#include <linux/time64.h>
#include <errno.h>
#include "futex.h"
-#include <internal/cpumap.h>
#include <perf/cpumap.h>
#include <err.h>
#include <stdlib.h>
#include <sys/time.h>
+#include <sys/mman.h>
struct thread_data {
pthread_t worker;
@@ -48,20 +49,27 @@ static unsigned int nwakes = 1;
static u_int32_t futex = 0;
static pthread_t *blocked_worker;
-static bool done = false, silent = false, fshared = false;
-static unsigned int nblocked_threads = 0, nwaking_threads = 0;
-static pthread_mutex_t thread_lock;
-static pthread_cond_t thread_parent, thread_worker;
+static bool done = false;
+static struct mutex thread_lock;
+static struct cond thread_parent, thread_worker;
static pthread_barrier_t barrier;
static struct stats waketime_stats, wakeup_stats;
static unsigned int threads_starting;
static int futex_flag = 0;
+static struct bench_futex_parameters params = {
+ .nbuckets = -1,
+};
+
static const struct option options[] = {
- OPT_UINTEGER('t', "threads", &nblocked_threads, "Specify amount of threads"),
- OPT_UINTEGER('w', "nwakers", &nwaking_threads, "Specify amount of waking threads"),
- OPT_BOOLEAN( 's', "silent", &silent, "Silent mode: do not display data/details"),
- OPT_BOOLEAN( 'S', "shared", &fshared, "Use shared futexes instead of private ones"),
+ OPT_INTEGER( 'b', "buckets", &params.nbuckets, "Specify amount of hash buckets"),
+ OPT_BOOLEAN( 'I', "immutable", &params.buckets_immutable, "Make the hash buckets immutable"),
+ OPT_UINTEGER('t', "threads", &params.nthreads, "Specify amount of threads"),
+ OPT_UINTEGER('w', "nwakers", &params.nwakes, "Specify amount of waking threads"),
+ OPT_BOOLEAN( 's', "silent", &params.silent, "Silent mode: do not display data/details"),
+ OPT_BOOLEAN( 'S', "shared", &params.fshared, "Use shared futexes instead of private ones"),
+ OPT_BOOLEAN( 'm', "mlockall", &params.mlockall, "Lock all current and future memory"),
+
OPT_END()
};
@@ -91,16 +99,18 @@ static void *waking_workerfn(void *arg)
return NULL;
}
-static void wakeup_threads(struct thread_data *td, pthread_attr_t thread_attr)
+static void wakeup_threads(struct thread_data *td)
{
unsigned int i;
+ pthread_attr_t thread_attr;
+ pthread_attr_init(&thread_attr);
pthread_attr_setdetachstate(&thread_attr, PTHREAD_CREATE_JOINABLE);
- pthread_barrier_init(&barrier, NULL, nwaking_threads + 1);
+ pthread_barrier_init(&barrier, NULL, params.nwakes + 1);
/* create and block all threads */
- for (i = 0; i < nwaking_threads; i++) {
+ for (i = 0; i < params.nwakes; i++) {
/*
* Thread creation order will impact per-thread latency
* as it will affect the order to acquire the hb spinlock.
@@ -113,21 +123,22 @@ static void wakeup_threads(struct thread_data *td, pthread_attr_t thread_attr)
pthread_barrier_wait(&barrier);
- for (i = 0; i < nwaking_threads; i++)
+ for (i = 0; i < params.nwakes; i++)
if (pthread_join(td[i].worker, NULL))
err(EXIT_FAILURE, "pthread_join");
pthread_barrier_destroy(&barrier);
+ pthread_attr_destroy(&thread_attr);
}
static void *blocked_workerfn(void *arg __maybe_unused)
{
- pthread_mutex_lock(&thread_lock);
+ mutex_lock(&thread_lock);
threads_starting--;
if (!threads_starting)
- pthread_cond_signal(&thread_parent);
- pthread_cond_wait(&thread_worker, &thread_lock);
- pthread_mutex_unlock(&thread_lock);
+ cond_signal(&thread_parent);
+ cond_wait(&thread_worker, &thread_lock);
+ mutex_unlock(&thread_lock);
while (1) { /* handle spurious wakeups */
if (futex_wait(&futex, 0, NULL, futex_flag) != EINTR)
@@ -138,25 +149,39 @@ static void *blocked_workerfn(void *arg __maybe_unused)
return NULL;
}
-static void block_threads(pthread_t *w, pthread_attr_t thread_attr,
- struct perf_cpu_map *cpu)
+static void block_threads(pthread_t *w, struct perf_cpu_map *cpu)
{
- cpu_set_t cpuset;
+ cpu_set_t *cpuset;
unsigned int i;
+ int nrcpus = cpu__max_cpu().cpu;
+ size_t size;
- threads_starting = nblocked_threads;
+ threads_starting = params.nthreads;
+
+ cpuset = CPU_ALLOC(nrcpus);
+ BUG_ON(!cpuset);
+ size = CPU_ALLOC_SIZE(nrcpus);
/* create and block all threads */
- for (i = 0; i < nblocked_threads; i++) {
- CPU_ZERO(&cpuset);
- CPU_SET(cpu->map[i % cpu->nr], &cpuset);
+ for (i = 0; i < params.nthreads; i++) {
+ pthread_attr_t thread_attr;
+
+ pthread_attr_init(&thread_attr);
+ CPU_ZERO_S(size, cpuset);
+ CPU_SET_S(perf_cpu_map__cpu(cpu, i % perf_cpu_map__nr(cpu)).cpu, size, cpuset);
- if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset))
+ if (pthread_attr_setaffinity_np(&thread_attr, size, cpuset)) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_attr_setaffinity_np");
+ }
- if (pthread_create(&w[i], &thread_attr, blocked_workerfn, NULL))
+ if (pthread_create(&w[i], &thread_attr, blocked_workerfn, NULL)) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_create");
+ }
+ pthread_attr_destroy(&thread_attr);
}
+ CPU_FREE(cpuset);
}
static void print_run(struct thread_data *waking_worker, unsigned int run_num)
@@ -168,7 +193,7 @@ static void print_run(struct thread_data *waking_worker, unsigned int run_num)
init_stats(&__wakeup_stats);
init_stats(&__waketime_stats);
- for (i = 0; i < nwaking_threads; i++) {
+ for (i = 0; i < params.nwakes; i++) {
update_stats(&__waketime_stats, waking_worker[i].runtime.tv_usec);
update_stats(&__wakeup_stats, waking_worker[i].nwoken);
}
@@ -179,7 +204,7 @@ static void print_run(struct thread_data *waking_worker, unsigned int run_num)
printf("[Run %d]: Avg per-thread latency (waking %d/%d threads) "
"in %.4f ms (+-%.2f%%)\n", run_num + 1, wakeup_avg,
- nblocked_threads, waketime_avg / USEC_PER_MSEC,
+ params.nthreads, waketime_avg / USEC_PER_MSEC,
rel_stddev_stats(waketime_stddev, waketime_avg));
}
@@ -194,9 +219,10 @@ static void print_summary(void)
printf("Avg per-thread latency (waking %d/%d threads) in %.4f ms (+-%.2f%%)\n",
wakeup_avg,
- nblocked_threads,
+ params.nthreads,
waketime_avg / USEC_PER_MSEC,
rel_stddev_stats(waketime_stddev, waketime_avg));
+ futex_print_nbuckets(&params);
}
@@ -204,7 +230,7 @@ static void do_run_stats(struct thread_data *waking_worker)
{
unsigned int i;
- for (i = 0; i < nwaking_threads; i++) {
+ for (i = 0; i < params.nwakes; i++) {
update_stats(&waketime_stats, waking_worker[i].runtime.tv_usec);
update_stats(&wakeup_stats, waking_worker[i].nwoken);
}
@@ -223,7 +249,6 @@ int bench_futex_wake_parallel(int argc, const char **argv)
int ret = 0;
unsigned int i, j;
struct sigaction act;
- pthread_attr_t thread_attr;
struct thread_data *waking_worker;
struct perf_cpu_map *cpu;
@@ -239,87 +264,94 @@ int bench_futex_wake_parallel(int argc, const char **argv)
act.sa_sigaction = toggle_done;
sigaction(SIGINT, &act, NULL);
- cpu = perf_cpu_map__new(NULL);
+ if (params.mlockall) {
+ if (mlockall(MCL_CURRENT | MCL_FUTURE))
+ err(EXIT_FAILURE, "mlockall");
+ }
+
+ cpu = perf_cpu_map__new_online_cpus();
if (!cpu)
err(EXIT_FAILURE, "calloc");
- if (!nblocked_threads)
- nblocked_threads = cpu->nr;
+ if (!params.nthreads)
+ params.nthreads = perf_cpu_map__nr(cpu);
/* some sanity checks */
- if (nwaking_threads > nblocked_threads || !nwaking_threads)
- nwaking_threads = nblocked_threads;
+ if (params.nwakes > params.nthreads ||
+ !params.nwakes)
+ params.nwakes = params.nthreads;
- if (nblocked_threads % nwaking_threads)
+ if (params.nthreads % params.nwakes)
errx(EXIT_FAILURE, "Must be perfectly divisible");
/*
* Each thread will wakeup nwakes tasks in
* a single futex_wait call.
*/
- nwakes = nblocked_threads/nwaking_threads;
+ nwakes = params.nthreads/params.nwakes;
- blocked_worker = calloc(nblocked_threads, sizeof(*blocked_worker));
+ blocked_worker = calloc(params.nthreads, sizeof(*blocked_worker));
if (!blocked_worker)
err(EXIT_FAILURE, "calloc");
- if (!fshared)
+ if (!params.fshared)
futex_flag = FUTEX_PRIVATE_FLAG;
+ futex_set_nbuckets_param(&params);
+
printf("Run summary [PID %d]: blocking on %d threads (at [%s] "
"futex %p), %d threads waking up %d at a time.\n\n",
- getpid(), nblocked_threads, fshared ? "shared":"private",
- &futex, nwaking_threads, nwakes);
+ getpid(), params.nthreads, params.fshared ? "shared":"private",
+ &futex, params.nwakes, nwakes);
init_stats(&wakeup_stats);
init_stats(&waketime_stats);
- pthread_attr_init(&thread_attr);
- pthread_mutex_init(&thread_lock, NULL);
- pthread_cond_init(&thread_parent, NULL);
- pthread_cond_init(&thread_worker, NULL);
+ mutex_init(&thread_lock);
+ cond_init(&thread_parent);
+ cond_init(&thread_worker);
for (j = 0; j < bench_repeat && !done; j++) {
- waking_worker = calloc(nwaking_threads, sizeof(*waking_worker));
+ waking_worker = calloc(params.nwakes, sizeof(*waking_worker));
if (!waking_worker)
err(EXIT_FAILURE, "calloc");
/* create, launch & block all threads */
- block_threads(blocked_worker, thread_attr, cpu);
+ block_threads(blocked_worker, cpu);
/* make sure all threads are already blocked */
- pthread_mutex_lock(&thread_lock);
+ mutex_lock(&thread_lock);
while (threads_starting)
- pthread_cond_wait(&thread_parent, &thread_lock);
- pthread_cond_broadcast(&thread_worker);
- pthread_mutex_unlock(&thread_lock);
+ cond_wait(&thread_parent, &thread_lock);
+ cond_broadcast(&thread_worker);
+ mutex_unlock(&thread_lock);
- usleep(100000);
+ usleep(200000);
/* Ok, all threads are patiently blocked, start waking folks up */
- wakeup_threads(waking_worker, thread_attr);
+ wakeup_threads(waking_worker);
- for (i = 0; i < nblocked_threads; i++) {
+ for (i = 0; i < params.nthreads; i++) {
ret = pthread_join(blocked_worker[i], NULL);
if (ret)
err(EXIT_FAILURE, "pthread_join");
}
do_run_stats(waking_worker);
- if (!silent)
+ if (!params.silent)
print_run(waking_worker, j);
free(waking_worker);
}
/* cleanup & report results */
- pthread_cond_destroy(&thread_parent);
- pthread_cond_destroy(&thread_worker);
- pthread_mutex_destroy(&thread_lock);
- pthread_attr_destroy(&thread_attr);
+ cond_destroy(&thread_parent);
+ cond_destroy(&thread_worker);
+ mutex_destroy(&thread_lock);
print_summary();
free(blocked_worker);
+ perf_cpu_map__put(cpu);
return ret;
}
#endif /* HAVE_PTHREAD_BARRIER */
diff --git a/tools/perf/bench/futex-wake.c b/tools/perf/bench/futex-wake.c
index 2dfcef3e371e..8d6107f7cd94 100644
--- a/tools/perf/bench/futex-wake.c
+++ b/tools/perf/bench/futex-wake.c
@@ -14,13 +14,13 @@
#include <pthread.h>
#include <signal.h>
+#include "../util/mutex.h"
#include "../util/stat.h"
#include <subcmd/parse-options.h>
#include <linux/compiler.h>
#include <linux/kernel.h>
#include <linux/time64.h>
#include <errno.h>
-#include <internal/cpumap.h>
#include <perf/cpumap.h>
#include "bench.h"
#include "futex.h"
@@ -28,29 +28,37 @@
#include <err.h>
#include <stdlib.h>
#include <sys/time.h>
+#include <sys/mman.h>
/* all threads will block on the same futex */
static u_int32_t futex1 = 0;
-/*
- * How many wakeups to do at a time.
- * Default to 1 in order to make the kernel work more.
- */
-static unsigned int nwakes = 1;
-
-pthread_t *worker;
-static bool done = false, silent = false, fshared = false;
-static pthread_mutex_t thread_lock;
-static pthread_cond_t thread_parent, thread_worker;
+static pthread_t *worker;
+static bool done = false;
+static struct mutex thread_lock;
+static struct cond thread_parent, thread_worker;
static struct stats waketime_stats, wakeup_stats;
-static unsigned int threads_starting, nthreads = 0;
+static unsigned int threads_starting;
static int futex_flag = 0;
+static struct bench_futex_parameters params = {
+ .nbuckets = -1,
+ /*
+ * How many wakeups to do at a time.
+ * Default to 1 in order to make the kernel work more.
+ */
+ .nwakes = 1,
+};
+
static const struct option options[] = {
- OPT_UINTEGER('t', "threads", &nthreads, "Specify amount of threads"),
- OPT_UINTEGER('w', "nwakes", &nwakes, "Specify amount of threads to wake at once"),
- OPT_BOOLEAN( 's', "silent", &silent, "Silent mode: do not display data/details"),
- OPT_BOOLEAN( 'S', "shared", &fshared, "Use shared futexes instead of private ones"),
+ OPT_INTEGER( 'b', "buckets", &params.nbuckets, "Specify amount of hash buckets"),
+ OPT_BOOLEAN( 'I', "immutable", &params.buckets_immutable, "Make the hash buckets immutable"),
+ OPT_UINTEGER('t', "threads", &params.nthreads, "Specify amount of threads"),
+ OPT_UINTEGER('w', "nwakes", &params.nwakes, "Specify amount of threads to wake at once"),
+ OPT_BOOLEAN( 's', "silent", &params.silent, "Silent mode: do not display data/details"),
+ OPT_BOOLEAN( 'S', "shared", &params.fshared, "Use shared futexes instead of private ones"),
+ OPT_BOOLEAN( 'm', "mlockall", &params.mlockall, "Lock all current and future memory"),
+
OPT_END()
};
@@ -61,12 +69,12 @@ static const char * const bench_futex_wake_usage[] = {
static void *workerfn(void *arg __maybe_unused)
{
- pthread_mutex_lock(&thread_lock);
+ mutex_lock(&thread_lock);
threads_starting--;
if (!threads_starting)
- pthread_cond_signal(&thread_parent);
- pthread_cond_wait(&thread_worker, &thread_lock);
- pthread_mutex_unlock(&thread_lock);
+ cond_signal(&thread_parent);
+ cond_wait(&thread_worker, &thread_lock);
+ mutex_unlock(&thread_lock);
while (1) {
if (futex_wait(&futex1, 0, NULL, futex_flag) != EINTR)
@@ -85,30 +93,44 @@ static void print_summary(void)
printf("Wokeup %d of %d threads in %.4f ms (+-%.2f%%)\n",
wakeup_avg,
- nthreads,
+ params.nthreads,
waketime_avg / USEC_PER_MSEC,
rel_stddev_stats(waketime_stddev, waketime_avg));
+ futex_print_nbuckets(&params);
}
-static void block_threads(pthread_t *w,
- pthread_attr_t thread_attr, struct perf_cpu_map *cpu)
+static void block_threads(pthread_t *w, struct perf_cpu_map *cpu)
{
- cpu_set_t cpuset;
+ cpu_set_t *cpuset;
unsigned int i;
+ size_t size;
+ int nrcpus = cpu__max_cpu().cpu;
+ threads_starting = params.nthreads;
- threads_starting = nthreads;
+ cpuset = CPU_ALLOC(nrcpus);
+ BUG_ON(!cpuset);
+ size = CPU_ALLOC_SIZE(nrcpus);
/* create and block all threads */
- for (i = 0; i < nthreads; i++) {
- CPU_ZERO(&cpuset);
- CPU_SET(cpu->map[i % cpu->nr], &cpuset);
+ for (i = 0; i < params.nthreads; i++) {
+ pthread_attr_t thread_attr;
- if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset))
+ pthread_attr_init(&thread_attr);
+ CPU_ZERO_S(size, cpuset);
+ CPU_SET_S(perf_cpu_map__cpu(cpu, i % perf_cpu_map__nr(cpu)).cpu, size, cpuset);
+
+ if (pthread_attr_setaffinity_np(&thread_attr, size, cpuset)) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_attr_setaffinity_np");
+ }
- if (pthread_create(&w[i], &thread_attr, workerfn, NULL))
+ if (pthread_create(&w[i], &thread_attr, workerfn, NULL)) {
+ CPU_FREE(cpuset);
err(EXIT_FAILURE, "pthread_create");
+ }
+ pthread_attr_destroy(&thread_attr);
}
+ CPU_FREE(cpuset);
}
static void toggle_done(int sig __maybe_unused,
@@ -123,7 +145,6 @@ int bench_futex_wake(int argc, const char **argv)
int ret = 0;
unsigned int i, j;
struct sigaction act;
- pthread_attr_t thread_attr;
struct perf_cpu_map *cpu;
argc = parse_options(argc, argv, options, bench_futex_wake_usage, 0);
@@ -132,7 +153,7 @@ int bench_futex_wake(int argc, const char **argv)
exit(EXIT_FAILURE);
}
- cpu = perf_cpu_map__new(NULL);
+ cpu = perf_cpu_map__new_online_cpus();
if (!cpu)
err(EXIT_FAILURE, "calloc");
@@ -141,59 +162,66 @@ int bench_futex_wake(int argc, const char **argv)
act.sa_sigaction = toggle_done;
sigaction(SIGINT, &act, NULL);
- if (!nthreads)
- nthreads = cpu->nr;
+ if (params.mlockall) {
+ if (mlockall(MCL_CURRENT | MCL_FUTURE))
+ err(EXIT_FAILURE, "mlockall");
+ }
+
+ if (!params.nthreads)
+ params.nthreads = perf_cpu_map__nr(cpu);
- worker = calloc(nthreads, sizeof(*worker));
+ worker = calloc(params.nthreads, sizeof(*worker));
if (!worker)
err(EXIT_FAILURE, "calloc");
- if (!fshared)
+ if (!params.fshared)
futex_flag = FUTEX_PRIVATE_FLAG;
printf("Run summary [PID %d]: blocking on %d threads (at [%s] futex %p), "
"waking up %d at a time.\n\n",
- getpid(), nthreads, fshared ? "shared":"private", &futex1, nwakes);
+ getpid(), params.nthreads, params.fshared ? "shared":"private",
+ &futex1, params.nwakes);
init_stats(&wakeup_stats);
init_stats(&waketime_stats);
- pthread_attr_init(&thread_attr);
- pthread_mutex_init(&thread_lock, NULL);
- pthread_cond_init(&thread_parent, NULL);
- pthread_cond_init(&thread_worker, NULL);
+ mutex_init(&thread_lock);
+ cond_init(&thread_parent);
+ cond_init(&thread_worker);
for (j = 0; j < bench_repeat && !done; j++) {
unsigned int nwoken = 0;
struct timeval start, end, runtime;
/* create, launch & block all threads */
- block_threads(worker, thread_attr, cpu);
+ block_threads(worker, cpu);
/* make sure all threads are already blocked */
- pthread_mutex_lock(&thread_lock);
+ mutex_lock(&thread_lock);
while (threads_starting)
- pthread_cond_wait(&thread_parent, &thread_lock);
- pthread_cond_broadcast(&thread_worker);
- pthread_mutex_unlock(&thread_lock);
+ cond_wait(&thread_parent, &thread_lock);
+ cond_broadcast(&thread_worker);
+ mutex_unlock(&thread_lock);
usleep(100000);
/* Ok, all threads are patiently blocked, start waking folks up */
gettimeofday(&start, NULL);
- while (nwoken != nthreads)
- nwoken += futex_wake(&futex1, nwakes, futex_flag);
+ while (nwoken != params.nthreads)
+ nwoken += futex_wake(&futex1,
+ params.nwakes, futex_flag);
gettimeofday(&end, NULL);
timersub(&end, &start, &runtime);
update_stats(&wakeup_stats, nwoken);
update_stats(&waketime_stats, runtime.tv_usec);
- if (!silent) {
+ if (!params.silent) {
printf("[Run %d]: Wokeup %d of %d threads in %.4f ms\n",
- j + 1, nwoken, nthreads, runtime.tv_usec / (double)USEC_PER_MSEC);
+ j + 1, nwoken, params.nthreads,
+ runtime.tv_usec / (double)USEC_PER_MSEC);
}
- for (i = 0; i < nthreads; i++) {
+ for (i = 0; i < params.nthreads; i++) {
ret = pthread_join(worker[i], NULL);
if (ret)
err(EXIT_FAILURE, "pthread_join");
@@ -202,13 +230,13 @@ int bench_futex_wake(int argc, const char **argv)
}
/* cleanup & report results */
- pthread_cond_destroy(&thread_parent);
- pthread_cond_destroy(&thread_worker);
- pthread_mutex_destroy(&thread_lock);
- pthread_attr_destroy(&thread_attr);
+ cond_destroy(&thread_parent);
+ cond_destroy(&thread_worker);
+ mutex_destroy(&thread_lock);
print_summary();
free(worker);
+ perf_cpu_map__put(cpu);
return ret;
}
diff --git a/tools/perf/bench/futex.c b/tools/perf/bench/futex.c
new file mode 100644
index 000000000000..4c4fee107e59
--- /dev/null
+++ b/tools/perf/bench/futex.c
@@ -0,0 +1,74 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <err.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/prctl.h>
+
+#include "futex.h"
+
+#ifndef PR_FUTEX_HASH
+#define PR_FUTEX_HASH 78
+# define PR_FUTEX_HASH_SET_SLOTS 1
+# define FH_FLAG_IMMUTABLE (1ULL << 0)
+# define PR_FUTEX_HASH_GET_SLOTS 2
+# define PR_FUTEX_HASH_GET_IMMUTABLE 3
+#endif // PR_FUTEX_HASH
+
+void futex_set_nbuckets_param(struct bench_futex_parameters *params)
+{
+ unsigned long flags;
+ int ret;
+
+ if (params->nbuckets < 0)
+ return;
+
+ flags = params->buckets_immutable ? FH_FLAG_IMMUTABLE : 0;
+ ret = prctl(PR_FUTEX_HASH, PR_FUTEX_HASH_SET_SLOTS, params->nbuckets, flags);
+ if (ret) {
+ printf("Requesting %d hash buckets failed: %d/%m\n",
+ params->nbuckets, ret);
+ err(EXIT_FAILURE, "prctl(PR_FUTEX_HASH)");
+ }
+}
+
+void futex_print_nbuckets(struct bench_futex_parameters *params)
+{
+ char *futex_hash_mode;
+ int ret;
+
+ ret = prctl(PR_FUTEX_HASH, PR_FUTEX_HASH_GET_SLOTS);
+ if (params->nbuckets >= 0) {
+ if (ret != params->nbuckets) {
+ if (ret < 0) {
+ printf("Can't query number of buckets: %m\n");
+ err(EXIT_FAILURE, "prctl(PR_FUTEX_HASH)");
+ }
+ printf("Requested number of hash buckets does not currently used.\n");
+ printf("Requested: %d in usage: %d\n", params->nbuckets, ret);
+ err(EXIT_FAILURE, "prctl(PR_FUTEX_HASH)");
+ }
+ if (params->nbuckets == 0) {
+ ret = asprintf(&futex_hash_mode, "Futex hashing: global hash");
+ } else {
+ ret = prctl(PR_FUTEX_HASH, PR_FUTEX_HASH_GET_IMMUTABLE);
+ if (ret < 0) {
+ printf("Can't check if the hash is immutable: %m\n");
+ err(EXIT_FAILURE, "prctl(PR_FUTEX_HASH)");
+ }
+ ret = asprintf(&futex_hash_mode, "Futex hashing: %d hash buckets %s",
+ params->nbuckets,
+ ret == 1 ? "(immutable)" : "");
+ }
+ } else {
+ if (ret <= 0) {
+ ret = asprintf(&futex_hash_mode, "Futex hashing: global hash");
+ } else {
+ ret = asprintf(&futex_hash_mode, "Futex hashing: auto resized to %d buckets",
+ ret);
+ }
+ }
+ if (ret < 0)
+ err(EXIT_FAILURE, "ENOMEM, futex_hash_mode");
+ printf("%s\n", futex_hash_mode);
+ free(futex_hash_mode);
+}
diff --git a/tools/perf/bench/futex.h b/tools/perf/bench/futex.h
index 31b53cc7d5bc..9c9a73f9d865 100644
--- a/tools/perf/bench/futex.h
+++ b/tools/perf/bench/futex.h
@@ -13,28 +13,53 @@
#include <sys/types.h>
#include <linux/futex.h>
+struct bench_futex_parameters {
+ bool silent;
+ bool fshared;
+ bool mlockall;
+ bool multi; /* lock-pi */
+ bool pi; /* requeue-pi */
+ bool broadcast; /* requeue */
+ unsigned int runtime; /* seconds*/
+ unsigned int nthreads;
+ unsigned int nfutexes;
+ unsigned int nwakes;
+ unsigned int nrequeue;
+ int nbuckets;
+ bool buckets_immutable;
+};
+
/**
- * futex() - SYS_futex syscall wrapper
+ * futex_syscall() - SYS_futex syscall wrapper
* @uaddr: address of first futex
* @op: futex op code
* @val: typically expected value of uaddr, but varies by op
* @timeout: typically an absolute struct timespec (except where noted
* otherwise). Overloaded by some ops
- * @uaddr2: address of second futex for some ops\
+ * @uaddr2: address of second futex for some ops
* @val3: varies by op
* @opflags: flags to be bitwise OR'd with op, such as FUTEX_PRIVATE_FLAG
*
- * futex() is used by all the following futex op wrappers. It can also be
+ * futex_syscall() is used by all the following futex op wrappers. It can also be
* used for misuse and abuse testing. Generally, the specific op wrappers
- * should be used instead. It is a macro instead of an static inline function as
- * some of the types over overloaded (timeout is used for nr_requeue for
- * example).
+ * should be used instead.
*
* These argument descriptions are the defaults for all
* like-named arguments in the following wrappers except where noted below.
*/
-#define futex(uaddr, op, val, timeout, uaddr2, val3, opflags) \
- syscall(SYS_futex, uaddr, op | opflags, val, timeout, uaddr2, val3)
+static inline int
+futex_syscall(volatile u_int32_t *uaddr, int op, u_int32_t val, struct timespec *timeout,
+ volatile u_int32_t *uaddr2, int val3, int opflags)
+{
+ return syscall(SYS_futex, uaddr, op | opflags, val, timeout, uaddr2, val3);
+}
+
+static inline int
+futex_syscall_nr_requeue(volatile u_int32_t *uaddr, int op, u_int32_t val, int nr_requeue,
+ volatile u_int32_t *uaddr2, int val3, int opflags)
+{
+ return syscall(SYS_futex, uaddr, op | opflags, val, nr_requeue, uaddr2, val3);
+}
/**
* futex_wait() - block on uaddr with optional timeout
@@ -43,7 +68,7 @@
static inline int
futex_wait(u_int32_t *uaddr, u_int32_t val, struct timespec *timeout, int opflags)
{
- return futex(uaddr, FUTEX_WAIT, val, timeout, NULL, 0, opflags);
+ return futex_syscall(uaddr, FUTEX_WAIT, val, timeout, NULL, 0, opflags);
}
/**
@@ -53,7 +78,7 @@ futex_wait(u_int32_t *uaddr, u_int32_t val, struct timespec *timeout, int opflag
static inline int
futex_wake(u_int32_t *uaddr, int nr_wake, int opflags)
{
- return futex(uaddr, FUTEX_WAKE, nr_wake, NULL, NULL, 0, opflags);
+ return futex_syscall(uaddr, FUTEX_WAKE, nr_wake, NULL, NULL, 0, opflags);
}
/**
@@ -62,7 +87,7 @@ futex_wake(u_int32_t *uaddr, int nr_wake, int opflags)
static inline int
futex_lock_pi(u_int32_t *uaddr, struct timespec *timeout, int opflags)
{
- return futex(uaddr, FUTEX_LOCK_PI, 0, timeout, NULL, 0, opflags);
+ return futex_syscall(uaddr, FUTEX_LOCK_PI, 0, timeout, NULL, 0, opflags);
}
/**
@@ -71,19 +96,56 @@ futex_lock_pi(u_int32_t *uaddr, struct timespec *timeout, int opflags)
static inline int
futex_unlock_pi(u_int32_t *uaddr, int opflags)
{
- return futex(uaddr, FUTEX_UNLOCK_PI, 0, NULL, NULL, 0, opflags);
+ return futex_syscall(uaddr, FUTEX_UNLOCK_PI, 0, NULL, NULL, 0, opflags);
}
/**
* futex_cmp_requeue() - requeue tasks from uaddr to uaddr2
* @nr_wake: wake up to this many tasks
-* @nr_requeue: requeue up to this many tasks
+* @nr_requeue: requeue up to this many tasks
*/
static inline int
futex_cmp_requeue(u_int32_t *uaddr, u_int32_t val, u_int32_t *uaddr2, int nr_wake,
int nr_requeue, int opflags)
{
- return futex(uaddr, FUTEX_CMP_REQUEUE, nr_wake, nr_requeue, uaddr2,
- val, opflags);
+ return futex_syscall_nr_requeue(uaddr, FUTEX_CMP_REQUEUE, nr_wake, nr_requeue, uaddr2,
+ val, opflags);
+}
+
+/**
+ * futex_wait_requeue_pi() - block on uaddr and prepare to requeue to uaddr2
+ * @uaddr: non-PI futex source
+ * @uaddr2: PI futex target
+ *
+ * This is the first half of the requeue_pi mechanism. It shall always be
+ * paired with futex_cmp_requeue_pi().
+ */
+static inline int
+futex_wait_requeue_pi(u_int32_t *uaddr, u_int32_t val, u_int32_t *uaddr2,
+ struct timespec *timeout, int opflags)
+{
+ return futex_syscall(uaddr, FUTEX_WAIT_REQUEUE_PI, val, timeout, uaddr2, 0,
+ opflags);
+}
+
+/**
+ * futex_cmp_requeue_pi() - requeue tasks from uaddr to uaddr2
+ * @uaddr: non-PI futex source
+ * @uaddr2: PI futex target
+ * @nr_requeue: requeue up to this many tasks
+ *
+ * This is the second half of the requeue_pi mechanism. It shall always be
+ * paired with futex_wait_requeue_pi(). The first waker is always awoken.
+ */
+static inline int
+futex_cmp_requeue_pi(u_int32_t *uaddr, u_int32_t val, u_int32_t *uaddr2,
+ int nr_requeue, int opflags)
+{
+ return futex_syscall_nr_requeue(uaddr, FUTEX_CMP_REQUEUE_PI, 1, nr_requeue, uaddr2,
+ val, opflags);
}
+
+void futex_set_nbuckets_param(struct bench_futex_parameters *params);
+void futex_print_nbuckets(struct bench_futex_parameters *params);
+
#endif /* _FUTEX_H */
diff --git a/tools/perf/bench/inject-buildid.c b/tools/perf/bench/inject-buildid.c
new file mode 100644
index 000000000000..f55c07e4be94
--- /dev/null
+++ b/tools/perf/bench/inject-buildid.c
@@ -0,0 +1,486 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <stdlib.h>
+#include <stddef.h>
+#include <ftw.h>
+#include <fcntl.h>
+#include <errno.h>
+#include <unistd.h>
+#include <pthread.h>
+#include <sys/mman.h>
+#include <sys/wait.h>
+#include <linux/kernel.h>
+#include <linux/time64.h>
+#include <linux/list.h>
+#include <linux/err.h>
+#include <linux/zalloc.h>
+#include <internal/lib.h>
+#include <subcmd/parse-options.h>
+
+#include "bench.h"
+#include "util/data.h"
+#include "util/stat.h"
+#include "util/debug.h"
+#include "util/symbol.h"
+#include "util/session.h"
+#include "util/build-id.h"
+#include "util/sample.h"
+#include "util/synthetic-events.h"
+
+#define MMAP_DEV_MAJOR 8
+#define DSO_MMAP_RATIO 4
+
+static unsigned int iterations = 100;
+static unsigned int nr_mmaps = 100;
+static unsigned int nr_samples = 100; /* samples per mmap */
+
+static u64 bench_sample_type;
+static u16 bench_id_hdr_size;
+
+struct bench_data {
+ int pid;
+ int input_pipe[2];
+ int output_pipe[2];
+ pthread_t th;
+};
+
+struct bench_dso {
+ struct list_head list;
+ char *name;
+ int ino;
+};
+
+static int nr_dsos;
+static struct bench_dso *dsos;
+
+extern int main(int argc, const char **argv);
+
+static const struct option options[] = {
+ OPT_UINTEGER('i', "iterations", &iterations,
+ "Number of iterations used to compute average (default: 100)"),
+ OPT_UINTEGER('m', "nr-mmaps", &nr_mmaps,
+ "Number of mmap events for each iteration (default: 100)"),
+ OPT_UINTEGER('n', "nr-samples", &nr_samples,
+ "Number of sample events per mmap event (default: 100)"),
+ OPT_INCR('v', "verbose", &verbose,
+ "be more verbose (show iteration count, DSO name, etc)"),
+ OPT_END()
+};
+
+static const char *const bench_usage[] = {
+ "perf bench internals inject-build-id <options>",
+ NULL
+};
+
+/*
+ * Helper for collect_dso that adds the given file as a dso to dso_list
+ * if it contains a build-id. Stops after collecting 4 times more than
+ * we need (for MMAP2 events).
+ */
+static int add_dso(const char *fpath, const struct stat *sb __maybe_unused,
+ int typeflag, struct FTW *ftwbuf __maybe_unused)
+{
+ struct bench_dso *dso = &dsos[nr_dsos];
+ struct build_id bid;
+
+ if (typeflag == FTW_D || typeflag == FTW_SL)
+ return 0;
+
+ if (filename__read_build_id(fpath, &bid) < 0)
+ return 0;
+
+ dso->name = realpath(fpath, NULL);
+ if (dso->name == NULL)
+ return -1;
+
+ dso->ino = nr_dsos++;
+ pr_debug2(" Adding DSO: %s\n", fpath);
+
+ /* stop if we collected enough DSOs */
+ if ((unsigned int)nr_dsos == DSO_MMAP_RATIO * nr_mmaps)
+ return 1;
+
+ return 0;
+}
+
+static void collect_dso(void)
+{
+ dsos = calloc(nr_mmaps * DSO_MMAP_RATIO, sizeof(*dsos));
+ if (dsos == NULL) {
+ printf(" Memory allocation failed\n");
+ exit(1);
+ }
+
+ if (nftw("/usr/lib/", add_dso, 10, FTW_PHYS) < 0)
+ return;
+
+ pr_debug(" Collected %d DSOs\n", nr_dsos);
+}
+
+static void release_dso(void)
+{
+ int i;
+
+ for (i = 0; i < nr_dsos; i++) {
+ struct bench_dso *dso = &dsos[i];
+
+ zfree(&dso->name);
+ }
+ free(dsos);
+}
+
+/* Fake address used by mmap and sample events */
+static u64 dso_map_addr(struct bench_dso *dso)
+{
+ return 0x400000ULL + dso->ino * 8192ULL;
+}
+
+static ssize_t synthesize_attr(struct bench_data *data)
+{
+ union perf_event event;
+
+ memset(&event, 0, sizeof(event.attr) + sizeof(u64));
+
+ event.header.type = PERF_RECORD_HEADER_ATTR;
+ event.header.size = sizeof(event.attr) + sizeof(u64);
+
+ event.attr.attr.type = PERF_TYPE_SOFTWARE;
+ event.attr.attr.config = PERF_COUNT_SW_TASK_CLOCK;
+ event.attr.attr.exclude_kernel = 1;
+ event.attr.attr.sample_id_all = 1;
+ event.attr.attr.sample_type = bench_sample_type;
+
+ return writen(data->input_pipe[1], &event, event.header.size);
+}
+
+static ssize_t synthesize_fork(struct bench_data *data)
+{
+ union perf_event event;
+
+ memset(&event, 0, sizeof(event.fork) + bench_id_hdr_size);
+
+ event.header.type = PERF_RECORD_FORK;
+ event.header.misc = PERF_RECORD_MISC_FORK_EXEC;
+ event.header.size = sizeof(event.fork) + bench_id_hdr_size;
+
+ event.fork.ppid = 1;
+ event.fork.ptid = 1;
+ event.fork.pid = data->pid;
+ event.fork.tid = data->pid;
+
+ return writen(data->input_pipe[1], &event, event.header.size);
+}
+
+static ssize_t synthesize_mmap(struct bench_data *data, struct bench_dso *dso, u64 timestamp)
+{
+ union perf_event event;
+ size_t len = offsetof(struct perf_record_mmap2, filename);
+ u64 *id_hdr_ptr = (void *)&event;
+ int ts_idx;
+
+ len += roundup(strlen(dso->name) + 1, 8) + bench_id_hdr_size;
+
+ memset(&event, 0, min(len, sizeof(event.mmap2)));
+
+ event.header.type = PERF_RECORD_MMAP2;
+ event.header.misc = PERF_RECORD_MISC_USER;
+ event.header.size = len;
+
+ event.mmap2.pid = data->pid;
+ event.mmap2.tid = data->pid;
+ event.mmap2.maj = MMAP_DEV_MAJOR;
+ event.mmap2.ino = dso->ino;
+
+ strcpy(event.mmap2.filename, dso->name);
+
+ event.mmap2.start = dso_map_addr(dso);
+ event.mmap2.len = 4096;
+ event.mmap2.prot = PROT_EXEC;
+
+ if (len > sizeof(event.mmap2)) {
+ /* write mmap2 event first */
+ if (writen(data->input_pipe[1], &event, len - bench_id_hdr_size) < 0)
+ return -1;
+ /* zero-fill sample id header */
+ memset(id_hdr_ptr, 0, bench_id_hdr_size);
+ /* put timestamp in the right position */
+ ts_idx = (bench_id_hdr_size / sizeof(u64)) - 2;
+ id_hdr_ptr[ts_idx] = timestamp;
+ if (writen(data->input_pipe[1], id_hdr_ptr, bench_id_hdr_size) < 0)
+ return -1;
+
+ return len;
+ }
+
+ ts_idx = (len / sizeof(u64)) - 2;
+ id_hdr_ptr[ts_idx] = timestamp;
+ return writen(data->input_pipe[1], &event, len);
+}
+
+static ssize_t synthesize_sample(struct bench_data *data, struct bench_dso *dso, u64 timestamp)
+{
+ union perf_event event;
+ struct perf_sample sample = {
+ .tid = data->pid,
+ .pid = data->pid,
+ .ip = dso_map_addr(dso),
+ .time = timestamp,
+ };
+
+ event.header.type = PERF_RECORD_SAMPLE;
+ event.header.misc = PERF_RECORD_MISC_USER;
+ event.header.size = perf_event__sample_event_size(&sample, bench_sample_type, 0);
+
+ perf_event__synthesize_sample(&event, bench_sample_type, 0, &sample);
+
+ return writen(data->input_pipe[1], &event, event.header.size);
+}
+
+static ssize_t synthesize_flush(struct bench_data *data)
+{
+ struct perf_event_header header = {
+ .size = sizeof(header),
+ .type = PERF_RECORD_FINISHED_ROUND,
+ };
+
+ return writen(data->input_pipe[1], &header, header.size);
+}
+
+static void *data_reader(void *arg)
+{
+ struct bench_data *data = arg;
+ char buf[8192];
+ int flag;
+ int n;
+
+ flag = fcntl(data->output_pipe[0], F_GETFL);
+ fcntl(data->output_pipe[0], F_SETFL, flag | O_NONBLOCK);
+
+ /* read out data from child */
+ while (true) {
+ n = read(data->output_pipe[0], buf, sizeof(buf));
+ if (n > 0)
+ continue;
+ if (n == 0)
+ break;
+
+ if (errno != EINTR && errno != EAGAIN)
+ break;
+
+ usleep(100);
+ }
+
+ close(data->output_pipe[0]);
+ return NULL;
+}
+
+static int setup_injection(struct bench_data *data, bool build_id_all)
+{
+ int ready_pipe[2];
+ int dev_null_fd;
+ char buf;
+
+ if (pipe(ready_pipe) < 0)
+ return -1;
+
+ if (pipe(data->input_pipe) < 0)
+ return -1;
+
+ if (pipe(data->output_pipe) < 0)
+ return -1;
+
+ data->pid = fork();
+ if (data->pid < 0)
+ return -1;
+
+ if (data->pid == 0) {
+ const char **inject_argv;
+ int inject_argc = 3;
+
+ close(data->input_pipe[1]);
+ close(data->output_pipe[0]);
+ close(ready_pipe[0]);
+
+ dup2(data->input_pipe[0], STDIN_FILENO);
+ close(data->input_pipe[0]);
+ dup2(data->output_pipe[1], STDOUT_FILENO);
+ close(data->output_pipe[1]);
+
+ dev_null_fd = open("/dev/null", O_WRONLY);
+ if (dev_null_fd < 0)
+ exit(1);
+
+ dup2(dev_null_fd, STDERR_FILENO);
+
+ if (build_id_all)
+ inject_argc++;
+
+ inject_argv = calloc(inject_argc + 1, sizeof(*inject_argv));
+ if (inject_argv == NULL)
+ exit(1);
+
+ inject_argv[0] = strdup("perf");
+ inject_argv[1] = strdup("inject");
+ inject_argv[2] = strdup("-b");
+ if (build_id_all)
+ inject_argv[3] = strdup("--buildid-all");
+
+ /* signal that we're ready to go */
+ close(ready_pipe[1]);
+
+ main(inject_argc, inject_argv);
+
+ exit(0);
+ }
+
+ pthread_create(&data->th, NULL, data_reader, data);
+
+ close(ready_pipe[1]);
+ close(data->input_pipe[0]);
+ close(data->output_pipe[1]);
+
+ /* wait for child ready */
+ if (read(ready_pipe[0], &buf, 1) < 0)
+ return -1;
+ close(ready_pipe[0]);
+
+ return 0;
+}
+
+static int inject_build_id(struct bench_data *data, u64 *max_rss)
+{
+ int status;
+ unsigned int i, k;
+ struct rusage rusage;
+
+ /* this makes the child to run */
+ if (perf_header__write_pipe(data->input_pipe[1]) < 0)
+ return -1;
+
+ if (synthesize_attr(data) < 0)
+ return -1;
+
+ if (synthesize_fork(data) < 0)
+ return -1;
+
+ for (i = 0; i < nr_mmaps; i++) {
+ int idx = rand() % nr_dsos;
+ struct bench_dso *dso = &dsos[idx];
+ u64 timestamp = rand() % 1000000;
+
+ pr_debug2(" [%d] injecting: %s\n", i+1, dso->name);
+ if (synthesize_mmap(data, dso, timestamp) < 0)
+ return -1;
+
+ for (k = 0; k < nr_samples; k++) {
+ if (synthesize_sample(data, dso, timestamp + k * 1000) < 0)
+ return -1;
+ }
+
+ if ((i + 1) % 10 == 0) {
+ if (synthesize_flush(data) < 0)
+ return -1;
+ }
+ }
+
+ /* this makes the child to finish */
+ close(data->input_pipe[1]);
+
+ wait4(data->pid, &status, 0, &rusage);
+ *max_rss = rusage.ru_maxrss;
+
+ pr_debug(" Child %d exited with %d\n", data->pid, status);
+
+ return 0;
+}
+
+static void do_inject_loop(struct bench_data *data, bool build_id_all)
+{
+ unsigned int i;
+ struct stats time_stats, mem_stats;
+ double time_average, time_stddev;
+ double mem_average, mem_stddev;
+
+ init_stats(&time_stats);
+ init_stats(&mem_stats);
+
+ pr_debug(" Build-id%s injection benchmark\n", build_id_all ? "-all" : "");
+
+ for (i = 0; i < iterations; i++) {
+ struct timeval start, end, diff;
+ u64 runtime_us, max_rss;
+
+ pr_debug(" Iteration #%d\n", i+1);
+
+ if (setup_injection(data, build_id_all) < 0) {
+ printf(" Build-id injection setup failed\n");
+ break;
+ }
+
+ gettimeofday(&start, NULL);
+ if (inject_build_id(data, &max_rss) < 0) {
+ printf(" Build-id injection failed\n");
+ break;
+ }
+
+ gettimeofday(&end, NULL);
+ timersub(&end, &start, &diff);
+ runtime_us = diff.tv_sec * USEC_PER_SEC + diff.tv_usec;
+ update_stats(&time_stats, runtime_us);
+ update_stats(&mem_stats, max_rss);
+
+ pthread_join(data->th, NULL);
+ }
+
+ time_average = avg_stats(&time_stats) / USEC_PER_MSEC;
+ time_stddev = stddev_stats(&time_stats) / USEC_PER_MSEC;
+ printf(" Average build-id%s injection took: %.3f msec (+- %.3f msec)\n",
+ build_id_all ? "-all" : "", time_average, time_stddev);
+
+ /* each iteration, it processes MMAP2 + BUILD_ID + nr_samples * SAMPLE */
+ time_average = avg_stats(&time_stats) / (nr_mmaps * (nr_samples + 2));
+ time_stddev = stddev_stats(&time_stats) / (nr_mmaps * (nr_samples + 2));
+ printf(" Average time per event: %.3f usec (+- %.3f usec)\n",
+ time_average, time_stddev);
+
+ mem_average = avg_stats(&mem_stats);
+ mem_stddev = stddev_stats(&mem_stats);
+ printf(" Average memory usage: %.0f KB (+- %.0f KB)\n",
+ mem_average, mem_stddev);
+}
+
+static int do_inject_loops(struct bench_data *data)
+{
+
+ srand(time(NULL));
+ symbol__init(NULL);
+
+ bench_sample_type = PERF_SAMPLE_IDENTIFIER | PERF_SAMPLE_IP;
+ bench_sample_type |= PERF_SAMPLE_TID | PERF_SAMPLE_TIME;
+ bench_id_hdr_size = 32;
+
+ collect_dso();
+ if (nr_dsos == 0) {
+ printf(" Cannot collect DSOs for injection\n");
+ return -1;
+ }
+
+ do_inject_loop(data, false);
+ do_inject_loop(data, true);
+
+ release_dso();
+ return 0;
+}
+
+int bench_inject_build_id(int argc, const char **argv)
+{
+ struct bench_data data;
+
+ argc = parse_options(argc, argv, options, bench_usage, 0);
+ if (argc) {
+ usage_with_options(bench_usage, options);
+ exit(EXIT_FAILURE);
+ }
+
+ return do_inject_loops(&data);
+}
+
diff --git a/tools/perf/bench/mem-functions.c b/tools/perf/bench/mem-functions.c
index 9235b76501be..19d45c377ac1 100644
--- a/tools/perf/bench/mem-functions.c
+++ b/tools/perf/bench/mem-functions.c
@@ -223,12 +223,8 @@ static int bench_mem_common(int argc, const char **argv, struct bench_mem_info *
return 0;
}
-static u64 do_memcpy_cycles(const struct function *r, size_t size, void *src, void *dst)
+static void memcpy_prefault(memcpy_t fn, size_t size, void *src, void *dst)
{
- u64 cycle_start = 0ULL, cycle_end = 0ULL;
- memcpy_t fn = r->fn.memcpy;
- int i;
-
/* Make sure to always prefault zero pages even if MMAP_THRESH is crossed: */
memset(src, 0, size);
@@ -237,6 +233,15 @@ static u64 do_memcpy_cycles(const struct function *r, size_t size, void *src, vo
* to not measure page fault overhead:
*/
fn(dst, src, size);
+}
+
+static u64 do_memcpy_cycles(const struct function *r, size_t size, void *src, void *dst)
+{
+ u64 cycle_start = 0ULL, cycle_end = 0ULL;
+ memcpy_t fn = r->fn.memcpy;
+ int i;
+
+ memcpy_prefault(fn, size, src, dst);
cycle_start = get_cycles();
for (i = 0; i < nr_loops; ++i)
@@ -252,11 +257,7 @@ static double do_memcpy_gettimeofday(const struct function *r, size_t size, void
memcpy_t fn = r->fn.memcpy;
int i;
- /*
- * We prefault the freshly allocated memory range here,
- * to not measure page fault overhead:
- */
- fn(dst, src, size);
+ memcpy_prefault(fn, size, src, dst);
BUG_ON(gettimeofday(&tv_start, NULL));
for (i = 0; i < nr_loops; ++i)
diff --git a/tools/perf/bench/mem-memcpy-x86-64-asm-def.h b/tools/perf/bench/mem-memcpy-x86-64-asm-def.h
index 50ae8bd58296..6188e19d3129 100644
--- a/tools/perf/bench/mem-memcpy-x86-64-asm-def.h
+++ b/tools/perf/bench/mem-memcpy-x86-64-asm-def.h
@@ -7,7 +7,3 @@ MEMCPY_FN(memcpy_orig,
MEMCPY_FN(__memcpy,
"x86-64-movsq",
"movsq-based memcpy() in arch/x86/lib/memcpy_64.S")
-
-MEMCPY_FN(memcpy_erms,
- "x86-64-movsb",
- "movsb-based memcpy() in arch/x86/lib/memcpy_64.S")
diff --git a/tools/perf/bench/mem-memcpy-x86-64-asm.S b/tools/perf/bench/mem-memcpy-x86-64-asm.S
index 9ad015a1e202..1b9fef7efcdc 100644
--- a/tools/perf/bench/mem-memcpy-x86-64-asm.S
+++ b/tools/perf/bench/mem-memcpy-x86-64-asm.S
@@ -2,6 +2,9 @@
/* Various wrappers to make the kernel .S file build in user-space: */
+// memcpy_orig is being defined as SYM_L_LOCAL but we need it
+#define SYM_FUNC_START_LOCAL(name) \
+ SYM_START(name, SYM_L_GLOBAL, SYM_A_ALIGN)
#define memcpy MEMCPY /* don't hide glibc's memcpy() */
#define altinstr_replacement text
#define globl p2align 4; .globl
diff --git a/tools/perf/bench/mem-memcpy-x86-64-lib.c b/tools/perf/bench/mem-memcpy-x86-64-lib.c
deleted file mode 100644
index 4130734dde84..000000000000
--- a/tools/perf/bench/mem-memcpy-x86-64-lib.c
+++ /dev/null
@@ -1,24 +0,0 @@
-/*
- * From code in arch/x86/lib/usercopy_64.c, copied to keep tools/ copy
- * of the kernel's arch/x86/lib/memcpy_64.s used in 'perf bench mem memcpy'
- * happy.
- */
-#include <linux/types.h>
-
-unsigned long __memcpy_mcsafe(void *dst, const void *src, size_t cnt);
-unsigned long mcsafe_handle_tail(char *to, char *from, unsigned len);
-
-unsigned long mcsafe_handle_tail(char *to, char *from, unsigned len)
-{
- for (; len; --len, to++, from++) {
- /*
- * Call the assembly routine back directly since
- * memcpy_mcsafe() may silently fallback to memcpy.
- */
- unsigned long rem = __memcpy_mcsafe(to, from, 1);
-
- if (rem)
- break;
- }
- return len;
-}
diff --git a/tools/perf/bench/mem-memset-x86-64-asm-def.h b/tools/perf/bench/mem-memset-x86-64-asm-def.h
index dac6d2b7c39b..247c72fdfb9d 100644
--- a/tools/perf/bench/mem-memset-x86-64-asm-def.h
+++ b/tools/perf/bench/mem-memset-x86-64-asm-def.h
@@ -7,7 +7,3 @@ MEMSET_FN(memset_orig,
MEMSET_FN(__memset,
"x86-64-stosq",
"movsq-based memset() in arch/x86/lib/memset_64.S")
-
-MEMSET_FN(memset_erms,
- "x86-64-stosb",
- "movsb-based memset() in arch/x86/lib/memset_64.S")
diff --git a/tools/perf/bench/mem-memset-x86-64-asm.S b/tools/perf/bench/mem-memset-x86-64-asm.S
index d550bd526162..abd26c95f1aa 100644
--- a/tools/perf/bench/mem-memset-x86-64-asm.S
+++ b/tools/perf/bench/mem-memset-x86-64-asm.S
@@ -1,4 +1,7 @@
/* SPDX-License-Identifier: GPL-2.0 */
+// memset_orig is being defined as SYM_L_LOCAL but we need it
+#define SYM_FUNC_START_LOCAL(name) \
+ SYM_START(name, SYM_L_GLOBAL, SYM_A_ALIGN)
#define memset MEMSET /* don't hide glibc's memset() */
#define altinstr_replacement text
#define globl p2align 4; .globl
diff --git a/tools/perf/bench/numa.c b/tools/perf/bench/numa.c
index 5797253b9700..19be2aaf4dc0 100644
--- a/tools/perf/bench/numa.c
+++ b/tools/perf/bench/numa.c
@@ -6,8 +6,6 @@
*/
#include <inttypes.h>
-/* For the CLR_() macros */
-#include <pthread.h>
#include <subcmd/parse-options.h>
#include "../util/cloexec.h"
@@ -18,6 +16,7 @@
#include <sched.h>
#include <stdio.h>
#include <assert.h>
+#include <debug.h>
#include <malloc.h>
#include <signal.h>
#include <stdlib.h>
@@ -28,12 +27,16 @@
#include <sys/resource.h>
#include <sys/wait.h>
#include <sys/prctl.h>
+#include <sys/stat.h>
#include <sys/types.h>
#include <linux/kernel.h>
#include <linux/time64.h>
#include <linux/numa.h>
#include <linux/zalloc.h>
+#include "../util/header.h"
+#include "../util/mutex.h"
+#include <api/fs/fs.h>
#include <numa.h>
#include <numaif.h>
@@ -42,7 +45,7 @@
#endif
/*
- * Regular printout to the terminal, supressed if -q is specified:
+ * Regular printout to the terminal, suppressed if -q is specified:
*/
#define tprintf(x...) do { if (g && g->p.show_details >= 0) printf(x); } while (0)
@@ -54,7 +57,7 @@
struct thread_data {
int curr_cpu;
- cpu_set_t bind_cpumask;
+ cpu_set_t *bind_cpumask;
int bind_node;
u8 *process_data;
int process_nr;
@@ -66,7 +69,7 @@ struct thread_data {
u64 system_time_ns;
u64 user_time_ns;
double speed_gbs;
- pthread_mutex_t *process_lock;
+ struct mutex *process_lock;
};
/* Parameters set by options: */
@@ -116,7 +119,6 @@ struct params {
long bytes_thread;
int nr_tasks;
- bool show_quiet;
bool show_convergence;
bool measure_convergence;
@@ -136,15 +138,16 @@ struct params {
struct global_info {
u8 *data;
- pthread_mutex_t startup_mutex;
+ struct mutex startup_mutex;
+ struct cond startup_cond;
int nr_tasks_started;
- pthread_mutex_t startup_done_mutex;
-
- pthread_mutex_t start_work_mutex;
+ struct mutex start_work_mutex;
+ struct cond start_work_cond;
int nr_tasks_working;
+ bool start_work;
- pthread_mutex_t stop_work_mutex;
+ struct mutex stop_work_mutex;
u64 bytes_done;
struct thread_data *threads;
@@ -196,7 +199,8 @@ static const struct option options[] = {
OPT_BOOLEAN('c', "show_convergence", &p0.show_convergence, "show convergence details, "
"convergence is reached when each process (all its threads) is running on a single NUMA node."),
OPT_BOOLEAN('m', "measure_convergence", &p0.measure_convergence, "measure convergence latency"),
- OPT_BOOLEAN('q', "quiet" , &p0.show_quiet, "quiet mode"),
+ OPT_BOOLEAN('q', "quiet" , &quiet,
+ "quiet mode (do not show any warnings or messages)"),
OPT_BOOLEAN('S', "serialize-startup", &p0.serialize_startup,"serialize thread startup"),
/* Special option string parsing callbacks: */
@@ -247,85 +251,135 @@ static int is_node_present(int node)
*/
static bool node_has_cpus(int node)
{
- struct bitmask *cpu = numa_allocate_cpumask();
- unsigned int i;
+ struct bitmask *cpumask = numa_allocate_cpumask();
+ bool ret = false; /* fall back to nocpus */
+ int cpu;
- if (cpu && !numa_node_to_cpus(node, cpu)) {
- for (i = 0; i < cpu->size; i++) {
- if (numa_bitmask_isbitset(cpu, i))
- return true;
+ BUG_ON(!cpumask);
+ if (!numa_node_to_cpus(node, cpumask)) {
+ for (cpu = 0; cpu < (int)cpumask->size; cpu++) {
+ if (numa_bitmask_isbitset(cpumask, cpu)) {
+ ret = true;
+ break;
+ }
}
}
+ numa_free_cpumask(cpumask);
- return false; /* lets fall back to nocpus safely */
+ return ret;
}
-static cpu_set_t bind_to_cpu(int target_cpu)
+static cpu_set_t *bind_to_cpu(int target_cpu)
{
- cpu_set_t orig_mask, mask;
- int ret;
+ int nrcpus = numa_num_possible_cpus();
+ cpu_set_t *orig_mask, *mask;
+ size_t size;
- ret = sched_getaffinity(0, sizeof(orig_mask), &orig_mask);
- BUG_ON(ret);
+ orig_mask = CPU_ALLOC(nrcpus);
+ BUG_ON(!orig_mask);
+ size = CPU_ALLOC_SIZE(nrcpus);
+ CPU_ZERO_S(size, orig_mask);
+
+ if (sched_getaffinity(0, size, orig_mask))
+ goto err_out;
+
+ mask = CPU_ALLOC(nrcpus);
+ if (!mask)
+ goto err_out;
- CPU_ZERO(&mask);
+ CPU_ZERO_S(size, mask);
if (target_cpu == -1) {
int cpu;
for (cpu = 0; cpu < g->p.nr_cpus; cpu++)
- CPU_SET(cpu, &mask);
+ CPU_SET_S(cpu, size, mask);
} else {
- BUG_ON(target_cpu < 0 || target_cpu >= g->p.nr_cpus);
- CPU_SET(target_cpu, &mask);
+ if (target_cpu < 0 || target_cpu >= g->p.nr_cpus)
+ goto err;
+
+ CPU_SET_S(target_cpu, size, mask);
}
- ret = sched_setaffinity(0, sizeof(mask), &mask);
- BUG_ON(ret);
+ if (sched_setaffinity(0, size, mask))
+ goto err;
return orig_mask;
+
+err:
+ CPU_FREE(mask);
+err_out:
+ CPU_FREE(orig_mask);
+
+ /* BUG_ON due to failure in allocation of orig_mask/mask */
+ BUG_ON(-1);
+ return NULL;
}
-static cpu_set_t bind_to_node(int target_node)
+static cpu_set_t *bind_to_node(int target_node)
{
- int cpus_per_node = g->p.nr_cpus / nr_numa_nodes();
- cpu_set_t orig_mask, mask;
+ int nrcpus = numa_num_possible_cpus();
+ size_t size;
+ cpu_set_t *orig_mask, *mask;
int cpu;
- int ret;
- BUG_ON(cpus_per_node * nr_numa_nodes() != g->p.nr_cpus);
- BUG_ON(!cpus_per_node);
+ orig_mask = CPU_ALLOC(nrcpus);
+ BUG_ON(!orig_mask);
+ size = CPU_ALLOC_SIZE(nrcpus);
+ CPU_ZERO_S(size, orig_mask);
- ret = sched_getaffinity(0, sizeof(orig_mask), &orig_mask);
- BUG_ON(ret);
+ if (sched_getaffinity(0, size, orig_mask))
+ goto err_out;
+
+ mask = CPU_ALLOC(nrcpus);
+ if (!mask)
+ goto err_out;
- CPU_ZERO(&mask);
+ CPU_ZERO_S(size, mask);
if (target_node == NUMA_NO_NODE) {
for (cpu = 0; cpu < g->p.nr_cpus; cpu++)
- CPU_SET(cpu, &mask);
+ CPU_SET_S(cpu, size, mask);
} else {
- int cpu_start = (target_node + 0) * cpus_per_node;
- int cpu_stop = (target_node + 1) * cpus_per_node;
+ struct bitmask *cpumask = numa_allocate_cpumask();
- BUG_ON(cpu_stop > g->p.nr_cpus);
+ if (!cpumask)
+ goto err;
- for (cpu = cpu_start; cpu < cpu_stop; cpu++)
- CPU_SET(cpu, &mask);
+ if (!numa_node_to_cpus(target_node, cpumask)) {
+ for (cpu = 0; cpu < (int)cpumask->size; cpu++) {
+ if (numa_bitmask_isbitset(cpumask, cpu))
+ CPU_SET_S(cpu, size, mask);
+ }
+ }
+ numa_free_cpumask(cpumask);
}
- ret = sched_setaffinity(0, sizeof(mask), &mask);
- BUG_ON(ret);
+ if (sched_setaffinity(0, size, mask))
+ goto err;
return orig_mask;
+
+err:
+ CPU_FREE(mask);
+err_out:
+ CPU_FREE(orig_mask);
+
+ /* BUG_ON due to failure in allocation of orig_mask/mask */
+ BUG_ON(-1);
+ return NULL;
}
-static void bind_to_cpumask(cpu_set_t mask)
+static void bind_to_cpumask(cpu_set_t *mask)
{
int ret;
+ size_t size = CPU_ALLOC_SIZE(numa_num_possible_cpus());
- ret = sched_setaffinity(0, sizeof(mask), &mask);
- BUG_ON(ret);
+ ret = sched_setaffinity(0, size, mask);
+ if (ret) {
+ CPU_FREE(mask);
+ BUG_ON(ret);
+ }
}
static void mempol_restore(void)
@@ -339,18 +393,22 @@ static void mempol_restore(void)
static void bind_to_memnode(int node)
{
- unsigned long nodemask;
+ struct bitmask *node_mask;
int ret;
if (node == NUMA_NO_NODE)
return;
- BUG_ON(g->p.nr_nodes > (int)sizeof(nodemask)*8);
- nodemask = 1L << node;
+ node_mask = numa_allocate_nodemask();
+ BUG_ON(!node_mask);
+
+ numa_bitmask_clearall(node_mask);
+ numa_bitmask_setbit(node_mask, node);
- ret = set_mempolicy(MPOL_BIND, &nodemask, sizeof(nodemask)*8);
- dprintf("binding to node %d, mask: %016lx => %d\n", node, nodemask, ret);
+ ret = set_mempolicy(MPOL_BIND, node_mask->maskp, node_mask->size + 1);
+ dprintf("binding to node %d, mask: %016lx => %d\n", node, *node_mask->maskp, ret);
+ numa_bitmask_free(node_mask);
BUG_ON(ret);
}
@@ -367,7 +425,7 @@ do { \
static u8 *alloc_data(ssize_t bytes0, int map_flags,
int init_zero, int init_cpu0, int thp, int init_random)
{
- cpu_set_t orig_mask;
+ cpu_set_t *orig_mask = NULL;
ssize_t bytes;
u8 *buf;
int ret;
@@ -425,6 +483,7 @@ static u8 *alloc_data(ssize_t bytes0, int map_flags,
/* Restore affinity: */
if (init_cpu0) {
bind_to_cpumask(orig_mask);
+ CPU_FREE(orig_mask);
mempol_restore();
}
@@ -467,18 +526,6 @@ static void * setup_private_data(ssize_t bytes)
return alloc_data(bytes, MAP_PRIVATE, 0, g->p.init_cpu0, g->p.thp, g->p.init_random);
}
-/*
- * Return a process-shared (global) mutex:
- */
-static void init_global_mutex(pthread_mutex_t *mutex)
-{
- pthread_mutexattr_t attr;
-
- pthread_mutexattr_init(&attr);
- pthread_mutexattr_setpshared(&attr, PTHREAD_PROCESS_SHARED);
- pthread_mutex_init(mutex, &attr);
-}
-
static int parse_cpu_list(const char *arg)
{
p0.cpu_list_str = strdup(arg);
@@ -488,6 +535,57 @@ static int parse_cpu_list(const char *arg)
return 0;
}
+/*
+ * Check whether a CPU is online
+ *
+ * Returns:
+ * 1 -> if CPU is online
+ * 0 -> if CPU is offline
+ * -1 -> error case
+ */
+static int is_cpu_online(unsigned int cpu)
+{
+ char *str;
+ size_t strlen;
+ char buf[256];
+ int status = -1;
+ struct stat statbuf;
+
+ snprintf(buf, sizeof(buf),
+ "/sys/devices/system/cpu/cpu%d", cpu);
+ if (stat(buf, &statbuf) != 0)
+ return 0;
+
+ /*
+ * Check if /sys/devices/system/cpu/cpux/online file
+ * exists. Some cases cpu0 won't have online file since
+ * it is not expected to be turned off generally.
+ * In kernels without CONFIG_HOTPLUG_CPU, this
+ * file won't exist
+ */
+ snprintf(buf, sizeof(buf),
+ "/sys/devices/system/cpu/cpu%d/online", cpu);
+ if (stat(buf, &statbuf) != 0)
+ return 1;
+
+ /*
+ * Read online file using sysfs__read_str.
+ * If read or open fails, return -1.
+ * If read succeeds, return value from file
+ * which gets stored in "str"
+ */
+ snprintf(buf, sizeof(buf),
+ "devices/system/cpu/cpu%d/online", cpu);
+
+ if (sysfs__read_str(buf, &str, &strlen) < 0)
+ return status;
+
+ status = atoi(str);
+
+ free(str);
+ return status;
+}
+
static int parse_setup_cpu_list(void)
{
struct thread_data *td;
@@ -564,10 +662,16 @@ static int parse_setup_cpu_list(void)
return -1;
}
+ if (is_cpu_online(bind_cpu_0) != 1 || is_cpu_online(bind_cpu_1) != 1) {
+ printf("\nTest not applicable, bind_cpu_0 or bind_cpu_1 is offline\n");
+ return -1;
+ }
+
BUG_ON(bind_cpu_0 < 0 || bind_cpu_1 < 0);
BUG_ON(bind_cpu_0 > bind_cpu_1);
for (bind_cpu = bind_cpu_0; bind_cpu <= bind_cpu_1; bind_cpu += step) {
+ size_t size = CPU_ALLOC_SIZE(g->p.nr_cpus);
int i;
for (i = 0; i < mul; i++) {
@@ -587,10 +691,15 @@ static int parse_setup_cpu_list(void)
tprintf("%2d", bind_cpu);
}
- CPU_ZERO(&td->bind_cpumask);
+ td->bind_cpumask = CPU_ALLOC(g->p.nr_cpus);
+ BUG_ON(!td->bind_cpumask);
+ CPU_ZERO_S(size, td->bind_cpumask);
for (cpu = bind_cpu; cpu < bind_cpu+bind_len; cpu++) {
- BUG_ON(cpu < 0 || cpu >= g->p.nr_cpus);
- CPU_SET(cpu, &td->bind_cpumask);
+ if (cpu < 0 || cpu >= g->p.nr_cpus) {
+ CPU_FREE(td->bind_cpumask);
+ BUG_ON(-1);
+ }
+ CPU_SET_S(cpu, size, td->bind_cpumask);
}
t++;
}
@@ -729,12 +838,8 @@ static int parse_nodes_opt(const struct option *opt __maybe_unused,
return -1;
return parse_node_list(arg);
-
- return 0;
}
-#define BIT(x) (1ul << x)
-
static inline uint32_t lfsr_32(uint32_t lfsr)
{
const uint32_t taps = BIT(1) | BIT(5) | BIT(6) | BIT(31);
@@ -795,7 +900,7 @@ static u64 do_work(u8 *__data, long bytes, int nr, int nr_max, int loop, u64 val
if (g->p.data_rand_walk) {
u32 lfsr = nr + loop + val;
- int j;
+ long j;
for (i = 0; i < words/1024; i++) {
long start, end;
@@ -813,12 +918,12 @@ static u64 do_work(u8 *__data, long bytes, int nr, int nr_max, int loop, u64 val
}
}
} else if (!g->p.data_backwards || (nr + loop) & 1) {
+ /* Process data forwards: */
d0 = data + off;
d = data + off + 1;
d1 = data + words;
- /* Process data forwards: */
for (;;) {
if (unlikely(d >= d1))
d = data;
@@ -836,7 +941,6 @@ static u64 do_work(u8 *__data, long bytes, int nr, int nr_max, int loop, u64 val
d = data + off - 1;
d1 = data + words;
- /* Process data forwards: */
for (;;) {
if (unlikely(d < data))
d = data + words-1;
@@ -862,8 +966,6 @@ static void update_curr_cpu(int task_nr, unsigned long bytes_worked)
prctl(0, bytes_worked);
}
-#define MAX_NR_NODES 64
-
/*
* Count the number of nodes a process's threads
* are spread out on.
@@ -874,10 +976,15 @@ static void update_curr_cpu(int task_nr, unsigned long bytes_worked)
*/
static int count_process_nodes(int process_nr)
{
- char node_present[MAX_NR_NODES] = { 0, };
+ char *node_present;
int nodes;
int n, t;
+ node_present = (char *)malloc(g->p.nr_nodes * sizeof(char));
+ BUG_ON(!node_present);
+ for (nodes = 0; nodes < g->p.nr_nodes; nodes++)
+ node_present[nodes] = 0;
+
for (t = 0; t < g->p.nr_threads; t++) {
struct thread_data *td;
int task_nr;
@@ -887,17 +994,20 @@ static int count_process_nodes(int process_nr)
td = g->threads + task_nr;
node = numa_node_of_cpu(td->curr_cpu);
- if (node < 0) /* curr_cpu was likely still -1 */
+ if (node < 0) /* curr_cpu was likely still -1 */ {
+ free(node_present);
return 0;
+ }
node_present[node] = 1;
}
nodes = 0;
- for (n = 0; n < MAX_NR_NODES; n++)
+ for (n = 0; n < g->p.nr_nodes; n++)
nodes += node_present[n];
+ free(node_present);
return nodes;
}
@@ -966,7 +1076,7 @@ static void calc_convergence(double runtime_ns_max, double *convergence)
{
unsigned int loops_done_min, loops_done_max;
int process_groups;
- int nodes[MAX_NR_NODES];
+ int *nodes;
int distance;
int nr_min;
int nr_max;
@@ -980,6 +1090,8 @@ static void calc_convergence(double runtime_ns_max, double *convergence)
if (!g->p.show_convergence && !g->p.measure_convergence)
return;
+ nodes = (int *)malloc(g->p.nr_nodes * sizeof(int));
+ BUG_ON(!nodes);
for (node = 0; node < g->p.nr_nodes; node++)
nodes[node] = 0;
@@ -1021,8 +1133,10 @@ static void calc_convergence(double runtime_ns_max, double *convergence)
BUG_ON(sum > g->p.nr_tasks);
- if (0 && (sum < g->p.nr_tasks))
+ if (0 && (sum < g->p.nr_tasks)) {
+ free(nodes);
return;
+ }
/*
* Count the number of distinct process groups present
@@ -1074,6 +1188,8 @@ static void calc_convergence(double runtime_ns_max, double *convergence)
}
tprintf("\n");
}
+
+ free(nodes);
}
static void show_summary(double runtime_ns_max, int l, double *convergence)
@@ -1133,19 +1249,22 @@ static void *worker_thread(void *__tdata)
}
if (g->p.serialize_startup) {
- pthread_mutex_lock(&g->startup_mutex);
+ mutex_lock(&g->startup_mutex);
g->nr_tasks_started++;
- pthread_mutex_unlock(&g->startup_mutex);
+ /* The last thread wakes the main process. */
+ if (g->nr_tasks_started == g->p.nr_tasks)
+ cond_signal(&g->startup_cond);
+
+ mutex_unlock(&g->startup_mutex);
/* Here we will wait for the main process to start us all at once: */
- pthread_mutex_lock(&g->start_work_mutex);
+ mutex_lock(&g->start_work_mutex);
+ g->start_work = false;
g->nr_tasks_working++;
+ while (!g->start_work)
+ cond_wait(&g->start_work_cond, &g->start_work_mutex);
- /* Last one wake the main process: */
- if (g->nr_tasks_working == g->p.nr_tasks)
- pthread_mutex_unlock(&g->startup_done_mutex);
-
- pthread_mutex_unlock(&g->start_work_mutex);
+ mutex_unlock(&g->start_work_mutex);
}
gettimeofday(&start0, NULL);
@@ -1164,17 +1283,17 @@ static void *worker_thread(void *__tdata)
val += do_work(thread_data, g->p.bytes_thread, 0, 1, l, val);
if (g->p.sleep_usecs) {
- pthread_mutex_lock(td->process_lock);
+ mutex_lock(td->process_lock);
usleep(g->p.sleep_usecs);
- pthread_mutex_unlock(td->process_lock);
+ mutex_unlock(td->process_lock);
}
/*
* Amount of work to be done under a process-global lock:
*/
if (g->p.bytes_process_locked) {
- pthread_mutex_lock(td->process_lock);
+ mutex_lock(td->process_lock);
val += do_work(process_data, g->p.bytes_process_locked, thread_nr, g->p.nr_threads, l, val);
- pthread_mutex_unlock(td->process_lock);
+ mutex_unlock(td->process_lock);
}
work_done = g->p.bytes_global + g->p.bytes_process +
@@ -1208,7 +1327,7 @@ static void *worker_thread(void *__tdata)
* by migrating to CPU#0:
*/
if (first_task && g->p.perturb_secs && (int)(stop.tv_sec - last_perturbance) >= g->p.perturb_secs) {
- cpu_set_t orig_mask;
+ cpu_set_t *orig_mask;
int target_cpu;
int this_cpu;
@@ -1232,6 +1351,7 @@ static void *worker_thread(void *__tdata)
printf(" (injecting perturbalance, moved to CPU#%d)\n", target_cpu);
bind_to_cpumask(orig_mask);
+ CPU_FREE(orig_mask);
}
if (details >= 3) {
@@ -1270,9 +1390,9 @@ static void *worker_thread(void *__tdata)
free_data(thread_data, g->p.bytes_thread);
- pthread_mutex_lock(&g->stop_work_mutex);
+ mutex_lock(&g->stop_work_mutex);
g->bytes_done += bytes_done;
- pthread_mutex_unlock(&g->stop_work_mutex);
+ mutex_unlock(&g->stop_work_mutex);
return NULL;
}
@@ -1282,7 +1402,7 @@ static void *worker_thread(void *__tdata)
*/
static void worker_process(int process_nr)
{
- pthread_mutex_t process_lock;
+ struct mutex process_lock;
struct thread_data *td;
pthread_t *pthreads;
u8 *process_data;
@@ -1290,7 +1410,7 @@ static void worker_process(int process_nr)
int ret;
int t;
- pthread_mutex_init(&process_lock, NULL);
+ mutex_init(&process_lock);
set_taskname("process %d", process_nr);
/*
@@ -1365,21 +1485,31 @@ static void init_thread_data(void)
for (t = 0; t < g->p.nr_tasks; t++) {
struct thread_data *td = g->threads + t;
+ size_t cpuset_size = CPU_ALLOC_SIZE(g->p.nr_cpus);
int cpu;
/* Allow all nodes by default: */
td->bind_node = NUMA_NO_NODE;
/* Allow all CPUs by default: */
- CPU_ZERO(&td->bind_cpumask);
+ td->bind_cpumask = CPU_ALLOC(g->p.nr_cpus);
+ BUG_ON(!td->bind_cpumask);
+ CPU_ZERO_S(cpuset_size, td->bind_cpumask);
for (cpu = 0; cpu < g->p.nr_cpus; cpu++)
- CPU_SET(cpu, &td->bind_cpumask);
+ CPU_SET_S(cpu, cpuset_size, td->bind_cpumask);
}
}
static void deinit_thread_data(void)
{
ssize_t size = sizeof(*g->threads)*g->p.nr_tasks;
+ int t;
+
+ /* Free the bind_cpumask allocated for thread_data */
+ for (t = 0; t < g->p.nr_tasks; t++) {
+ struct thread_data *td = g->threads + t;
+ CPU_FREE(td->bind_cpumask);
+ }
free_data(g->threads, size);
}
@@ -1396,9 +1526,9 @@ static int init(void)
g->p.nr_nodes = numa_max_node() + 1;
/* char array in count_process_nodes(): */
- BUG_ON(g->p.nr_nodes > MAX_NR_NODES || g->p.nr_nodes < 0);
+ BUG_ON(g->p.nr_nodes < 0);
- if (g->p.show_quiet && !g->p.show_details)
+ if (quiet && !g->p.show_details)
g->p.show_details = -1;
/* Some memory should be specified: */
@@ -1439,10 +1569,11 @@ static int init(void)
g->data = setup_shared_data(g->p.bytes_global);
/* Startup serialization: */
- init_global_mutex(&g->start_work_mutex);
- init_global_mutex(&g->startup_mutex);
- init_global_mutex(&g->startup_done_mutex);
- init_global_mutex(&g->stop_work_mutex);
+ mutex_init_pshared(&g->start_work_mutex);
+ cond_init_pshared(&g->start_work_cond);
+ mutex_init_pshared(&g->startup_mutex);
+ cond_init_pshared(&g->startup_cond);
+ mutex_init_pshared(&g->stop_work_mutex);
init_thread_data();
@@ -1476,7 +1607,7 @@ static void print_res(const char *name, double val,
if (!name)
name = "main,";
- if (!g->p.show_quiet)
+ if (!quiet)
printf(" %-30s %15.3f, %-15s %s\n", name, val, txt_unit, txt_short);
else
printf(" %14.3f %s\n", val, txt_long);
@@ -1501,9 +1632,6 @@ static int __bench_numa(const char *name)
pids = zalloc(g->p.nr_proc * sizeof(*pids));
pid = -1;
- /* All threads try to acquire it, this way we can wait for them to start up: */
- pthread_mutex_lock(&g->start_work_mutex);
-
if (g->p.serialize_startup) {
tprintf(" #\n");
tprintf(" # Startup synchronization: ..."); fflush(stdout);
@@ -1525,22 +1653,29 @@ static int __bench_numa(const char *name)
pids[i] = pid;
}
- /* Wait for all the threads to start up: */
- while (g->nr_tasks_started != g->p.nr_tasks)
- usleep(USEC_PER_MSEC);
-
- BUG_ON(g->nr_tasks_started != g->p.nr_tasks);
if (g->p.serialize_startup) {
+ bool threads_ready = false;
double startup_sec;
- pthread_mutex_lock(&g->startup_done_mutex);
-
- /* This will start all threads: */
- pthread_mutex_unlock(&g->start_work_mutex);
-
- /* This mutex is locked - the last started thread will wake us: */
- pthread_mutex_lock(&g->startup_done_mutex);
+ /*
+ * Wait for all the threads to start up. The last thread will
+ * signal this process.
+ */
+ mutex_lock(&g->startup_mutex);
+ while (g->nr_tasks_started != g->p.nr_tasks)
+ cond_wait(&g->startup_cond, &g->startup_mutex);
+
+ mutex_unlock(&g->startup_mutex);
+
+ /* Wait for all threads to be at the start_work_cond. */
+ while (!threads_ready) {
+ mutex_lock(&g->start_work_mutex);
+ threads_ready = (g->nr_tasks_working == g->p.nr_tasks);
+ mutex_unlock(&g->start_work_mutex);
+ if (!threads_ready)
+ usleep(1);
+ }
gettimeofday(&stop, NULL);
@@ -1554,7 +1689,11 @@ static int __bench_numa(const char *name)
tprintf(" #\n");
start = stop;
- pthread_mutex_unlock(&g->startup_done_mutex);
+ /* Start all threads running. */
+ mutex_lock(&g->start_work_mutex);
+ g->start_work = true;
+ mutex_unlock(&g->start_work_mutex);
+ cond_broadcast(&g->start_work_cond);
} else {
gettimeofday(&start, NULL);
}
@@ -1630,7 +1769,7 @@ static int __bench_numa(const char *name)
"GB/sec,", "total-speed", "GB/sec total speed");
if (g->p.show_details >= 2) {
- char tname[14 + 2 * 10 + 1];
+ char tname[14 + 2 * 11 + 1];
struct thread_data *td;
for (p = 0; p < g->p.nr_proc; p++) {
for (t = 0; t < g->p.nr_threads; t++) {
@@ -1733,12 +1872,12 @@ err:
*/
static const char *tests[][MAX_ARGS] = {
/* Basic single-stream NUMA bandwidth measurements: */
- { "RAM-bw-local,", "mem", "-p", "1", "-t", "1", "-P", "1024",
+ { "RAM-bw-local,", "mem", "-p", "1", "-t", "1", "-P", "1024",
"-C" , "0", "-M", "0", OPT_BW_RAM },
{ "RAM-bw-local-NOTHP,",
"mem", "-p", "1", "-t", "1", "-P", "1024",
"-C" , "0", "-M", "0", OPT_BW_RAM_NOTHP },
- { "RAM-bw-remote,", "mem", "-p", "1", "-t", "1", "-P", "1024",
+ { "RAM-bw-remote,", "mem", "-p", "1", "-t", "1", "-P", "1024",
"-C" , "0", "-M", "1", OPT_BW_RAM },
/* 2-stream NUMA bandwidth measurements: */
@@ -1755,7 +1894,7 @@ static const char *tests[][MAX_ARGS] = {
{ " 1x3-convergence,", "mem", "-p", "1", "-t", "3", "-P", "512", OPT_CONV },
{ " 1x4-convergence,", "mem", "-p", "1", "-t", "4", "-P", "512", OPT_CONV },
{ " 1x6-convergence,", "mem", "-p", "1", "-t", "6", "-P", "1020", OPT_CONV },
- { " 2x3-convergence,", "mem", "-p", "3", "-t", "3", "-P", "1020", OPT_CONV },
+ { " 2x3-convergence,", "mem", "-p", "2", "-t", "3", "-P", "1020", OPT_CONV },
{ " 3x3-convergence,", "mem", "-p", "3", "-t", "3", "-P", "1020", OPT_CONV },
{ " 4x4-convergence,", "mem", "-p", "4", "-t", "4", "-P", "512", OPT_CONV },
{ " 4x4-convergence-NOTHP,",
@@ -1780,24 +1919,24 @@ static const char *tests[][MAX_ARGS] = {
"mem", "-p", "8", "-t", "1", "-P", " 512", OPT_BW_NOTHP },
{ "16x1-bw-process,", "mem", "-p", "16", "-t", "1", "-P", "256", OPT_BW },
- { " 4x1-bw-thread,", "mem", "-p", "1", "-t", "4", "-T", "256", OPT_BW },
- { " 8x1-bw-thread,", "mem", "-p", "1", "-t", "8", "-T", "256", OPT_BW },
- { "16x1-bw-thread,", "mem", "-p", "1", "-t", "16", "-T", "128", OPT_BW },
- { "32x1-bw-thread,", "mem", "-p", "1", "-t", "32", "-T", "64", OPT_BW },
+ { " 1x4-bw-thread,", "mem", "-p", "1", "-t", "4", "-T", "256", OPT_BW },
+ { " 1x8-bw-thread,", "mem", "-p", "1", "-t", "8", "-T", "256", OPT_BW },
+ { "1x16-bw-thread,", "mem", "-p", "1", "-t", "16", "-T", "128", OPT_BW },
+ { "1x32-bw-thread,", "mem", "-p", "1", "-t", "32", "-T", "64", OPT_BW },
- { " 2x3-bw-thread,", "mem", "-p", "2", "-t", "3", "-P", "512", OPT_BW },
- { " 4x4-bw-thread,", "mem", "-p", "4", "-t", "4", "-P", "512", OPT_BW },
- { " 4x6-bw-thread,", "mem", "-p", "4", "-t", "6", "-P", "512", OPT_BW },
- { " 4x8-bw-thread,", "mem", "-p", "4", "-t", "8", "-P", "512", OPT_BW },
- { " 4x8-bw-thread-NOTHP,",
+ { " 2x3-bw-process,", "mem", "-p", "2", "-t", "3", "-P", "512", OPT_BW },
+ { " 4x4-bw-process,", "mem", "-p", "4", "-t", "4", "-P", "512", OPT_BW },
+ { " 4x6-bw-process,", "mem", "-p", "4", "-t", "6", "-P", "512", OPT_BW },
+ { " 4x8-bw-process,", "mem", "-p", "4", "-t", "8", "-P", "512", OPT_BW },
+ { " 4x8-bw-process-NOTHP,",
"mem", "-p", "4", "-t", "8", "-P", "512", OPT_BW_NOTHP },
- { " 3x3-bw-thread,", "mem", "-p", "3", "-t", "3", "-P", "512", OPT_BW },
- { " 5x5-bw-thread,", "mem", "-p", "5", "-t", "5", "-P", "512", OPT_BW },
+ { " 3x3-bw-process,", "mem", "-p", "3", "-t", "3", "-P", "512", OPT_BW },
+ { " 5x5-bw-process,", "mem", "-p", "5", "-t", "5", "-P", "512", OPT_BW },
- { "2x16-bw-thread,", "mem", "-p", "2", "-t", "16", "-P", "512", OPT_BW },
- { "1x32-bw-thread,", "mem", "-p", "1", "-t", "32", "-P", "2048", OPT_BW },
+ { "2x16-bw-process,", "mem", "-p", "2", "-t", "16", "-P", "512", OPT_BW },
+ { "1x32-bw-process,", "mem", "-p", "1", "-t", "32", "-P", "2048", OPT_BW },
- { "numa02-bw,", "mem", "-p", "1", "-t", "32", "-T", "32", OPT_BW },
+ { "numa02-bw,", "mem", "-p", "1", "-t", "32", "-T", "32", OPT_BW },
{ "numa02-bw-NOTHP,", "mem", "-p", "1", "-t", "32", "-T", "32", OPT_BW_NOTHP },
{ "numa01-bw-thread,", "mem", "-p", "2", "-t", "16", "-T", "192", OPT_BW },
{ "numa01-bw-thread-NOTHP,",
diff --git a/tools/perf/bench/pmu-scan.c b/tools/perf/bench/pmu-scan.c
new file mode 100644
index 000000000000..9e4d36486f62
--- /dev/null
+++ b/tools/perf/bench/pmu-scan.c
@@ -0,0 +1,186 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Benchmark scanning sysfs files for PMU information.
+ *
+ * Copyright 2023 Google LLC.
+ */
+#include <stdio.h>
+#include "bench.h"
+#include "util/debug.h"
+#include "util/pmu.h"
+#include "util/pmus.h"
+#include "util/stat.h"
+#include <linux/atomic.h>
+#include <linux/err.h>
+#include <linux/time64.h>
+#include <subcmd/parse-options.h>
+
+static unsigned int iterations = 100;
+
+struct pmu_scan_result {
+ char *name;
+ int nr_aliases;
+ int nr_formats;
+ int nr_caps;
+ bool is_core;
+};
+
+static const struct option options[] = {
+ OPT_UINTEGER('i', "iterations", &iterations,
+ "Number of iterations used to compute average"),
+ OPT_END()
+};
+
+static const char *const bench_usage[] = {
+ "perf bench internals pmu-scan <options>",
+ NULL
+};
+
+static int nr_pmus;
+static struct pmu_scan_result *results;
+
+static int save_result(void)
+{
+ struct perf_pmu *pmu = NULL;
+ struct list_head *list;
+ struct pmu_scan_result *r;
+
+ while ((pmu = perf_pmus__scan(pmu)) != NULL) {
+ r = realloc(results, (nr_pmus + 1) * sizeof(*r));
+ if (r == NULL)
+ return -ENOMEM;
+
+ results = r;
+ r = results + nr_pmus;
+
+ r->name = strdup(pmu->name);
+ r->is_core = pmu->is_core;
+ r->nr_caps = pmu->nr_caps;
+
+ r->nr_aliases = perf_pmu__num_events(pmu);
+
+ r->nr_formats = 0;
+ list_for_each(list, &pmu->format)
+ r->nr_formats++;
+
+ pr_debug("pmu[%d] name=%s, nr_caps=%d, nr_aliases=%d, nr_formats=%d\n",
+ nr_pmus, r->name, r->nr_caps, r->nr_aliases, r->nr_formats);
+ nr_pmus++;
+ }
+
+ perf_pmus__destroy();
+ return 0;
+}
+
+static int check_result(bool core_only)
+{
+ struct pmu_scan_result *r;
+ struct perf_pmu *pmu;
+ struct list_head *list;
+ int nr;
+
+ for (int i = 0; i < nr_pmus; i++) {
+ r = &results[i];
+ if (core_only && !r->is_core)
+ continue;
+
+ pmu = perf_pmus__find(r->name);
+ if (pmu == NULL) {
+ pr_err("Cannot find PMU %s\n", r->name);
+ return -1;
+ }
+
+ if (pmu->nr_caps != (u32)r->nr_caps) {
+ pr_err("Unmatched number of event caps in %s: expect %d vs got %d\n",
+ pmu->name, r->nr_caps, pmu->nr_caps);
+ return -1;
+ }
+
+ nr = perf_pmu__num_events(pmu);
+ if (nr != r->nr_aliases) {
+ pr_err("Unmatched number of event aliases in %s: expect %d vs got %d\n",
+ pmu->name, r->nr_aliases, nr);
+ return -1;
+ }
+
+ nr = 0;
+ list_for_each(list, &pmu->format)
+ nr++;
+ if (nr != r->nr_formats) {
+ pr_err("Unmatched number of event formats in %s: expect %d vs got %d\n",
+ pmu->name, r->nr_formats, nr);
+ return -1;
+ }
+ }
+ return 0;
+}
+
+static void delete_result(void)
+{
+ for (int i = 0; i < nr_pmus; i++)
+ free(results[i].name);
+ free(results);
+
+ results = NULL;
+ nr_pmus = 0;
+}
+
+static int run_pmu_scan(void)
+{
+ struct stats stats;
+ struct timeval start, end, diff;
+ double time_average, time_stddev;
+ u64 runtime_us;
+ int ret;
+
+ init_stats(&stats);
+ pr_info("Computing performance of sysfs PMU event scan for %u times\n",
+ iterations);
+
+ if (save_result() < 0) {
+ pr_err("Failed to initialize PMU scan result\n");
+ return -1;
+ }
+
+ for (int j = 0; j < 2; j++) {
+ bool core_only = (j == 0);
+
+ for (unsigned int i = 0; i < iterations; i++) {
+ gettimeofday(&start, NULL);
+ if (core_only)
+ perf_pmus__scan_core(NULL);
+ else
+ perf_pmus__scan(NULL);
+ gettimeofday(&end, NULL);
+ timersub(&end, &start, &diff);
+ runtime_us = diff.tv_sec * USEC_PER_SEC + diff.tv_usec;
+ update_stats(&stats, runtime_us);
+
+ ret = check_result(core_only);
+ perf_pmus__destroy();
+ if (ret < 0)
+ break;
+ }
+ time_average = avg_stats(&stats);
+ time_stddev = stddev_stats(&stats);
+ pr_info(" Average%s PMU scanning took: %.3f usec (+- %.3f usec)\n",
+ core_only ? " core" : "", time_average, time_stddev);
+ }
+ delete_result();
+ return 0;
+}
+
+int bench_pmu_scan(int argc, const char **argv)
+{
+ int err = 0;
+
+ argc = parse_options(argc, argv, options, bench_usage, 0);
+ if (argc) {
+ usage_with_options(bench_usage, options);
+ exit(EXIT_FAILURE);
+ }
+
+ err = run_pmu_scan();
+
+ return err;
+}
diff --git a/tools/perf/bench/sched-messaging.c b/tools/perf/bench/sched-messaging.c
index 71d830d7b923..93dcd9dba3d0 100644
--- a/tools/perf/bench/sched-messaging.c
+++ b/tools/perf/bench/sched-messaging.c
@@ -27,6 +27,7 @@
#include <poll.h>
#include <limits.h>
#include <err.h>
+#include <linux/list.h>
#include <linux/time64.h>
#define DATASIZE 100
@@ -35,8 +36,12 @@ static bool use_pipes = false;
static unsigned int nr_loops = 100;
static bool thread_mode = false;
static unsigned int num_groups = 10;
+static unsigned int total_children = 0;
+static struct list_head sender_contexts = LIST_HEAD_INIT(sender_contexts);
+static struct list_head receiver_contexts = LIST_HEAD_INIT(receiver_contexts);
struct sender_context {
+ struct list_head list;
unsigned int num_fds;
int ready_out;
int wakefd;
@@ -44,12 +49,20 @@ struct sender_context {
};
struct receiver_context {
+ struct list_head list;
unsigned int num_packets;
int in_fds[2];
int ready_out;
int wakefd;
};
+union messaging_worker {
+ pthread_t thread;
+ pid_t pid;
+};
+
+static union messaging_worker *worker_tab;
+
static void fdpair(int fds[2])
{
if (use_pipes) {
@@ -66,11 +79,10 @@ static void fdpair(int fds[2])
/* Block until we're ready to go */
static void ready(int ready_out, int wakefd)
{
- char dummy;
struct pollfd pollfd = { .fd = wakefd, .events = POLLIN };
/* Tell them we're ready. */
- if (write(ready_out, &dummy, 1) != 1)
+ if (write(ready_out, "R", 1) != 1)
err(EXIT_FAILURE, "CLIENT: ready write");
/* Wait for "GO" signal */
@@ -85,6 +97,7 @@ static void *sender(struct sender_context *ctx)
unsigned int i, j;
ready(ctx->ready_out, ctx->wakefd);
+ memset(data, 'S', sizeof(data));
/* Now pump to every receiver. */
for (i = 0; i < nr_loops; i++) {
@@ -93,7 +106,7 @@ static void *sender(struct sender_context *ctx)
again:
ret = write(ctx->out_fds[j], data + done,
- sizeof(data)-done);
+ sizeof(data) - done);
if (ret < 0)
err(EXIT_FAILURE, "SENDER: write");
done += ret;
@@ -134,30 +147,12 @@ again:
return NULL;
}
-static pthread_t create_worker(void *ctx, void *(*func)(void *))
+static void create_thread_worker(union messaging_worker *worker,
+ void *ctx, void *(*func)(void *))
{
pthread_attr_t attr;
- pthread_t childid;
int ret;
- if (!thread_mode) {
- /* process mode */
- /* Fork the receiver. */
- switch (fork()) {
- case -1:
- err(EXIT_FAILURE, "fork()");
- break;
- case 0:
- (*func) (ctx);
- exit(0);
- break;
- default:
- break;
- }
-
- return (pthread_t)0;
- }
-
if (pthread_attr_init(&attr) != 0)
err(EXIT_FAILURE, "pthread_attr_init:");
@@ -166,14 +161,37 @@ static pthread_t create_worker(void *ctx, void *(*func)(void *))
err(EXIT_FAILURE, "pthread_attr_setstacksize");
#endif
- ret = pthread_create(&childid, &attr, func, ctx);
+ ret = pthread_create(&worker->thread, &attr, func, ctx);
if (ret != 0)
err(EXIT_FAILURE, "pthread_create failed");
- return childid;
+ pthread_attr_destroy(&attr);
+}
+
+static void create_process_worker(union messaging_worker *worker,
+ void *ctx, void *(*func)(void *))
+{
+ /* Fork the receiver. */
+ worker->pid = fork();
+
+ if (worker->pid == -1) {
+ err(EXIT_FAILURE, "fork()");
+ } else if (worker->pid == 0) {
+ (*func) (ctx);
+ exit(0);
+ }
}
-static void reap_worker(pthread_t id)
+static void create_worker(union messaging_worker *worker,
+ void *ctx, void *(*func)(void *))
+{
+ if (!thread_mode)
+ return create_process_worker(worker, ctx, func);
+ else
+ return create_thread_worker(worker, ctx, func);
+}
+
+static void reap_worker(union messaging_worker *worker)
{
int proc_status;
void *thread_status;
@@ -184,23 +202,24 @@ static void reap_worker(pthread_t id)
if (!WIFEXITED(proc_status))
exit(1);
} else {
- pthread_join(id, &thread_status);
+ pthread_join(worker->thread, &thread_status);
}
}
/* One group of senders and receivers */
-static unsigned int group(pthread_t *pth,
+static unsigned int group(union messaging_worker *worker,
unsigned int num_fds,
int ready_out,
int wakefd)
{
unsigned int i;
- struct sender_context *snd_ctx = malloc(sizeof(struct sender_context)
- + num_fds * sizeof(int));
+ struct sender_context *snd_ctx = malloc(sizeof(struct sender_context) +
+ num_fds * sizeof(int));
if (!snd_ctx)
err(EXIT_FAILURE, "malloc()");
+ list_add(&snd_ctx->list, &sender_contexts);
for (i = 0; i < num_fds; i++) {
int fds[2];
struct receiver_context *ctx = malloc(sizeof(*ctx));
@@ -208,6 +227,7 @@ static unsigned int group(pthread_t *pth,
if (!ctx)
err(EXIT_FAILURE, "malloc()");
+ list_add(&ctx->list, &receiver_contexts);
/* Create the pipe between client and server */
fdpair(fds);
@@ -218,7 +238,7 @@ static unsigned int group(pthread_t *pth,
ctx->ready_out = ready_out;
ctx->wakefd = wakefd;
- pth[i] = create_worker(ctx, (void *)receiver);
+ create_worker(worker + i, ctx, (void *)receiver);
snd_ctx->out_fds[i] = fds[1];
if (!thread_mode)
@@ -231,7 +251,7 @@ static unsigned int group(pthread_t *pth,
snd_ctx->wakefd = wakefd;
snd_ctx->num_fds = num_fds;
- pth[num_fds+i] = create_worker(snd_ctx, (void *)sender);
+ create_worker(worker + num_fds + i, snd_ctx, (void *)sender);
}
/* Close the fds we have left */
@@ -243,6 +263,17 @@ static unsigned int group(pthread_t *pth,
return num_fds * 2;
}
+static void sig_handler(int sig __maybe_unused)
+{
+ unsigned int i;
+
+ /*
+ * When exit abnormally, kill all forked child processes.
+ */
+ for (i = 0; i < total_children; i++)
+ kill(worker_tab[i].pid, SIGKILL);
+}
+
static const struct option options[] = {
OPT_BOOLEAN('p', "pipe", &use_pipes,
"Use pipe() instead of socketpair()"),
@@ -260,26 +291,30 @@ static const char * const bench_sched_message_usage[] = {
int bench_sched_messaging(int argc, const char **argv)
{
- unsigned int i, total_children;
+ unsigned int i;
struct timeval start, stop, diff;
unsigned int num_fds = 20;
int readyfds[2], wakefds[2];
char dummy;
- pthread_t *pth_tab;
+ struct sender_context *pos, *n;
argc = parse_options(argc, argv, options,
bench_sched_message_usage, 0);
- pth_tab = malloc(num_fds * 2 * num_groups * sizeof(pthread_t));
- if (!pth_tab)
+ worker_tab = malloc(num_fds * 2 * num_groups * sizeof(union messaging_worker));
+ if (!worker_tab)
err(EXIT_FAILURE, "main:malloc()");
fdpair(readyfds);
fdpair(wakefds);
- total_children = 0;
+ if (!thread_mode) {
+ signal(SIGINT, sig_handler);
+ signal(SIGTERM, sig_handler);
+ }
+
for (i = 0; i < num_groups; i++)
- total_children += group(pth_tab+total_children, num_fds,
+ total_children += group(worker_tab + total_children, num_fds,
readyfds[1], wakefds[0]);
/* Wait for everyone to be ready */
@@ -295,7 +330,7 @@ int bench_sched_messaging(int argc, const char **argv)
/* Reap them all */
for (i = 0; i < total_children; i++)
- reap_worker(pth_tab[i]);
+ reap_worker(worker_tab + i);
gettimeofday(&stop, NULL);
@@ -309,11 +344,11 @@ int bench_sched_messaging(int argc, const char **argv)
num_groups, num_groups * 2 * num_fds,
thread_mode ? "threads" : "processes");
printf(" %14s: %lu.%03lu [sec]\n", "Total time",
- diff.tv_sec,
+ (unsigned long) diff.tv_sec,
(unsigned long) (diff.tv_usec / USEC_PER_MSEC));
break;
case BENCH_FORMAT_SIMPLE:
- printf("%lu.%03lu\n", diff.tv_sec,
+ printf("%lu.%03lu\n", (unsigned long) diff.tv_sec,
(unsigned long) (diff.tv_usec / USEC_PER_MSEC));
break;
default:
@@ -323,7 +358,14 @@ int bench_sched_messaging(int argc, const char **argv)
break;
}
- free(pth_tab);
-
+ free(worker_tab);
+ list_for_each_entry_safe(pos, n, &sender_contexts, list) {
+ list_del_init(&pos->list);
+ free(pos);
+ }
+ list_for_each_entry_safe(pos, n, &receiver_contexts, list) {
+ list_del_init(&pos->list);
+ free(pos);
+ }
return 0;
}
diff --git a/tools/perf/bench/sched-pipe.c b/tools/perf/bench/sched-pipe.c
index 3c88d1f201f1..70139036d68f 100644
--- a/tools/perf/bench/sched-pipe.c
+++ b/tools/perf/bench/sched-pipe.c
@@ -10,7 +10,9 @@
* Ported to perf by Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
*/
#include <subcmd/parse-options.h>
+#include <api/fs/fs.h>
#include "bench.h"
+#include "util/cgroup.h"
#include <unistd.h>
#include <stdio.h>
@@ -19,7 +21,9 @@
#include <sys/wait.h>
#include <string.h>
#include <errno.h>
+#include <fcntl.h>
#include <assert.h>
+#include <sys/epoll.h>
#include <sys/time.h>
#include <sys/types.h>
#include <sys/syscall.h>
@@ -31,6 +35,9 @@ struct thread_data {
int nr;
int pipe_read;
int pipe_write;
+ struct epoll_event epoll_ev;
+ int epoll_fd;
+ bool cgroup_failed;
pthread_t pthread;
};
@@ -40,9 +47,50 @@ static int loops = LOOPS_DEFAULT;
/* Use processes by default: */
static bool threaded;
+static bool nonblocking;
+static char *cgrp_names[2];
+static struct cgroup *cgrps[2];
+
+static int parse_two_cgroups(const struct option *opt __maybe_unused,
+ const char *str, int unset __maybe_unused)
+{
+ char *p = strdup(str);
+ char *q;
+ int ret = -1;
+
+ if (p == NULL) {
+ fprintf(stderr, "memory allocation failure\n");
+ return -1;
+ }
+
+ q = strchr(p, ',');
+ if (q == NULL) {
+ fprintf(stderr, "it should have two cgroup names: %s\n", p);
+ goto out;
+ }
+ *q = '\0';
+
+ cgrp_names[0] = strdup(p);
+ cgrp_names[1] = strdup(q + 1);
+
+ if (cgrp_names[0] == NULL || cgrp_names[1] == NULL) {
+ fprintf(stderr, "memory allocation failure\n");
+ goto out;
+ }
+ ret = 0;
+
+out:
+ free(p);
+ return ret;
+}
+
static const struct option options[] = {
+ OPT_BOOLEAN('n', "nonblocking", &nonblocking, "Use non-blocking operations"),
OPT_INTEGER('l', "loop", &loops, "Specify number of loops"),
OPT_BOOLEAN('T', "threaded", &threaded, "Specify threads/process based task setup"),
+ OPT_CALLBACK('G', "cgroups", NULL, "SEND,RECV",
+ "Put sender and receivers in given cgroups",
+ parse_two_cgroups),
OPT_END()
};
@@ -51,24 +99,115 @@ static const char * const bench_sched_pipe_usage[] = {
NULL
};
+static int enter_cgroup(int nr)
+{
+ char buf[32];
+ int fd, len, ret;
+ int saved_errno;
+ struct cgroup *cgrp;
+ pid_t pid;
+
+ if (cgrp_names[nr] == NULL)
+ return 0;
+
+ if (cgrps[nr] == NULL) {
+ cgrps[nr] = cgroup__new(cgrp_names[nr], /*do_open=*/true);
+ if (cgrps[nr] == NULL)
+ goto err;
+ }
+ cgrp = cgrps[nr];
+
+ if (threaded)
+ pid = syscall(__NR_gettid);
+ else
+ pid = getpid();
+
+ snprintf(buf, sizeof(buf), "%d\n", pid);
+ len = strlen(buf);
+
+ /* try cgroup v2 interface first */
+ if (threaded)
+ fd = openat(cgrp->fd, "cgroup.threads", O_WRONLY);
+ else
+ fd = openat(cgrp->fd, "cgroup.procs", O_WRONLY);
+
+ /* try cgroup v1 if failed */
+ if (fd < 0 && errno == ENOENT)
+ fd = openat(cgrp->fd, "tasks", O_WRONLY);
+
+ if (fd < 0)
+ goto err;
+
+ ret = write(fd, buf, len);
+ close(fd);
+
+ if (ret != len) {
+ printf("Cannot enter to cgroup: %s\n", cgrp->name);
+ return -1;
+ }
+ return 0;
+
+err:
+ saved_errno = errno;
+ printf("Failed to open cgroup file in %s\n", cgrp_names[nr]);
+
+ if (saved_errno == ENOENT) {
+ char mnt[PATH_MAX];
+
+ if (cgroupfs_find_mountpoint(mnt, sizeof(mnt), "perf_event") == 0)
+ printf(" Hint: create the cgroup first, like 'mkdir %s/%s'\n",
+ mnt, cgrp_names[nr]);
+ } else if (saved_errno == EACCES && geteuid() > 0) {
+ printf(" Hint: try to run as root\n");
+ }
+
+ return -1;
+}
+
+static void exit_cgroup(int nr)
+{
+ cgroup__put(cgrps[nr]);
+ free(cgrp_names[nr]);
+}
+
+static inline int read_pipe(struct thread_data *td)
+{
+ int ret, m;
+retry:
+ if (nonblocking) {
+ ret = epoll_wait(td->epoll_fd, &td->epoll_ev, 1, -1);
+ if (ret < 0)
+ return ret;
+ }
+ ret = read(td->pipe_read, &m, sizeof(int));
+ if (nonblocking && ret < 0 && errno == EWOULDBLOCK)
+ goto retry;
+ return ret;
+}
+
static void *worker_thread(void *__tdata)
{
struct thread_data *td = __tdata;
- int m = 0, i;
- int ret;
+ int i, ret, m = 0;
+
+ ret = enter_cgroup(td->nr);
+ if (ret < 0) {
+ td->cgroup_failed = true;
+ return NULL;
+ }
+
+ if (nonblocking) {
+ td->epoll_ev.events = EPOLLIN;
+ td->epoll_fd = epoll_create(1);
+ BUG_ON(td->epoll_fd < 0);
+ BUG_ON(epoll_ctl(td->epoll_fd, EPOLL_CTL_ADD, td->pipe_read, &td->epoll_ev) < 0);
+ }
for (i = 0; i < loops; i++) {
- if (!td->nr) {
- ret = read(td->pipe_read, &m, sizeof(int));
- BUG_ON(ret != sizeof(int));
- ret = write(td->pipe_write, &m, sizeof(int));
- BUG_ON(ret != sizeof(int));
- } else {
- ret = write(td->pipe_write, &m, sizeof(int));
- BUG_ON(ret != sizeof(int));
- ret = read(td->pipe_read, &m, sizeof(int));
- BUG_ON(ret != sizeof(int));
- }
+ ret = write(td->pipe_write, &m, sizeof(int));
+ BUG_ON(ret != sizeof(int));
+ ret = read_pipe(td);
+ BUG_ON(ret != sizeof(int));
}
return NULL;
@@ -76,7 +215,8 @@ static void *worker_thread(void *__tdata)
int bench_sched_pipe(int argc, const char **argv)
{
- struct thread_data threads[2], *td;
+ struct thread_data threads[2] = {};
+ struct thread_data *td;
int pipe_1[2], pipe_2[2];
struct timeval start, stop, diff;
unsigned long long result_usec = 0;
@@ -88,13 +228,16 @@ int bench_sched_pipe(int argc, const char **argv)
* discarding returned value of read(), write()
* causes error in building environment for perf
*/
- int __maybe_unused ret, wait_stat;
+ int __maybe_unused ret, wait_stat, flags = 0;
pid_t pid, retpid __maybe_unused;
argc = parse_options(argc, argv, options, bench_sched_pipe_usage, 0);
- BUG_ON(pipe(pipe_1));
- BUG_ON(pipe(pipe_2));
+ if (nonblocking)
+ flags |= O_NONBLOCK;
+
+ BUG_ON(pipe2(pipe_1, flags));
+ BUG_ON(pipe2(pipe_2, flags));
gettimeofday(&start, NULL);
@@ -112,9 +255,7 @@ int bench_sched_pipe(int argc, const char **argv)
}
}
-
if (threaded) {
-
for (t = 0; t < nr_threads; t++) {
td = threads + t;
@@ -128,7 +269,6 @@ int bench_sched_pipe(int argc, const char **argv)
ret = pthread_join(td->pthread, NULL);
BUG_ON(ret);
}
-
} else {
pid = fork();
assert(pid >= 0);
@@ -147,6 +287,12 @@ int bench_sched_pipe(int argc, const char **argv)
gettimeofday(&stop, NULL);
timersub(&stop, &start, &diff);
+ exit_cgroup(0);
+ exit_cgroup(1);
+
+ if (threads[0].cgroup_failed || threads[1].cgroup_failed)
+ return 0;
+
switch (bench_format) {
case BENCH_FORMAT_DEFAULT:
printf("# Executed %d pipe operations between two %s\n\n",
@@ -156,7 +302,7 @@ int bench_sched_pipe(int argc, const char **argv)
result_usec += diff.tv_usec;
printf(" %14s: %lu.%03lu [sec]\n\n", "Total time",
- diff.tv_sec,
+ (unsigned long) diff.tv_sec,
(unsigned long) (diff.tv_usec / USEC_PER_MSEC));
printf(" %14lf usecs/op\n",
@@ -168,7 +314,7 @@ int bench_sched_pipe(int argc, const char **argv)
case BENCH_FORMAT_SIMPLE:
printf("%lu.%03lu\n",
- diff.tv_sec,
+ (unsigned long) diff.tv_sec,
(unsigned long) (diff.tv_usec / USEC_PER_MSEC));
break;
diff --git a/tools/perf/bench/sched-seccomp-notify.c b/tools/perf/bench/sched-seccomp-notify.c
new file mode 100644
index 000000000000..269c1f4a6852
--- /dev/null
+++ b/tools/perf/bench/sched-seccomp-notify.c
@@ -0,0 +1,178 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <subcmd/parse-options.h>
+#include "bench.h"
+
+#include <uapi/linux/filter.h>
+#include <sys/types.h>
+#include <sys/time.h>
+#include <linux/unistd.h>
+#include <sys/syscall.h>
+#include <sys/ioctl.h>
+#include <linux/time64.h>
+#include <uapi/linux/seccomp.h>
+#include <sys/prctl.h>
+
+#include <unistd.h>
+#include <limits.h>
+#include <stddef.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <sys/wait.h>
+#include <string.h>
+#include <errno.h>
+#include <err.h>
+#include <inttypes.h>
+
+#define LOOPS_DEFAULT 1000000UL
+static uint64_t loops = LOOPS_DEFAULT;
+static bool sync_mode;
+
+static const struct option options[] = {
+ OPT_U64('l', "loop", &loops, "Specify number of loops"),
+ OPT_BOOLEAN('s', "sync-mode", &sync_mode,
+ "Enable the synchronous mode for seccomp notifications"),
+ OPT_END()
+};
+
+static const char * const bench_seccomp_usage[] = {
+ "perf bench sched secccomp-notify <options>",
+ NULL
+};
+
+static int seccomp(unsigned int op, unsigned int flags, void *args)
+{
+ return syscall(__NR_seccomp, op, flags, args);
+}
+
+static int user_notif_syscall(int nr, unsigned int flags)
+{
+ struct sock_filter filter[] = {
+ BPF_STMT(BPF_LD|BPF_W|BPF_ABS,
+ offsetof(struct seccomp_data, nr)),
+ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, nr, 0, 1),
+ BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_USER_NOTIF),
+ BPF_STMT(BPF_RET|BPF_K, SECCOMP_RET_ALLOW),
+ };
+
+ struct sock_fprog prog = {
+ .len = (unsigned short)ARRAY_SIZE(filter),
+ .filter = filter,
+ };
+
+ return seccomp(SECCOMP_SET_MODE_FILTER, flags, &prog);
+}
+
+#define USER_NOTIF_MAGIC INT_MAX
+static void user_notification_sync_loop(int listener)
+{
+ struct seccomp_notif_resp resp;
+ struct seccomp_notif req;
+ uint64_t nr;
+
+ for (nr = 0; nr < loops; nr++) {
+ memset(&req, 0, sizeof(req));
+ if (ioctl(listener, SECCOMP_IOCTL_NOTIF_RECV, &req))
+ err(EXIT_FAILURE, "SECCOMP_IOCTL_NOTIF_RECV failed");
+
+ if (req.data.nr != __NR_gettid)
+ errx(EXIT_FAILURE, "unexpected syscall: %d", req.data.nr);
+
+ resp.id = req.id;
+ resp.error = 0;
+ resp.val = USER_NOTIF_MAGIC;
+ resp.flags = 0;
+ if (ioctl(listener, SECCOMP_IOCTL_NOTIF_SEND, &resp))
+ err(EXIT_FAILURE, "SECCOMP_IOCTL_NOTIF_SEND failed");
+ }
+}
+
+#ifndef SECCOMP_USER_NOTIF_FD_SYNC_WAKE_UP
+#define SECCOMP_USER_NOTIF_FD_SYNC_WAKE_UP (1UL << 0)
+#define SECCOMP_IOCTL_NOTIF_SET_FLAGS SECCOMP_IOW(4, __u64)
+#endif
+int bench_sched_seccomp_notify(int argc, const char **argv)
+{
+ struct timeval start, stop, diff;
+ unsigned long long result_usec = 0;
+ int status, listener;
+ pid_t pid;
+ long ret;
+
+ argc = parse_options(argc, argv, options, bench_seccomp_usage, 0);
+
+ gettimeofday(&start, NULL);
+
+ prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0);
+ listener = user_notif_syscall(__NR_gettid,
+ SECCOMP_FILTER_FLAG_NEW_LISTENER);
+ if (listener < 0)
+ err(EXIT_FAILURE, "can't create a notification descriptor");
+
+ pid = fork();
+ if (pid < 0)
+ err(EXIT_FAILURE, "fork");
+ if (pid == 0) {
+ if (prctl(PR_SET_PDEATHSIG, SIGKILL, 0, 0, 0))
+ err(EXIT_FAILURE, "can't set the parent death signal");
+ while (1) {
+ ret = syscall(__NR_gettid);
+ if (ret == USER_NOTIF_MAGIC)
+ continue;
+ break;
+ }
+ _exit(1);
+ }
+
+ if (sync_mode) {
+ if (ioctl(listener, SECCOMP_IOCTL_NOTIF_SET_FLAGS,
+ SECCOMP_USER_NOTIF_FD_SYNC_WAKE_UP, 0))
+ err(EXIT_FAILURE,
+ "can't set SECCOMP_USER_NOTIF_FD_SYNC_WAKE_UP");
+ }
+ user_notification_sync_loop(listener);
+
+ kill(pid, SIGKILL);
+ if (waitpid(pid, &status, 0) != pid)
+ err(EXIT_FAILURE, "waitpid(%d) failed", pid);
+ if (!WIFSIGNALED(status) || WTERMSIG(status) != SIGKILL)
+ errx(EXIT_FAILURE, "unexpected exit code: %d", status);
+
+ gettimeofday(&stop, NULL);
+ timersub(&stop, &start, &diff);
+
+ switch (bench_format) {
+ case BENCH_FORMAT_DEFAULT:
+ printf("# Executed %" PRIu64 " system calls\n\n",
+ loops);
+
+ result_usec = diff.tv_sec * USEC_PER_SEC;
+ result_usec += diff.tv_usec;
+
+ printf(" %14s: %lu.%03lu [sec]\n\n", "Total time",
+ (unsigned long) diff.tv_sec,
+ (unsigned long) (diff.tv_usec / USEC_PER_MSEC));
+
+ printf(" %14lf usecs/op\n",
+ (double)result_usec / (double)loops);
+ printf(" %14d ops/sec\n",
+ (int)((double)loops /
+ ((double)result_usec / (double)USEC_PER_SEC)));
+ break;
+
+ case BENCH_FORMAT_SIMPLE:
+ printf("%lu.%03lu\n",
+ (unsigned long) diff.tv_sec,
+ (unsigned long) (diff.tv_usec / USEC_PER_MSEC));
+ break;
+
+ default:
+ /* reaching here is something disaster */
+ fprintf(stderr, "Unknown format:%d\n", bench_format);
+ exit(1);
+ break;
+ }
+
+ return 0;
+}
diff --git a/tools/perf/bench/synthesize.c b/tools/perf/bench/synthesize.c
index 8d624aea1c5e..9b333276cbdb 100644
--- a/tools/perf/bench/synthesize.c
+++ b/tools/perf/bench/synthesize.c
@@ -49,7 +49,7 @@ static const char *const bench_usage[] = {
static atomic_t event_count;
-static int process_synthesized_event(struct perf_tool *tool __maybe_unused,
+static int process_synthesized_event(const struct perf_tool *tool __maybe_unused,
union perf_event *event __maybe_unused,
struct perf_sample *sample __maybe_unused,
struct machine *machine __maybe_unused)
@@ -80,7 +80,7 @@ static int do_run_single_threaded(struct perf_session *session,
NULL,
target, threads,
process_synthesized_event,
- data_mmap,
+ true, data_mmap,
nr_threads_synthesize);
if (err)
return err;
@@ -117,7 +117,7 @@ static int run_single_threaded(void)
int err;
perf_set_singlethreaded();
- session = perf_session__new(NULL, false, NULL);
+ session = perf_session__new(NULL, NULL);
if (IS_ERR(session)) {
pr_err("Session creation failed.\n");
return PTR_ERR(session);
@@ -161,9 +161,9 @@ static int do_run_multi_threaded(struct target *target,
init_stats(&time_stats);
init_stats(&event_stats);
for (i = 0; i < multi_iterations; i++) {
- session = perf_session__new(NULL, false, NULL);
- if (!session)
- return -ENOMEM;
+ session = perf_session__new(NULL, NULL);
+ if (IS_ERR(session))
+ return PTR_ERR(session);
atomic_set(&event_count, 0);
gettimeofday(&start, NULL);
@@ -171,7 +171,7 @@ static int do_run_multi_threaded(struct target *target,
NULL,
target, NULL,
process_synthesized_event,
- false,
+ true, false,
nr_threads_synthesize);
if (err) {
perf_session__delete(session);
diff --git a/tools/perf/bench/syscall.c b/tools/perf/bench/syscall.c
new file mode 100644
index 000000000000..e7dc216f717f
--- /dev/null
+++ b/tools/perf/bench/syscall.c
@@ -0,0 +1,188 @@
+/*
+ *
+ * syscall.c
+ *
+ * syscall: Benchmark for system call performance
+ */
+#include "../perf.h"
+#include "../util/util.h"
+#include <subcmd/parse-options.h>
+#include "../builtin.h"
+#include "bench.h"
+
+#include <stdio.h>
+#include <sys/time.h>
+#include <sys/syscall.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+#include <stdlib.h>
+
+#ifndef __NR_fork
+#define __NR_fork -1
+#endif
+
+static int loops;
+
+static const struct option options[] = {
+ OPT_INTEGER('l', "loop", &loops, "Specify number of loops"),
+ OPT_END()
+};
+
+static const char * const bench_syscall_usage[] = {
+ "perf bench syscall <options>",
+ NULL
+};
+
+static void test_fork(void)
+{
+ pid_t pid = fork();
+
+ if (pid < 0) {
+ fprintf(stderr, "fork failed\n");
+ exit(1);
+ } else if (pid == 0) {
+ exit(0);
+ } else {
+ if (waitpid(pid, NULL, 0) < 0) {
+ fprintf(stderr, "waitpid failed\n");
+ exit(1);
+ }
+ }
+}
+
+static void test_execve(void)
+{
+ const char *pathname = "/bin/true";
+ char *const argv[] = { (char *)pathname, NULL };
+ pid_t pid = fork();
+
+ if (pid < 0) {
+ fprintf(stderr, "fork failed\n");
+ exit(1);
+ } else if (pid == 0) {
+ execve(pathname, argv, NULL);
+ fprintf(stderr, "execve /bin/true failed\n");
+ exit(1);
+ } else {
+ if (waitpid(pid, NULL, 0) < 0) {
+ fprintf(stderr, "waitpid failed\n");
+ exit(1);
+ }
+ }
+}
+
+static int bench_syscall_common(int argc, const char **argv, int syscall)
+{
+ struct timeval start, stop, diff;
+ unsigned long long result_usec = 0;
+ const char *name = NULL;
+ int i;
+
+ switch (syscall) {
+ case __NR_fork:
+ case __NR_execve:
+ /* Limit default loop to 10000 times to save time */
+ loops = 10000;
+ break;
+ default:
+ loops = 10000000;
+ break;
+ }
+
+ /* Options -l and --loops override default above */
+ argc = parse_options(argc, argv, options, bench_syscall_usage, 0);
+
+ gettimeofday(&start, NULL);
+
+ for (i = 0; i < loops; i++) {
+ switch (syscall) {
+ case __NR_getppid:
+ getppid();
+ break;
+ case __NR_getpgid:
+ getpgid(0);
+ break;
+ case __NR_fork:
+ test_fork();
+ break;
+ case __NR_execve:
+ test_execve();
+ default:
+ break;
+ }
+ }
+
+ gettimeofday(&stop, NULL);
+ timersub(&stop, &start, &diff);
+
+ switch (syscall) {
+ case __NR_getppid:
+ name = "getppid()";
+ break;
+ case __NR_getpgid:
+ name = "getpgid()";
+ break;
+ case __NR_fork:
+ name = "fork()";
+ break;
+ case __NR_execve:
+ name = "execve()";
+ break;
+ default:
+ break;
+ }
+
+ switch (bench_format) {
+ case BENCH_FORMAT_DEFAULT:
+ printf("# Executed %'d %s calls\n", loops, name);
+
+ result_usec = diff.tv_sec * 1000000;
+ result_usec += diff.tv_usec;
+
+ printf(" %14s: %lu.%03lu [sec]\n\n", "Total time",
+ (unsigned long) diff.tv_sec,
+ (unsigned long) (diff.tv_usec/1000));
+
+ printf(" %14lf usecs/op\n",
+ (double)result_usec / (double)loops);
+ printf(" %'14d ops/sec\n",
+ (int)((double)loops /
+ ((double)result_usec / (double)1000000)));
+ break;
+
+ case BENCH_FORMAT_SIMPLE:
+ printf("%lu.%03lu\n",
+ (unsigned long) diff.tv_sec,
+ (unsigned long) (diff.tv_usec / 1000));
+ break;
+
+ default:
+ /* reaching here is something disaster */
+ fprintf(stderr, "Unknown format:%d\n", bench_format);
+ exit(1);
+ break;
+ }
+
+ return 0;
+}
+
+int bench_syscall_basic(int argc, const char **argv)
+{
+ return bench_syscall_common(argc, argv, __NR_getppid);
+}
+
+int bench_syscall_getpgid(int argc, const char **argv)
+{
+ return bench_syscall_common(argc, argv, __NR_getpgid);
+}
+
+int bench_syscall_fork(int argc, const char **argv)
+{
+ return bench_syscall_common(argc, argv, __NR_fork);
+}
+
+int bench_syscall_execve(int argc, const char **argv)
+{
+ return bench_syscall_common(argc, argv, __NR_execve);
+}
diff --git a/tools/perf/bench/uprobe.c b/tools/perf/bench/uprobe.c
new file mode 100644
index 000000000000..0b90275862e1
--- /dev/null
+++ b/tools/perf/bench/uprobe.c
@@ -0,0 +1,213 @@
+// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+/*
+ * uprobe.c
+ *
+ * uprobe benchmarks
+ *
+ * Copyright (C) 2023, Red Hat Inc, Arnaldo Carvalho de Melo <acme@redhat.com>
+ */
+#include "../perf.h"
+#include "../util/util.h"
+#include <subcmd/parse-options.h>
+#include "../builtin.h"
+#include "bench.h"
+#include <linux/compiler.h>
+#include <linux/time64.h>
+
+#include <inttypes.h>
+#include <stdio.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <time.h>
+#include <unistd.h>
+#include <stdlib.h>
+
+#define LOOPS_DEFAULT 1000
+static int loops = LOOPS_DEFAULT;
+
+enum bench_uprobe {
+ BENCH_UPROBE__BASELINE,
+ BENCH_UPROBE__EMPTY,
+ BENCH_UPROBE__TRACE_PRINTK,
+ BENCH_UPROBE__EMPTY_RET,
+ BENCH_UPROBE__TRACE_PRINTK_RET,
+};
+
+static const struct option options[] = {
+ OPT_INTEGER('l', "loop", &loops, "Specify number of loops"),
+ OPT_END()
+};
+
+static const char * const bench_uprobe_usage[] = {
+ "perf bench uprobe <options>",
+ NULL
+};
+
+#ifdef HAVE_BPF_SKEL
+#include "bpf_skel/bench_uprobe.skel.h"
+
+#define bench_uprobe__attach_uprobe(prog) \
+ skel->links.prog = bpf_program__attach_uprobe_opts(/*prog=*/skel->progs.prog, \
+ /*pid=*/-1, \
+ /*binary_path=*/"libc.so.6", \
+ /*func_offset=*/0, \
+ /*opts=*/&uprobe_opts); \
+ if (!skel->links.prog) { \
+ err = -errno; \
+ fprintf(stderr, "Failed to attach bench uprobe \"%s\": %s\n", #prog, strerror(errno)); \
+ goto cleanup; \
+ }
+
+struct bench_uprobe_bpf *skel;
+
+static int bench_uprobe__setup_bpf_skel(enum bench_uprobe bench)
+{
+ DECLARE_LIBBPF_OPTS(bpf_uprobe_opts, uprobe_opts);
+ int err;
+
+ /* Load and verify BPF application */
+ skel = bench_uprobe_bpf__open();
+ if (!skel) {
+ fprintf(stderr, "Failed to open and load uprobes bench BPF skeleton\n");
+ return -1;
+ }
+
+ err = bench_uprobe_bpf__load(skel);
+ if (err) {
+ fprintf(stderr, "Failed to load and verify BPF skeleton\n");
+ goto cleanup;
+ }
+
+ uprobe_opts.func_name = "usleep";
+ switch (bench) {
+ case BENCH_UPROBE__BASELINE: break;
+ case BENCH_UPROBE__EMPTY: bench_uprobe__attach_uprobe(empty); break;
+ case BENCH_UPROBE__TRACE_PRINTK: bench_uprobe__attach_uprobe(trace_printk); break;
+ case BENCH_UPROBE__EMPTY_RET: bench_uprobe__attach_uprobe(empty_ret); break;
+ case BENCH_UPROBE__TRACE_PRINTK_RET: bench_uprobe__attach_uprobe(trace_printk_ret); break;
+ default:
+ fprintf(stderr, "Invalid bench: %d\n", bench);
+ goto cleanup;
+ }
+
+ return err;
+cleanup:
+ bench_uprobe_bpf__destroy(skel);
+ skel = NULL;
+ return err;
+}
+
+static void bench_uprobe__teardown_bpf_skel(void)
+{
+ if (skel) {
+ bench_uprobe_bpf__destroy(skel);
+ skel = NULL;
+ }
+}
+#else
+static int bench_uprobe__setup_bpf_skel(enum bench_uprobe bench __maybe_unused) { return 0; }
+static void bench_uprobe__teardown_bpf_skel(void) {};
+#endif
+
+static int bench_uprobe_format__default_fprintf(const char *name, const char *unit, u64 diff, FILE *fp)
+{
+ static u64 baseline, previous;
+ s64 diff_to_baseline = diff - baseline,
+ diff_to_previous = diff - previous;
+ int printed = fprintf(fp, "# Executed %'d %s calls\n", loops, name);
+
+ printed += fprintf(fp, " %14s: %'" PRIu64 " %ss", "Total time", diff, unit);
+
+ if (baseline) {
+ printed += fprintf(fp, " %s%'" PRId64 " to baseline", diff_to_baseline > 0 ? "+" : "", diff_to_baseline);
+
+ if (previous != baseline)
+ fprintf(stdout, " %s%'" PRId64 " to previous", diff_to_previous > 0 ? "+" : "", diff_to_previous);
+ }
+
+ printed += fprintf(fp, "\n\n %'.3f %ss/op", (double)diff / (double)loops, unit);
+
+ if (baseline) {
+ printed += fprintf(fp, " %'.3f %ss/op to baseline", (double)diff_to_baseline / (double)loops, unit);
+
+ if (previous != baseline)
+ printed += fprintf(fp, " %'.3f %ss/op to previous", (double)diff_to_previous / (double)loops, unit);
+ } else {
+ baseline = diff;
+ }
+
+ fputc('\n', fp);
+
+ previous = diff;
+
+ return printed + 1;
+}
+
+static int bench_uprobe(int argc, const char **argv, enum bench_uprobe bench)
+{
+ const char *name = "usleep(1000)", *unit = "usec";
+ struct timespec start, end;
+ u64 diff;
+ int i;
+
+ argc = parse_options(argc, argv, options, bench_uprobe_usage, 0);
+
+ if (bench != BENCH_UPROBE__BASELINE && bench_uprobe__setup_bpf_skel(bench) < 0)
+ return 0;
+
+ clock_gettime(CLOCK_REALTIME, &start);
+
+ for (i = 0; i < loops; i++) {
+ usleep(USEC_PER_MSEC);
+ }
+
+ clock_gettime(CLOCK_REALTIME, &end);
+
+ diff = end.tv_sec * NSEC_PER_SEC + end.tv_nsec - (start.tv_sec * NSEC_PER_SEC + start.tv_nsec);
+ diff /= NSEC_PER_USEC;
+
+ switch (bench_format) {
+ case BENCH_FORMAT_DEFAULT:
+ bench_uprobe_format__default_fprintf(name, unit, diff, stdout);
+ break;
+
+ case BENCH_FORMAT_SIMPLE:
+ printf("%" PRIu64 "\n", diff);
+ break;
+
+ default:
+ /* reaching here is something of a disaster */
+ fprintf(stderr, "Unknown format:%d\n", bench_format);
+ exit(1);
+ }
+
+ if (bench != BENCH_UPROBE__BASELINE)
+ bench_uprobe__teardown_bpf_skel();
+
+ return 0;
+}
+
+int bench_uprobe_baseline(int argc, const char **argv)
+{
+ return bench_uprobe(argc, argv, BENCH_UPROBE__BASELINE);
+}
+
+int bench_uprobe_empty(int argc, const char **argv)
+{
+ return bench_uprobe(argc, argv, BENCH_UPROBE__EMPTY);
+}
+
+int bench_uprobe_trace_printk(int argc, const char **argv)
+{
+ return bench_uprobe(argc, argv, BENCH_UPROBE__TRACE_PRINTK);
+}
+
+int bench_uprobe_empty_ret(int argc, const char **argv)
+{
+ return bench_uprobe(argc, argv, BENCH_UPROBE__EMPTY_RET);
+}
+
+int bench_uprobe_trace_printk_ret(int argc, const char **argv)
+{
+ return bench_uprobe(argc, argv, BENCH_UPROBE__TRACE_PRINTK_RET);
+}