Merge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core

Pull perf/core improvements and fixes from Arnaldo:

User visible:

  * Add several futex 'perf bench' microbenchmarks (Davidlohr Bueso)

  * Speed up thread map generation (Don Zickus)

  * Fix synthesizing mmaps for threads (Don Zickus)

  * Fix invalid output on event group stdio report  (Namhyung Kim)

  * Introduce 'perf kvm --list-cmds' command line option for use by
    scripts (Ramkumar Ramachandra)

Documentation:

  * Clarify load-latency information in the 'perf mem' docs (Andi Kleen)

  * Clarify x86 register naming in 'perf probe' docs (Andi Kleen)

Refactorings:

  * hists browser refactorings to reuse code accross UIs (Namhyung Kim)

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>

9 files changed, 204 insertions, 94 deletions

diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c
index b0f3ca850e9e..3e580be0f6fb 100644
--- a/tools/perf/util/event.c
+++ b/tools/perf/util/event.c
@@ -94,14 +94,10 @@ static pid_t perf_event__get_comm_tgid(pid_t pid, char *comm, size_t len)
 
 static pid_t perf_event__synthesize_comm(struct perf_tool *tool,
                                          union perf_event *event, pid_t pid,
-                                         int full,
                                          perf_event__handler_t process,
                                          struct machine *machine)
 {
-        char filename[PATH_MAX];
         size_t size;
-        DIR *tasks;
-        struct dirent dirent, *next;
         pid_t tgid;
 
         memset(&event->comm, 0, sizeof(event->comm));
@@ -124,55 +120,35 @@ static pid_t perf_event__synthesize_comm(struct perf_tool *tool,
         event->comm.header.size = (sizeof(event->comm) -
                                 (sizeof(event->comm.comm) - size) +
                                 machine->id_hdr_size);
-        if (!full) {
-                event->comm.tid = pid;
-
-                if (process(tool, event, &synth_sample, machine) != 0)
-                        return -1;
+        event->comm.tid = pid;
 
-                goto out;
-        }
-
-        if (machine__is_default_guest(machine))
-                return 0;
-
-        snprintf(filename, sizeof(filename), "%s/proc/%d/task",
-                 machine->root_dir, pid);
-
-        tasks = opendir(filename);
-        if (tasks == NULL) {
-                pr_debug("couldn't open %s\n", filename);
-                return 0;
-        }
+        if (process(tool, event, &synth_sample, machine) != 0)
+                return -1;
 
-        while (!readdir_r(tasks, &dirent, &next) && next) {
-                char *end;
-                pid = strtol(dirent.d_name, &end, 10);
-                if (*end)
-                        continue;
+out:
+        return tgid;
+}
 
-                /* already have tgid; jut want to update the comm */
-                (void) perf_event__get_comm_tgid(pid, event->comm.comm,
-                                         sizeof(event->comm.comm));
+static int perf_event__synthesize_fork(struct perf_tool *tool,
+                                       union perf_event *event, pid_t pid,
+                                       pid_t tgid, perf_event__handler_t process,
+                                       struct machine *machine)
+{
+        memset(&event->fork, 0, sizeof(event->fork) + machine->id_hdr_size);
 
-                size = strlen(event->comm.comm) + 1;
-                size = PERF_ALIGN(size, sizeof(u64));
-                memset(event->comm.comm + size, 0, machine->id_hdr_size);
-                event->comm.header.size = (sizeof(event->comm) -
-                                          (sizeof(event->comm.comm) - size) +
-                                          machine->id_hdr_size);
+        /* this is really a clone event but we use fork to synthesize it */
+        event->fork.ppid = tgid;
+        event->fork.ptid = tgid;
+        event->fork.pid  = tgid;
+        event->fork.tid  = pid;
+        event->fork.header.type = PERF_RECORD_FORK;
 
-                event->comm.tid = pid;
+        event->fork.header.size = (sizeof(event->fork) + machine->id_hdr_size);
 
-                if (process(tool, event, &synth_sample, machine) != 0) {
-                        tgid = -1;
-                        break;
-                }
-        }
+        if (process(tool, event, &synth_sample, machine) != 0)
+                return -1;
 
-        closedir(tasks);
-out:
-        return tgid;
+        return 0;
 }
 
 int perf_event__synthesize_mmap_events(struct perf_tool *tool,
@@ -324,17 +300,71 @@ int perf_event__synthesize_modules(struct perf_tool *tool,
 
 static int __event__synthesize_thread(union perf_event *comm_event,
                                       union perf_event *mmap_event,
+                                      union perf_event *fork_event,
                                       pid_t pid, int full,
                                           perf_event__handler_t process,
                                       struct perf_tool *tool,
                                       struct machine *machine, bool mmap_data)
 {
-        pid_t tgid = perf_event__synthesize_comm(tool, comm_event, pid, full,
+        char filename[PATH_MAX];
+        DIR *tasks;
+        struct dirent dirent, *next;
+        pid_t tgid;
+
+        /* special case: only send one comm event using passed in pid */
+        if (!full) {
+                tgid = perf_event__synthesize_comm(tool, comm_event, pid,
+                                                   process, machine);
+
+                if (tgid == -1)
+                        return -1;
+
+                return perf_event__synthesize_mmap_events(tool, mmap_event, pid, tgid,
+                                                          process, machine, mmap_data);
+        }
+
+        if (machine__is_default_guest(machine))
+                return 0;
+
+        snprintf(filename, sizeof(filename), "%s/proc/%d/task",
+                 machine->root_dir, pid);
+
+        tasks = opendir(filename);
+        if (tasks == NULL) {
+                pr_debug("couldn't open %s\n", filename);
+                return 0;
+        }
+
+        while (!readdir_r(tasks, &dirent, &next) && next) {
+                char *end;
+                int rc = 0;
+                pid_t _pid;
+
+                _pid = strtol(dirent.d_name, &end, 10);
+                if (*end)
+                        continue;
+
+                tgid = perf_event__synthesize_comm(tool, comm_event, _pid,
+                                                   process, machine);
+                if (tgid == -1)
+                        return -1;
+
+                if (_pid == pid) {
+                        /* process the parent's maps too */
+                        rc = perf_event__synthesize_mmap_events(tool, mmap_event, pid, tgid,
+                                                process, machine, mmap_data);
+                } else {
+                        /* only fork the tid's map, to save time */
+                        rc = perf_event__synthesize_fork(tool, fork_event, _pid, tgid,
                                                  process, machine);
-        if (tgid == -1)
-                return -1;
-        return perf_event__synthesize_mmap_events(tool, mmap_event, pid, tgid,
-                                                  process, machine, mmap_data);
+                }
+
+                if (rc)
+                        return rc;
+        }
+
+        closedir(tasks);
+        return 0;
 }
 
 int perf_event__synthesize_thread_map(struct perf_tool *tool,
@@ -343,7 +373,7 @@ int perf_event__synthesize_thread_map(struct perf_tool *tool,
                                       struct machine *machine,
                                       bool mmap_data)
 {
-        union perf_event *comm_event, *mmap_event;
+        union perf_event *comm_event, *mmap_event, *fork_event;
         int err = -1, thread, j;
 
         comm_event = malloc(sizeof(comm_event->comm) + machine->id_hdr_size);
@@ -354,9 +384,14 @@ int perf_event__synthesize_thread_map(struct perf_tool *tool,
         if (mmap_event == NULL)
                 goto out_free_comm;
 
+        fork_event = malloc(sizeof(fork_event->fork) + machine->id_hdr_size);
+        if (fork_event == NULL)
+                goto out_free_mmap;
+
         err = 0;
         for (thread = 0; thread < threads->nr; ++thread) {
                 if (__event__synthesize_thread(comm_event, mmap_event,
+                                               fork_event,
                                                threads->map[thread], 0,
                                                process, tool, machine,
                                                mmap_data)) {
@@ -382,6 +417,7 @@ int perf_event__synthesize_thread_map(struct perf_tool *tool,
                         /* if not, generate events for it */
                         if (need_leader &&
                             __event__synthesize_thread(comm_event, mmap_event,
+                                                       fork_event,
                                                        comm_event->comm.pid, 0,
                                                        process, tool, machine,
                                                        mmap_data)) {
@@ -390,6 +426,8 @@ int perf_event__synthesize_thread_map(struct perf_tool *tool,
                         }
                 }
         }
+        free(fork_event);
+out_free_mmap:
         free(mmap_event);
 out_free_comm:
         free(comm_event);
@@ -404,7 +442,7 @@ int perf_event__synthesize_threads(struct perf_tool *tool,
         DIR *proc;
         char proc_path[PATH_MAX];
         struct dirent dirent, *next;
-        union perf_event *comm_event, *mmap_event;
+        union perf_event *comm_event, *mmap_event, *fork_event;
         int err = -1;
 
         comm_event = malloc(sizeof(comm_event->comm) + machine->id_hdr_size);
@@ -415,6 +453,10 @@ int perf_event__synthesize_threads(struct perf_tool *tool,
         if (mmap_event == NULL)
                 goto out_free_comm;
 
+        fork_event = malloc(sizeof(fork_event->fork) + machine->id_hdr_size);
+        if (fork_event == NULL)
+                goto out_free_mmap;
+
         if (machine__is_default_guest(machine))
                 return 0;
 
@@ -422,7 +464,7 @@ int perf_event__synthesize_threads(struct perf_tool *tool,
         proc = opendir(proc_path);
 
         if (proc == NULL)
-                goto out_free_mmap;
+                goto out_free_fork;
 
         while (!readdir_r(proc, &dirent, &next) && next) {
                 char *end;
@@ -434,12 +476,14 @@ int perf_event__synthesize_threads(struct perf_tool *tool,
                  * We may race with exiting thread, so don't stop just because
                  * one thread couldn't be synthesized.
                  */
-                __event__synthesize_thread(comm_event, mmap_event, pid, 1,
-                                           process, tool, machine, mmap_data);
+                __event__synthesize_thread(comm_event, mmap_event, fork_event, pid,
+                                           1, process, tool, machine, mmap_data);
         }
 
         err = 0;
         closedir(proc);
+out_free_fork:
+        free(fork_event);
 out_free_mmap:
         free(mmap_event);
 out_free_comm:
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index a59743fa3ef7..0c76bf972736 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -132,8 +132,10 @@ struct perf_hpp {
 };
 
 struct perf_hpp_fmt {
-        int (*header)(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp);
-        int (*width)(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp);
+        int (*header)(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp,
+                      struct perf_evsel *evsel);
+        int (*width)(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp,
+                     struct perf_evsel *evsel);
         int (*color)(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp,
                      struct hist_entry *he);
         int (*entry)(struct perf_hpp_fmt *fmt, struct perf_hpp *hpp,
@@ -166,6 +168,20 @@ void perf_hpp__init(void);
 void perf_hpp__column_register(struct perf_hpp_fmt *format);
 void perf_hpp__column_enable(unsigned col);
 
+typedef u64 (*hpp_field_fn)(struct hist_entry *he);
+typedef int (*hpp_callback_fn)(struct perf_hpp *hpp, bool front);
+typedef int (*hpp_snprint_fn)(struct perf_hpp *hpp, const char *fmt, ...);
+
+int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he,
+               hpp_field_fn get_field, hpp_callback_fn callback,
+               const char *fmt, hpp_snprint_fn print_fn, bool fmt_percent);
+
+static inline void advance_hpp(struct perf_hpp *hpp, int inc)
+{
+        hpp->buf  += inc;
+        hpp->size -= inc;
+}
+
 static inline size_t perf_hpp__use_color(void)
 {
         return !symbol_conf.field_sep;
diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c
index 813e94e7cf29..5cecd98c1bc0 100644
--- a/tools/perf/util/machine.c
+++ b/tools/perf/util/machine.c
@@ -327,9 +327,10 @@ struct thread *machine__findnew_thread(struct machine *machine, pid_t pid,
         return __machine__findnew_thread(machine, pid, tid, true);
 }
 
-struct thread *machine__find_thread(struct machine *machine, pid_t tid)
+struct thread *machine__find_thread(struct machine *machine, pid_t pid,
+                                    pid_t tid)
 {
-        return __machine__findnew_thread(machine, 0, tid, false);
+        return __machine__findnew_thread(machine, pid, tid, false);
 }
 
 int machine__process_comm_event(struct machine *machine, union perf_event *event,
@@ -1114,7 +1115,9 @@ static void machine__remove_thread(struct machine *machine, struct thread *th)
 int machine__process_fork_event(struct machine *machine, union perf_event *event,
                                 struct perf_sample *sample)
 {
-        struct thread *thread = machine__find_thread(machine, event->fork.tid);
+        struct thread *thread = machine__find_thread(machine,
+                                                     event->fork.pid,
+                                                     event->fork.tid);
         struct thread *parent = machine__findnew_thread(machine,
                                                         event->fork.ppid,
                                                         event->fork.ptid);
@@ -1140,7 +1143,9 @@ int machine__process_fork_event(struct machine *machine, union perf_event *event
 int machine__process_exit_event(struct machine *machine, union perf_event *event,
                                 struct perf_sample *sample __maybe_unused)
 {
-        struct thread *thread = machine__find_thread(machine, event->fork.tid);
+        struct thread *thread = machine__find_thread(machine,
+                                                     event->fork.pid,
+                                                     event->fork.tid);
 
         if (dump_trace)
                 perf_event__fprintf_task(event, stdout);
@@ -1184,39 +1189,22 @@ static bool symbol__match_regex(struct symbol *sym, regex_t *regex)
         return 0;
 }
 
-static const u8 cpumodes[] = {
-        PERF_RECORD_MISC_USER,
-        PERF_RECORD_MISC_KERNEL,
-        PERF_RECORD_MISC_GUEST_USER,
-        PERF_RECORD_MISC_GUEST_KERNEL
-};
-#define NCPUMODES (sizeof(cpumodes)/sizeof(u8))
-
 static void ip__resolve_ams(struct machine *machine, struct thread *thread,
                             struct addr_map_symbol *ams,
                             u64 ip)
 {
         struct addr_location al;
-        size_t i;
-        u8 m;
 
         memset(&al, 0, sizeof(al));
+        /*
+         * We cannot use the header.misc hint to determine whether a
+         * branch stack address is user, kernel, guest, hypervisor.
+         * Branches may straddle the kernel/user/hypervisor boundaries.
+         * Thus, we have to try consecutively until we find a match
+         * or else, the symbol is unknown
+         */
+        thread__find_cpumode_addr_location(thread, machine, MAP__FUNCTION, ip, &al);
 
-        for (i = 0; i < NCPUMODES; i++) {
-                m = cpumodes[i];
-                /*
-                 * We cannot use the header.misc hint to determine whether a
-                 * branch stack address is user, kernel, guest, hypervisor.
-                 * Branches may straddle the kernel/user/hypervisor boundaries.
-                 * Thus, we have to try consecutively until we find a match
-                 * or else, the symbol is unknown
-                 */
-                thread__find_addr_location(thread, machine, m, MAP__FUNCTION,
-                                ip, &al);
-                if (al.map)
-                        goto found;
-        }
-found:
         ams->addr = ip;
         ams->al_addr = al.addr;
         ams->sym = al.sym;
diff --git a/tools/perf/util/machine.h b/tools/perf/util/machine.h
index 2e6c248c870f..c8c74a119398 100644
--- a/tools/perf/util/machine.h
+++ b/tools/perf/util/machine.h
@@ -41,7 +41,8 @@ struct map *machine__kernel_map(struct machine *machine, enum map_type type)
         return machine->vmlinux_maps[type];
 }
 
-struct thread *machine__find_thread(struct machine *machine, pid_t tid);
+struct thread *machine__find_thread(struct machine *machine, pid_t pid,
+                                    pid_t tid);
 
 int machine__process_comm_event(struct machine *machine, union perf_event *event,
                                 struct perf_sample *sample);
diff --git a/tools/perf/util/parse-options.c b/tools/perf/util/parse-options.c
index d22e3f8017dc..bf48092983c6 100644
--- a/tools/perf/util/parse-options.c
+++ b/tools/perf/util/parse-options.c
@@ -407,7 +407,9 @@ int parse_options_step(struct parse_opt_ctx_t *ctx,
                 if (internal_help && !strcmp(arg + 2, "help"))
                         return usage_with_options_internal(usagestr, options, 0);
                 if (!strcmp(arg + 2, "list-opts"))
-                        return PARSE_OPT_LIST;
+                        return PARSE_OPT_LIST_OPTS;
+                if (!strcmp(arg + 2, "list-cmds"))
+                        return PARSE_OPT_LIST_SUBCMDS;
                 switch (parse_long_opt(ctx, arg + 2, options)) {
                 case -1:
                         return parse_options_usage(usagestr, options, arg + 2, 0);
@@ -433,25 +435,45 @@ int parse_options_end(struct parse_opt_ctx_t *ctx)
         return ctx->cpidx + ctx->argc;
 }
 
-int parse_options(int argc, const char **argv, const struct option *options,
-                  const char * const usagestr[], int flags)
+int parse_options_subcommand(int argc, const char **argv, const struct option *options,
+                        const char *const subcommands[], const char *usagestr[], int flags)
 {
         struct parse_opt_ctx_t ctx;
 
         perf_header__set_cmdline(argc, argv);
 
+        /* build usage string if it's not provided */
+        if (subcommands && !usagestr[0]) {
+                struct strbuf buf = STRBUF_INIT;
+
+                strbuf_addf(&buf, "perf %s [<options>] {", argv[0]);
+                for (int i = 0; subcommands[i]; i++) {
+                        if (i)
+                                strbuf_addstr(&buf, "|");
+                        strbuf_addstr(&buf, subcommands[i]);
+                }
+                strbuf_addstr(&buf, "}");
+
+                usagestr[0] = strdup(buf.buf);
+                strbuf_release(&buf);
+        }
+
         parse_options_start(&ctx, argc, argv, flags);
         switch (parse_options_step(&ctx, options, usagestr)) {
         case PARSE_OPT_HELP:
                 exit(129);
         case PARSE_OPT_DONE:
                 break;
-        case PARSE_OPT_LIST:
+        case PARSE_OPT_LIST_OPTS:
                 while (options->type != OPTION_END) {
                         printf("--%s ", options->long_name);
                         options++;
                 }
                 exit(130);
+        case PARSE_OPT_LIST_SUBCMDS:
+                for (int i = 0; subcommands[i]; i++)
+                        printf("%s ", subcommands[i]);
+                exit(130);
         default: /* PARSE_OPT_UNKNOWN */
                 if (ctx.argv[0][1] == '-') {
                         error("unknown option `%s'", ctx.argv[0] + 2);
@@ -464,6 +486,13 @@ int parse_options(int argc, const char **argv, const struct option *options,
         return parse_options_end(&ctx);
 }
 
+int parse_options(int argc, const char **argv, const struct option *options,
+                  const char * const usagestr[], int flags)
+{
+        return parse_options_subcommand(argc, argv, options, NULL,
+                                        (const char **) usagestr, flags);
+}
+
 #define USAGE_OPTS_WIDTH 24
 #define USAGE_GAP         2
 
diff --git a/tools/perf/util/parse-options.h b/tools/perf/util/parse-options.h
index cbf0149cf221..d8dac8ac5f37 100644
--- a/tools/perf/util/parse-options.h
+++ b/tools/perf/util/parse-options.h
@@ -140,6 +140,11 @@ extern int parse_options(int argc, const char **argv,
                          const struct option *options,
                          const char * const usagestr[], int flags);
 
+extern int parse_options_subcommand(int argc, const char **argv,
+                                const struct option *options,
+                                const char *const subcommands[],
+                                const char *usagestr[], int flags);
+
 extern NORETURN void usage_with_options(const char * const *usagestr,
                                         const struct option *options);
 
@@ -148,7 +153,8 @@ extern NORETURN void usage_with_options(const char * const *usagestr,
 enum {
         PARSE_OPT_HELP = -1,
         PARSE_OPT_DONE,
-        PARSE_OPT_LIST,
+        PARSE_OPT_LIST_OPTS,
+        PARSE_OPT_LIST_SUBCMDS,
         PARSE_OPT_UNKNOWN,
 };
 
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index 1d555d652f58..55960f22233c 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -794,7 +794,7 @@ static void dump_sample(struct perf_evsel *evsel, union perf_event *event,
         if (!dump_trace)
                 return;
 
-        printf("(IP, %d): %d/%d: %#" PRIx64 " period: %" PRIu64 " addr: %#" PRIx64 "\n",
+        printf("(IP, 0x%x): %d/%d: %#" PRIx64 " period: %" PRIu64 " addr: %#" PRIx64 "\n",
                event->header.misc, sample->pid, sample->tid, sample->ip,
                sample->period, sample->addr);
 
diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c
index 0358882c8910..3ce0498bdae6 100644
--- a/tools/perf/util/thread.c
+++ b/tools/perf/util/thread.c
@@ -142,3 +142,24 @@ int thread__fork(struct thread *thread, struct thread *parent, u64 timestamp)
 
         return 0;
 }
+
+void thread__find_cpumode_addr_location(struct thread *thread,
+                                        struct machine *machine,
+                                        enum map_type type, u64 addr,
+                                        struct addr_location *al)
+{
+        size_t i;
+        const u8 const cpumodes[] = {
+                PERF_RECORD_MISC_USER,
+                PERF_RECORD_MISC_KERNEL,
+                PERF_RECORD_MISC_GUEST_USER,
+                PERF_RECORD_MISC_GUEST_KERNEL
+        };
+
+        for (i = 0; i < ARRAY_SIZE(cpumodes); i++) {
+                thread__find_addr_location(thread, machine, cpumodes[i], type,
+                                           addr, al);
+                if (al->map)
+                        break;
+        }
+}
diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h
index 5b856bf942e1..9a070743270c 100644
--- a/tools/perf/util/thread.h
+++ b/tools/perf/util/thread.h
@@ -58,6 +58,11 @@ void thread__find_addr_location(struct thread *thread, struct machine *machine,
                                 u8 cpumode, enum map_type type, u64 addr,
                                 struct addr_location *al);
 
+void thread__find_cpumode_addr_location(struct thread *thread,
+                                        struct machine *machine,
+                                        enum map_type type, u64 addr,
+                                        struct addr_location *al);
+
 static inline void *thread__priv(struct thread *thread)
 {
         return thread->priv;