perf evlist: Store pointer to the cpu and thread maps

So that we don't have to pass it around to the several methods that
needs it, simplifying usage.

There is one case where we don't have the thread/cpu map in advance,
which is in the parsing routines used by top, stat, record, that we have
to wait till all options are parsed to know if a cpu or thread list was
passed to then create those maps.

For that case consolidate the cpu and thread map creation via
perf_evlist__create_maps() out of the code in top and record, while also
providing a perf_evlist__set_maps() for cases where multiple evlists
share maps or for when maps that represent CPU sockets, for instance,
get crafted out of topology information or subsets of threads in a
particular application are to be monitored, providing more granularity
in specifying which cpus and threads to monitor.

Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Tom Zanussi <tzanussi@gmail.com>
LKML-Reference: <new-submission>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index edc3555098c8a0613d62c2a736edc891381eee73..07f8d6d852c296fe7dc5dd83ca44f2d6bbfb0c93 100644 (file)
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -42,7 +42,6 @@ static u64                    user_interval                   = ULLONG_MAX;
 static u64                     default_interval                =      0;
 static u64                     sample_type;
 
-static struct cpu_map          *cpus;
 static unsigned int            page_size;
 static unsigned int            mmap_pages                      =    128;
 static unsigned int            user_freq                       = UINT_MAX;
@@ -58,7 +57,6 @@ static bool                   sample_id_all_avail             =   true;
 static bool                    system_wide                     =  false;
 static pid_t                   target_pid                      =     -1;
 static pid_t                   target_tid                      =     -1;
-static struct thread_map       *threads;
 static pid_t                   child_pid                       =     -1;
 static bool                    no_inherit                      =  false;
 static enum write_mode_t       write_mode                      = WRITE_FORCE;
@@ -189,7 +187,7 @@ static void create_counter(struct perf_evsel *evsel, int cpu)
        int thread_index;
        int ret;
 
-       for (thread_index = 0; thread_index < threads->nr; thread_index++) {
+       for (thread_index = 0; thread_index < evsel_list->threads->nr; thread_index++) {
                h_attr = get_header_attr(attr, evsel->idx);
                if (h_attr == NULL)
                        die("nomem\n");
@@ -317,7 +315,8 @@ static void open_counters(struct perf_evlist *evlist)
 retry_sample_id:
                attr->sample_id_all = sample_id_all_avail ? 1 : 0;
 try_again:
-               if (perf_evsel__open(pos, cpus, threads, group, !no_inherit) < 0) {
+               if (perf_evsel__open(pos, evlist->cpus, evlist->threads, group,
+                                    !no_inherit) < 0) {
                        int err = errno;
 
                        if (err == EPERM || err == EACCES)
@@ -368,10 +367,10 @@ try_again:
                }
        }
 
-       if (perf_evlist__mmap(evlist, cpus, threads, mmap_pages, false) < 0)
+       if (perf_evlist__mmap(evlist, mmap_pages, false) < 0)
                die("failed to mmap with %d (%s)\n", errno, strerror(errno));
 
-       for (cpu = 0; cpu < cpus->nr; ++cpu) {
+       for (cpu = 0; cpu < evsel_list->cpus->nr; ++cpu) {
                list_for_each_entry(pos, &evlist->entries, node)
                        create_counter(pos, cpu);
        }
@@ -450,7 +449,7 @@ static void mmap_read_all(void)
 {
        int i;
 
-       for (i = 0; i < cpus->nr; i++) {
+       for (i = 0; i < evsel_list->cpus->nr; i++) {
                if (evsel_list->mmap[i].base)
                        mmap_read(&evsel_list->mmap[i]);
        }
@@ -584,7 +583,7 @@ static int __cmd_record(int argc, const char **argv)
                }
 
                if (!system_wide && target_tid == -1 && target_pid == -1)
-                       threads->map[0] = child_pid;
+                       evsel_list->threads->map[0] = child_pid;
 
                close(child_ready_pipe[1]);
                close(go_pipe[0]);
@@ -718,12 +717,12 @@ static int __cmd_record(int argc, const char **argv)
                }
 
                if (done) {
-                       for (i = 0; i < cpus->nr; i++) {
+                       for (i = 0; i < evsel_list->cpus->nr; i++) {
                                struct perf_evsel *pos;
 
                                list_for_each_entry(pos, &evsel_list->entries, node) {
                                        for (thread = 0;
-                                               thread < threads->nr;
+                                               thread < evsel_list->threads->nr;
                                                thread++)
                                                ioctl(FD(pos, i, thread),
                                                        PERF_EVENT_IOC_DISABLE);
@@ -816,7 +815,7 @@ int cmd_record(int argc, const char **argv, const char *prefix __used)
        int err = -ENOMEM;
        struct perf_evsel *pos;
 
-       evsel_list = perf_evlist__new();
+       evsel_list = perf_evlist__new(NULL, NULL);
        if (evsel_list == NULL)
                return -ENOMEM;
 
@@ -850,28 +849,19 @@ int cmd_record(int argc, const char **argv, const char *prefix __used)
        if (target_pid != -1)
                target_tid = target_pid;
 
-       threads = thread_map__new(target_pid, target_tid);
-       if (threads == NULL) {
-               pr_err("Problems finding threads of monitor\n");
-               usage_with_options(record_usage, record_options);
-       }
-
-       if (target_tid != -1)
-               cpus = cpu_map__dummy_new();
-       else
-               cpus = cpu_map__new(cpu_list);
-
-       if (cpus == NULL)
+       if (perf_evlist__create_maps(evsel_list, target_pid,
+                                    target_tid, cpu_list) < 0)
                usage_with_options(record_usage, record_options);
 
        list_for_each_entry(pos, &evsel_list->entries, node) {
-               if (perf_evsel__alloc_fd(pos, cpus->nr, threads->nr) < 0)
+               if (perf_evsel__alloc_fd(pos, evsel_list->cpus->nr,
+                                        evsel_list->threads->nr) < 0)
                        goto out_free_fd;
                if (perf_header__push_event(pos->attr.config, event_name(pos)))
                        goto out_free_fd;
        }
 
-       if (perf_evlist__alloc_pollfd(evsel_list, cpus->nr, threads->nr) < 0)
+       if (perf_evlist__alloc_pollfd(evsel_list) < 0)
                goto out_free_fd;
 
        if (user_interval != ULLONG_MAX)
@@ -893,10 +883,8 @@ int cmd_record(int argc, const char **argv, const char *prefix __used)
        }
 
        err = __cmd_record(argc, argv);
-
 out_free_fd:
-       thread_map__delete(threads);
-       threads = NULL;
+       perf_evlist__delete_maps(evsel_list);
 out_symbol_exit:
        symbol__exit();
        return err;

diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index 8906adfdbd8e38aa37791c74af6638b20c208b50..e0f95755361b5fcc10ac4a53fe97a6e221924a7c 100644 (file)
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -76,7 +76,6 @@ static struct perf_event_attr default_attrs[] = {
 struct perf_evlist             *evsel_list;
 
 static bool                    system_wide                     =  false;
-static struct cpu_map          *cpus;
 static int                     run_idx                         =  0;
 
 static int                     run_count                       =  1;
@@ -85,7 +84,6 @@ static bool                   scale                           =  true;
 static bool                    no_aggr                         = false;
 static pid_t                   target_pid                      = -1;
 static pid_t                   target_tid                      = -1;
-static struct thread_map       *threads;
 static pid_t                   child_pid                       = -1;
 static bool                    null_run                        =  false;
 static bool                    big_num                         =  true;
@@ -170,7 +168,7 @@ static int create_perf_stat_counter(struct perf_evsel *evsel)
                                    PERF_FORMAT_TOTAL_TIME_RUNNING;
 
        if (system_wide)
-               return perf_evsel__open_per_cpu(evsel, cpus, false, false);
+               return perf_evsel__open_per_cpu(evsel, evsel_list->cpus, false, false);
 
        attr->inherit = !no_inherit;
        if (target_pid == -1 && target_tid == -1) {
@@ -178,7 +176,7 @@ static int create_perf_stat_counter(struct perf_evsel *evsel)
                attr->enable_on_exec = 1;
        }
 
-       return perf_evsel__open_per_thread(evsel, threads, false, false);
+       return perf_evsel__open_per_thread(evsel, evsel_list->threads, false, false);
 }
 
 /*
@@ -203,7 +201,8 @@ static int read_counter_aggr(struct perf_evsel *counter)
        u64 *count = counter->counts->aggr.values;
        int i;
 
-       if (__perf_evsel__read(counter, cpus->nr, threads->nr, scale) < 0)
+       if (__perf_evsel__read(counter, evsel_list->cpus->nr,
+                              evsel_list->threads->nr, scale) < 0)
                return -1;
 
        for (i = 0; i < 3; i++)
@@ -236,7 +235,7 @@ static int read_counter(struct perf_evsel *counter)
        u64 *count;
        int cpu;
 
-       for (cpu = 0; cpu < cpus->nr; cpu++) {
+       for (cpu = 0; cpu < evsel_list->cpus->nr; cpu++) {
                if (__perf_evsel__read_on_cpu(counter, cpu, 0, scale) < 0)
                        return -1;
 
@@ -301,7 +300,7 @@ static int run_perf_stat(int argc __used, const char **argv)
                }
 
                if (target_tid == -1 && target_pid == -1 && !system_wide)
-                       threads->map[0] = child_pid;
+                       evsel_list->threads->map[0] = child_pid;
 
                /*
                 * Wait for the child to be ready to exec.
@@ -353,12 +352,13 @@ static int run_perf_stat(int argc __used, const char **argv)
        if (no_aggr) {
                list_for_each_entry(counter, &evsel_list->entries, node) {
                        read_counter(counter);
-                       perf_evsel__close_fd(counter, cpus->nr, 1);
+                       perf_evsel__close_fd(counter, evsel_list->cpus->nr, 1);
                }
        } else {
                list_for_each_entry(counter, &evsel_list->entries, node) {
                        read_counter_aggr(counter);
-                       perf_evsel__close_fd(counter, cpus->nr, threads->nr);
+                       perf_evsel__close_fd(counter, evsel_list->cpus->nr,
+                                            evsel_list->threads->nr);
                }
        }
 
@@ -386,7 +386,7 @@ static void nsec_printout(int cpu, struct perf_evsel *evsel, double avg)
        if (no_aggr)
                sprintf(cpustr, "CPU%*d%s",
                        csv_output ? 0 : -4,
-                       cpus->map[cpu], csv_sep);
+                       evsel_list->cpus->map[cpu], csv_sep);
 
        fprintf(stderr, fmt, cpustr, msecs, csv_sep, event_name(evsel));
 
@@ -414,7 +414,7 @@ static void abs_printout(int cpu, struct perf_evsel *evsel, double avg)
        if (no_aggr)
                sprintf(cpustr, "CPU%*d%s",
                        csv_output ? 0 : -4,
-                       cpus->map[cpu], csv_sep);
+                       evsel_list->cpus->map[cpu], csv_sep);
        else
                cpu = 0;
 
@@ -500,14 +500,14 @@ static void print_counter(struct perf_evsel *counter)
        u64 ena, run, val;
        int cpu;
 
-       for (cpu = 0; cpu < cpus->nr; cpu++) {
+       for (cpu = 0; cpu < evsel_list->cpus->nr; cpu++) {
                val = counter->counts->cpu[cpu].val;
                ena = counter->counts->cpu[cpu].ena;
                run = counter->counts->cpu[cpu].run;
                if (run == 0 || ena == 0) {
                        fprintf(stderr, "CPU%*d%s%*s%s%-24s",
                                csv_output ? 0 : -4,
-                               cpus->map[cpu], csv_sep,
+                               evsel_list->cpus->map[cpu], csv_sep,
                                csv_output ? 0 : 18,
                                "<not counted>", csv_sep,
                                event_name(counter));
@@ -652,7 +652,7 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used)
 
        setlocale(LC_ALL, "");
 
-       evsel_list = perf_evlist__new();
+       evsel_list = perf_evlist__new(NULL, NULL);
        if (evsel_list == NULL)
                return -ENOMEM;
 
@@ -701,18 +701,18 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used)
        if (target_pid != -1)
                target_tid = target_pid;
 
-       threads = thread_map__new(target_pid, target_tid);
-       if (threads == NULL) {
+       evsel_list->threads = thread_map__new(target_pid, target_tid);
+       if (evsel_list->threads == NULL) {
                pr_err("Problems finding threads of monitor\n");
                usage_with_options(stat_usage, options);
        }
 
        if (system_wide)
-               cpus = cpu_map__new(cpu_list);
+               evsel_list->cpus = cpu_map__new(cpu_list);
        else
-               cpus = cpu_map__dummy_new();
+               evsel_list->cpus = cpu_map__dummy_new();
 
-       if (cpus == NULL) {
+       if (evsel_list->cpus == NULL) {
                perror("failed to parse CPUs map");
                usage_with_options(stat_usage, options);
                return -1;
@@ -720,8 +720,8 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used)
 
        list_for_each_entry(pos, &evsel_list->entries, node) {
                if (perf_evsel__alloc_stat_priv(pos) < 0 ||
-                   perf_evsel__alloc_counts(pos, cpus->nr) < 0 ||
-                   perf_evsel__alloc_fd(pos, cpus->nr, threads->nr) < 0)
+                   perf_evsel__alloc_counts(pos, evsel_list->cpus->nr) < 0 ||
+                   perf_evsel__alloc_fd(pos, evsel_list->cpus->nr, evsel_list->threads->nr) < 0)
                        goto out_free_fd;
        }
 
@@ -750,7 +750,6 @@ out_free_fd:
                perf_evsel__free_stat_priv(pos);
        perf_evlist__delete(evsel_list);
 out:
-       thread_map__delete(threads);
-       threads = NULL;
+       perf_evlist__delete_maps(evsel_list);
        return status;
 }

diff --git a/tools/perf/builtin-test.c b/tools/perf/builtin-test.c
index 845b9bd54ed452e449fcc7a91b388351fa58b5da..1b2106c58f660c83a858636df0862c458d127eb8 100644 (file)
--- a/tools/perf/builtin-test.c
+++ b/tools/perf/builtin-test.c
@@ -509,7 +509,7 @@ static int test__basic_mmap(void)
                goto out_free_cpus;
        }
 
-       evlist = perf_evlist__new();
+       evlist = perf_evlist__new(cpus, threads);
        if (evlist == NULL) {
                pr_debug("perf_evlist__new\n");
                goto out_free_cpus;
@@ -537,7 +537,7 @@ static int test__basic_mmap(void)
                }
        }
 
-       if (perf_evlist__mmap(evlist, cpus, threads, 128, true) < 0) {
+       if (perf_evlist__mmap(evlist, 128, true) < 0) {
                pr_debug("failed to mmap events: %d (%s)\n", errno,
                         strerror(errno));
                goto out_close_fd;
@@ -579,7 +579,7 @@ static int test__basic_mmap(void)
 
        err = 0;
 out_munmap:
-       perf_evlist__munmap(evlist, 1);
+       perf_evlist__munmap(evlist);
 out_close_fd:
        for (i = 0; i < nsyscalls; ++i)
                perf_evsel__close_fd(evsels[i], 1, threads->nr);

diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index 2f4d1f244be1d5ac8f6b2c740d3a5fdfa7cf1d91..599036b06730a9ae887d8efa2600896275463497 100644 (file)
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -73,9 +73,7 @@ static int                    print_entries;
 
 static int                     target_pid                      =     -1;
 static int                     target_tid                      =     -1;
-static struct thread_map       *threads;
 static bool                    inherit                         =  false;
-static struct cpu_map          *cpus;
 static int                     realtime_prio                   =      0;
 static bool                    group                           =  false;
 static unsigned int            page_size;
@@ -567,12 +565,13 @@ static void print_sym_table(struct perf_session *session)
                printf(" (all");
 
        if (cpu_list)
-               printf(", CPU%s: %s)\n", cpus->nr > 1 ? "s" : "", cpu_list);
+               printf(", CPU%s: %s)\n", evsel_list->cpus->nr > 1 ? "s" : "", cpu_list);
        else {
                if (target_tid != -1)
                        printf(")\n");
                else
-                       printf(", %d CPU%s)\n", cpus->nr, cpus->nr > 1 ? "s" : "");
+                       printf(", %d CPU%s)\n", evsel_list->cpus->nr,
+                              evsel_list->cpus->nr > 1 ? "s" : "");
        }
 
        printf("%-*.*s\n", win_width, win_width, graph_dotted_line);
@@ -1124,7 +1123,7 @@ static void perf_session__mmap_read(struct perf_session *self)
 {
        int i;
 
-       for (i = 0; i < cpus->nr; i++)
+       for (i = 0; i < evsel_list->cpus->nr; i++)
                perf_session__mmap_read_cpu(self, i);
 }
 
@@ -1150,7 +1149,8 @@ static void start_counters(struct perf_evlist *evlist)
 
                attr->mmap = 1;
 try_again:
-               if (perf_evsel__open(counter, cpus, threads, group, inherit) < 0) {
+               if (perf_evsel__open(counter, evsel_list->cpus,
+                                    evsel_list->threads, group, inherit) < 0) {
                        int err = errno;
 
                        if (err == EPERM || err == EACCES)
@@ -1181,7 +1181,7 @@ try_again:
                }
        }
 
-       if (perf_evlist__mmap(evlist, cpus, threads, mmap_pages, false) < 0)
+       if (perf_evlist__mmap(evlist, mmap_pages, false) < 0)
                die("failed to mmap with %d (%s)\n", errno, strerror(errno));
 }
 
@@ -1296,7 +1296,7 @@ int cmd_top(int argc, const char **argv, const char *prefix __used)
        struct perf_evsel *pos;
        int status = -ENOMEM;
 
-       evsel_list = perf_evlist__new();
+       evsel_list = perf_evlist__new(NULL, NULL);
        if (evsel_list == NULL)
                return -ENOMEM;
 
@@ -1306,15 +1306,6 @@ int cmd_top(int argc, const char **argv, const char *prefix __used)
        if (argc)
                usage_with_options(top_usage, options);
 
-       if (target_pid != -1)
-               target_tid = target_pid;
-
-       threads = thread_map__new(target_pid, target_tid);
-       if (threads == NULL) {
-               pr_err("Problems finding threads of monitor\n");
-               usage_with_options(top_usage, options);
-       }
-
        /* CPU and PID are mutually exclusive */
        if (target_tid > 0 && cpu_list) {
                printf("WARNING: PID switch overriding CPU\n");
@@ -1322,6 +1313,13 @@ int cmd_top(int argc, const char **argv, const char *prefix __used)
                cpu_list = NULL;
        }
 
+       if (target_pid != -1)
+               target_tid = target_pid;
+
+       if (perf_evlist__create_maps(evsel_list, target_pid,
+                                    target_tid, cpu_list) < 0)
+               usage_with_options(top_usage, options);
+
        if (!evsel_list->nr_entries &&
            perf_evlist__add_default(evsel_list) < 0) {
                pr_err("Not enough memory for event selector list\n");
@@ -1343,16 +1341,9 @@ int cmd_top(int argc, const char **argv, const char *prefix __used)
                exit(EXIT_FAILURE);
        }
 
-       if (target_tid != -1)
-               cpus = cpu_map__dummy_new();
-       else
-               cpus = cpu_map__new(cpu_list);
-
-       if (cpus == NULL)
-               usage_with_options(top_usage, options);
-
        list_for_each_entry(pos, &evsel_list->entries, node) {
-               if (perf_evsel__alloc_fd(pos, cpus->nr, threads->nr) < 0)
+               if (perf_evsel__alloc_fd(pos, evsel_list->cpus->nr,
+                                        evsel_list->threads->nr) < 0)
                        goto out_free_fd;
                /*
                 * Fill in the ones not specifically initialized via -c:
@@ -1363,8 +1354,8 @@ int cmd_top(int argc, const char **argv, const char *prefix __used)
                pos->attr.sample_period = default_interval;
        }
 
-       if (perf_evlist__alloc_pollfd(evsel_list, cpus->nr, threads->nr) < 0 ||
-           perf_evlist__alloc_mmap(evsel_list, cpus->nr) < 0)
+       if (perf_evlist__alloc_pollfd(evsel_list) < 0 ||
+           perf_evlist__alloc_mmap(evsel_list) < 0)
                goto out_free_fd;
 
        sym_evsel = list_entry(evsel_list->entries.next, struct perf_evsel, node);

diff --git a/tools/perf/python/twatch.py b/tools/perf/python/twatch.py
index 5e9f3b7b7ee8f9a77e5e6590e887bfd54414087e..df638c438a9f90088e41c0e730095078f0a4670a 100755 (executable)
--- a/tools/perf/python/twatch.py
+++ b/tools/perf/python/twatch.py
@@ -23,9 +23,9 @@ def main():
                           sample_id_all = 1,
                           sample_type = perf.SAMPLE_PERIOD | perf.SAMPLE_TID | perf.SAMPLE_CPU | perf.SAMPLE_TID)
        evsel.open(cpus = cpus, threads = threads);
-       evlist = perf.evlist()
+       evlist = perf.evlist(cpus, threads)
        evlist.add(evsel)
-       evlist.mmap(cpus = cpus, threads = threads)
+       evlist.mmap()
        while True:
                evlist.poll(timeout = -1)
                for cpu in cpus:

diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index dcd59328bb49048d2b9a34d8eab488d1f8a9ac0f..95b21fece2cefe1ac7c513a7351f37dbb0abd4be 100644 (file)
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -21,21 +21,24 @@
 #define FD(e, x, y) (*(int *)xyarray__entry(e->fd, x, y))
 #define SID(e, x, y) xyarray__entry(e->id, x, y)
 
-void perf_evlist__init(struct perf_evlist *evlist)
+void perf_evlist__init(struct perf_evlist *evlist, struct cpu_map *cpus,
+                      struct thread_map *threads)
 {
        int i;
 
        for (i = 0; i < PERF_EVLIST__HLIST_SIZE; ++i)
                INIT_HLIST_HEAD(&evlist->heads[i]);
        INIT_LIST_HEAD(&evlist->entries);
+       perf_evlist__set_maps(evlist, cpus, threads);
 }
 
-struct perf_evlist *perf_evlist__new(void)
+struct perf_evlist *perf_evlist__new(struct cpu_map *cpus,
+                                    struct thread_map *threads)
 {
        struct perf_evlist *evlist = zalloc(sizeof(*evlist));
 
        if (evlist != NULL)
-               perf_evlist__init(evlist);
+               perf_evlist__init(evlist, cpus, threads);
 
        return evlist;
 }
@@ -88,9 +91,9 @@ int perf_evlist__add_default(struct perf_evlist *evlist)
        return 0;
 }
 
-int perf_evlist__alloc_pollfd(struct perf_evlist *evlist, int ncpus, int nthreads)
+int perf_evlist__alloc_pollfd(struct perf_evlist *evlist)
 {
-       int nfds = ncpus * nthreads * evlist->nr_entries;
+       int nfds = evlist->cpus->nr * evlist->threads->nr * evlist->nr_entries;
        evlist->pollfd = malloc(sizeof(struct pollfd) * nfds);
        return evlist->pollfd != NULL ? 0 : -ENOMEM;
 }
@@ -213,11 +216,11 @@ union perf_event *perf_evlist__read_on_cpu(struct perf_evlist *evlist, int cpu)
        return event;
 }
 
-void perf_evlist__munmap(struct perf_evlist *evlist, int ncpus)
+void perf_evlist__munmap(struct perf_evlist *evlist)
 {
        int cpu;
 
-       for (cpu = 0; cpu < ncpus; cpu++) {
+       for (cpu = 0; cpu < evlist->cpus->nr; cpu++) {
                if (evlist->mmap[cpu].base != NULL) {
                        munmap(evlist->mmap[cpu].base, evlist->mmap_len);
                        evlist->mmap[cpu].base = NULL;
@@ -225,9 +228,9 @@ void perf_evlist__munmap(struct perf_evlist *evlist, int ncpus)
        }
 }
 
-int perf_evlist__alloc_mmap(struct perf_evlist *evlist, int ncpus)
+int perf_evlist__alloc_mmap(struct perf_evlist *evlist)
 {
-       evlist->mmap = zalloc(ncpus * sizeof(struct perf_mmap));
+       evlist->mmap = zalloc(evlist->cpus->nr * sizeof(struct perf_mmap));
        return evlist->mmap != NULL ? 0 : -ENOMEM;
 }
 
@@ -248,8 +251,6 @@ static int __perf_evlist__mmap(struct perf_evlist *evlist, int cpu, int prot,
 /** perf_evlist__mmap - Create per cpu maps to receive events
  *
  * @evlist - list of events
- * @cpus - cpu map being monitored
- * @threads - threads map being monitored
  * @pages - map length in pages
  * @overwrite - overwrite older events?
  *
@@ -259,21 +260,22 @@ static int __perf_evlist__mmap(struct perf_evlist *evlist, int cpu, int prot,
  *     unsigned int head = perf_mmap__read_head(m);
  *
  *     perf_mmap__write_tail(m, head)
+ *
+ * Using perf_evlist__read_on_cpu does this automatically.
  */
-int perf_evlist__mmap(struct perf_evlist *evlist, struct cpu_map *cpus,
-                     struct thread_map *threads, int pages, bool overwrite)
+int perf_evlist__mmap(struct perf_evlist *evlist, int pages, bool overwrite)
 {
        unsigned int page_size = sysconf(_SC_PAGE_SIZE);
        int mask = pages * page_size - 1, cpu;
        struct perf_evsel *first_evsel, *evsel;
+       const struct cpu_map *cpus = evlist->cpus;
+       const struct thread_map *threads = evlist->threads;
        int thread, prot = PROT_READ | (overwrite ? 0 : PROT_WRITE);
 
-       if (evlist->mmap == NULL &&
-           perf_evlist__alloc_mmap(evlist, cpus->nr) < 0)
+       if (evlist->mmap == NULL && perf_evlist__alloc_mmap(evlist) < 0)
                return -ENOMEM;
 
-       if (evlist->pollfd == NULL &&
-           perf_evlist__alloc_pollfd(evlist, cpus->nr, threads->nr) < 0)
+       if (evlist->pollfd == NULL && perf_evlist__alloc_pollfd(evlist) < 0)
                return -ENOMEM;
 
        evlist->overwrite = overwrite;
@@ -315,3 +317,34 @@ out_unmap:
        }
        return -1;
 }
+
+int perf_evlist__create_maps(struct perf_evlist *evlist, pid_t target_pid,
+                            pid_t target_tid, const char *cpu_list)
+{
+       evlist->threads = thread_map__new(target_pid, target_tid);
+
+       if (evlist->threads == NULL)
+               return -1;
+
+       if (target_tid != -1)
+               evlist->cpus = cpu_map__dummy_new();
+       else
+               evlist->cpus = cpu_map__new(cpu_list);
+
+       if (evlist->cpus == NULL)
+               goto out_delete_threads;
+
+       return 0;
+
+out_delete_threads:
+       thread_map__delete(evlist->threads);
+       return -1;
+}
+
+void perf_evlist__delete_maps(struct perf_evlist *evlist)
+{
+       cpu_map__delete(evlist->cpus);
+       thread_map__delete(evlist->threads);
+       evlist->cpus    = NULL;
+       evlist->threads = NULL;
+}

diff --git a/tools/perf/util/evlist.h b/tools/perf/util/evlist.h
index 85aca6eba16bbba84be1766de0c74c07b519fedb..c9884056097c1bf617c15772a1e3f76266330727 100644 (file)
--- a/tools/perf/util/evlist.h
+++ b/tools/perf/util/evlist.h
@@ -22,28 +22,43 @@ struct perf_evlist {
        union perf_event event_copy;
        struct perf_mmap *mmap;
        struct pollfd    *pollfd;
+       struct thread_map *threads;
+       struct cpu_map    *cpus;
 };
 
 struct perf_evsel;
 
-struct perf_evlist *perf_evlist__new(void);
-void perf_evlist__init(struct perf_evlist *evlist);
+struct perf_evlist *perf_evlist__new(struct cpu_map *cpus,
+                                    struct thread_map *threads);
+void perf_evlist__init(struct perf_evlist *evlist, struct cpu_map *cpus,
+                      struct thread_map *threads);
 void perf_evlist__exit(struct perf_evlist *evlist);
 void perf_evlist__delete(struct perf_evlist *evlist);
 
 void perf_evlist__add(struct perf_evlist *evlist, struct perf_evsel *entry);
 int perf_evlist__add_default(struct perf_evlist *evlist);
 
-int perf_evlist__alloc_pollfd(struct perf_evlist *evlist, int ncpus, int nthreads);
+int perf_evlist__alloc_pollfd(struct perf_evlist *evlist);
 void perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd);
 
 struct perf_evsel *perf_evlist__id2evsel(struct perf_evlist *evlist, u64 id);
 
 union perf_event *perf_evlist__read_on_cpu(struct perf_evlist *self, int cpu);
 
-int perf_evlist__alloc_mmap(struct perf_evlist *evlist, int ncpus);
-int perf_evlist__mmap(struct perf_evlist *evlist, struct cpu_map *cpus,
-                     struct thread_map *threads, int pages, bool overwrite);
-void perf_evlist__munmap(struct perf_evlist *evlist, int ncpus);
+int perf_evlist__alloc_mmap(struct perf_evlist *evlist);
+int perf_evlist__mmap(struct perf_evlist *evlist, int pages, bool overwrite);
+void perf_evlist__munmap(struct perf_evlist *evlist);
+
+static inline void perf_evlist__set_maps(struct perf_evlist *evlist,
+                                        struct cpu_map *cpus,
+                                        struct thread_map *threads)
+{
+       evlist->cpus    = cpus;
+       evlist->threads = threads;
+}
+
+int perf_evlist__create_maps(struct perf_evlist *evlist, pid_t target_pid,
+                            pid_t target_tid, const char *cpu_list);
+void perf_evlist__delete_maps(struct perf_evlist *evlist);
 
 #endif /* __PERF_EVLIST_H */

diff --git a/tools/perf/util/python.c b/tools/perf/util/python.c
index 88d47895a79c84ba761d0430ebd2eb8557c12a19..d2d52175362e9b304b406f51a8474374fd8ec510 100644 (file)
--- a/tools/perf/util/python.c
+++ b/tools/perf/util/python.c
@@ -553,7 +553,16 @@ struct pyrf_evlist {
 static int pyrf_evlist__init(struct pyrf_evlist *pevlist,
                             PyObject *args, PyObject *kwargs)
 {
-       perf_evlist__init(&pevlist->evlist);
+       PyObject *pcpus = NULL, *pthreads = NULL;
+       struct cpu_map *cpus;
+       struct thread_map *threads;
+
+       if (!PyArg_ParseTuple(args, "OO", &pcpus, &pthreads))
+               return -1;
+
+       threads = ((struct pyrf_thread_map *)pthreads)->threads;
+       cpus = ((struct pyrf_cpu_map *)pcpus)->cpus;
+       perf_evlist__init(&pevlist->evlist, cpus, threads);
        return 0;
 }
 
@@ -567,21 +576,15 @@ static PyObject *pyrf_evlist__mmap(struct pyrf_evlist *pevlist,
                                   PyObject *args, PyObject *kwargs)
 {
        struct perf_evlist *evlist = &pevlist->evlist;
-       PyObject *pcpus = NULL, *pthreads = NULL;
-       struct cpu_map *cpus = NULL;
-       struct thread_map *threads = NULL;
-       static char *kwlist[] = {"cpus", "threads", "pages", "overwrite",
+       static char *kwlist[] = {"pages", "overwrite",
                                  NULL, NULL};
        int pages = 128, overwrite = false;
 
-       if (!PyArg_ParseTupleAndKeywords(args, kwargs, "OO|ii", kwlist,
-                                        &pcpus, &pthreads, &pages, &overwrite))
+       if (!PyArg_ParseTupleAndKeywords(args, kwargs, "|ii", kwlist,
+                                        &pages, &overwrite))
                return NULL;
 
-       threads = ((struct pyrf_thread_map *)pthreads)->threads;
-       cpus = ((struct pyrf_cpu_map *)pcpus)->cpus;
-
-       if (perf_evlist__mmap(evlist, cpus, threads, pages, overwrite) < 0) {
+       if (perf_evlist__mmap(evlist, pages, overwrite) < 0) {
                PyErr_SetFromErrno(PyExc_OSError);
                return NULL;
        }