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 Carvalho de Melo:

User visible changes:

  - Allow selecting the type of callchains per event, including disabling
    callchains in all but one entry in an event list, to save space, and also
    to ask for the callchains collected in one event to be used in other
    events. (Kan Liang)

  - Beautify more syscall arguments in 'perf trace': (Arnaldo Carvalho de Melo)
    - A bunch more translate file/pathnames from pointers to strings.
    - Convert numbers to strings for the 'keyctl' syscall 'option' arg.
    - Add missing 'clockid' entries.

  - Fix 'perf probe -L sys_*' as it was not showing all the source code for
    syscall functions in the kernel. (Masami Hiramatsu)

  - Make ESC unzoom as well in the hists browser, i.e. in 'report' and 'top',
    as we're considering repurposing the right and left arrow keys to use in
    horizontal scrolling, i.e. leave just ESC to be used for what <- works
    now, and ENTER for what -> does (they are already aliases for ages).
    (Arnaldo Carvalho de Melo)

Infrastructure fixes:

  - Check for SRCLINE_UNKNOWN case in "srcfile" processing (Andi Kleen)

  - Wrap the slsmg_{printf,write_nstring} slang functions behind ui_browser, so
    that we can make the ui_browser based browsers (annotate, menus, hists, etc) UI
    library agnostic and usable with multiple backends (slang now, GTK+ and others
    in the future, maybe) (Arnaldo Carvalho de Melo)

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

27 files changed, 307 insertions, 81 deletions

diff --git a/tools/perf/Documentation/perf-record.txt b/tools/perf/Documentation/perf-record.txt
index afbe45ef7e3e..347a27322ed8 100644
--- a/tools/perf/Documentation/perf-record.txt
+++ b/tools/perf/Documentation/perf-record.txt
@@ -53,6 +53,10 @@ OPTIONS
           - 'time': Disable/enable time stamping. Acceptable values are 1 for
                     enabling time stamping. 0 for disabling time stamping.
                     The default is 1.
+          - 'call-graph': Disable/enable callgraph. Acceptable str are "fp" for
+                         FP mode, "dwarf" for DWARF mode, "lbr" for LBR mode and
+                         "no" for disable callgraph.
+          - 'stack-size': user stack size for dwarf mode
           Note: If user explicitly sets options which conflict with the params,
           the value set by the params will be overridden.
 
diff --git a/tools/perf/Documentation/perf-report.txt b/tools/perf/Documentation/perf-report.txt
index 7b07d19e2d54..a18ba757a0ed 100644
--- a/tools/perf/Documentation/perf-report.txt
+++ b/tools/perf/Documentation/perf-report.txt
@@ -359,6 +359,17 @@ OPTIONS
 --full-source-path::
         Show the full path for source files for srcline output.
 
+--show-ref-call-graph::
+        When multiple events are sampled, it may not be needed to collect
+        callgraphs for all of them. The sample sites are usually nearby,
+        and it's enough to collect the callgraphs on a reference event.
+        So user can use "call-graph=no" event modifier to disable callgraph
+        for other events to reduce the overhead.
+        However, perf report cannot show callgraphs for the event which
+        disable the callgraph.
+        This option extends the perf report to show reference callgraphs,
+        which collected by reference event, in no callgraph event.
+
 include::callchain-overhead-calculation.txt[]
 
 SEE ALSO
diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index 467a23b14e2f..a32a64ef08e2 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -239,6 +239,8 @@ static int __cmd_annotate(struct perf_annotate *ann)
                 if (nr_samples > 0) {
                         total_nr_samples += nr_samples;
                         hists__collapse_resort(hists, NULL);
+                        /* Don't sort callchain */
+                        perf_evsel__reset_sample_bit(pos, CALLCHAIN);
                         hists__output_resort(hists, NULL);
 
                         if (symbol_conf.event_group &&
diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c
index daaa7dca9c3b..0b180a885ba3 100644
--- a/tools/perf/builtin-diff.c
+++ b/tools/perf/builtin-diff.c
@@ -722,6 +722,9 @@ static void data_process(void)
                 if (verbose || data__files_cnt > 2)
                         data__fprintf();
 
+                /* Don't sort callchain for perf diff */
+                perf_evsel__reset_sample_bit(evsel_base, CALLCHAIN);
+
                 hists__process(hists_base);
         }
 }
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index f301e865001f..62b285e32aa5 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -316,6 +316,11 @@ static size_t hists__fprintf_nr_sample_events(struct hists *hists, struct report
         if (evname != NULL)
                 ret += fprintf(fp, " of event '%s'", evname);
 
+        if (symbol_conf.show_ref_callgraph &&
+            strstr(evname, "call-graph=no")) {
+                ret += fprintf(fp, ", show reference callgraph");
+        }
+
         if (rep->mem_mode) {
                 ret += fprintf(fp, "\n# Total weight : %" PRIu64, nr_events);
                 ret += fprintf(fp, "\n# Sort order   : %s", sort_order ? : default_mem_sort_order);
@@ -740,6 +745,8 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
                             itrace_parse_synth_opts),
         OPT_BOOLEAN(0, "full-source-path", &srcline_full_filename,
                         "Show full source file name path for source lines"),
+        OPT_BOOLEAN(0, "show-ref-call-graph", &symbol_conf.show_ref_callgraph,
+                    "Show callgraph from reference event"),
         OPT_END()
         };
         struct perf_data_file file = {
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index a25048c85b76..489cc118a36a 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -569,6 +569,15 @@ static DEFINE_STRARRAY_OFFSET(epoll_ctl_ops, 1);
 static const char *itimers[] = { "REAL", "VIRTUAL", "PROF", };
 static DEFINE_STRARRAY(itimers);
 
+static const char *keyctl_options[] = {
+        "GET_KEYRING_ID", "JOIN_SESSION_KEYRING", "UPDATE", "REVOKE", "CHOWN",
+        "SETPERM", "DESCRIBE", "CLEAR", "LINK", "UNLINK", "SEARCH", "READ",
+        "INSTANTIATE", "NEGATE", "SET_REQKEY_KEYRING", "SET_TIMEOUT",
+        "ASSUME_AUTHORITY", "GET_SECURITY", "SESSION_TO_PARENT", "REJECT",
+        "INSTANTIATE_IOV", "INVALIDATE", "GET_PERSISTENT",
+};
+static DEFINE_STRARRAY(keyctl_options);
+
 static const char *whences[] = { "SET", "CUR", "END",
 #ifdef SEEK_DATA
 "DATA",
@@ -599,7 +608,8 @@ static DEFINE_STRARRAY(sighow);
 
 static const char *clockid[] = {
         "REALTIME", "MONOTONIC", "PROCESS_CPUTIME_ID", "THREAD_CPUTIME_ID",
-        "MONOTONIC_RAW", "REALTIME_COARSE", "MONOTONIC_COARSE",
+        "MONOTONIC_RAW", "REALTIME_COARSE", "MONOTONIC_COARSE", "BOOTTIME",
+        "REALTIME_ALARM", "BOOTTIME_ALARM", "SGI_CYCLE", "TAI"
 };
 static DEFINE_STRARRAY(clockid);
 
@@ -1014,7 +1024,8 @@ static struct syscall_fmt {
         { .name     = "fchmod",     .errmsg = true,
           .arg_scnprintf = { [0] = SCA_FD, /* fd */ }, },
         { .name     = "fchmodat",   .errmsg = true,
-          .arg_scnprintf = { [0] = SCA_FDAT, /* fd */ }, },
+          .arg_scnprintf = { [0] = SCA_FDAT, /* fd */
+                             [1] = SCA_FILENAME, /* filename */ }, },
         { .name     = "fchown",     .errmsg = true,
           .arg_scnprintf = { [0] = SCA_FD, /* fd */ }, },
         { .name     = "fchownat",   .errmsg = true,
@@ -1045,7 +1056,8 @@ static struct syscall_fmt {
         { .name     = "futex",      .errmsg = true,
           .arg_scnprintf = { [1] = SCA_FUTEX_OP, /* op */ }, },
         { .name     = "futimesat", .errmsg = true,
-          .arg_scnprintf = { [0] = SCA_FDAT, /* fd */ }, },
+          .arg_scnprintf = { [0] = SCA_FDAT, /* fd */
+                             [1] = SCA_FILENAME, /* filename */ }, },
         { .name     = "getdents",   .errmsg = true,
           .arg_scnprintf = { [0] = SCA_FD, /* fd */ }, },
         { .name     = "getdents64", .errmsg = true,
@@ -1068,6 +1080,7 @@ static struct syscall_fmt {
 #else
                              [2] = SCA_HEX, /* arg */ }, },
 #endif
+        { .name     = "keyctl",     .errmsg = true, STRARRAY(0, option, keyctl_options), },
         { .name     = "kill",       .errmsg = true,
           .arg_scnprintf = { [1] = SCA_SIGNUM, /* sig */ }, },
         { .name     = "lchown",    .errmsg = true,
@@ -1078,13 +1091,18 @@ static struct syscall_fmt {
           .arg_scnprintf = { [0] = SCA_FDAT, /* fd */ }, },
         { .name     = "listxattr",  .errmsg = true,
           .arg_scnprintf = { [0] = SCA_FILENAME, /* pathname */ }, },
+        { .name     = "llistxattr", .errmsg = true,
+          .arg_scnprintf = { [0] = SCA_FILENAME, /* pathname */ }, },
+        { .name     = "lremovexattr",  .errmsg = true,
+          .arg_scnprintf = { [0] = SCA_FILENAME, /* pathname */ }, },
         { .name     = "lseek",      .errmsg = true,
           .arg_scnprintf = { [0] = SCA_FD, /* fd */
                              [2] = SCA_STRARRAY, /* whence */ },
           .arg_parm      = { [2] = &strarray__whences, /* whence */ }, },
         { .name     = "lsetxattr",  .errmsg = true,
           .arg_scnprintf = { [0] = SCA_FILENAME, /* pathname */ }, },
-        { .name     = "lstat",      .errmsg = true, .alias = "newlstat", },
+        { .name     = "lstat",      .errmsg = true, .alias = "newlstat",
+          .arg_scnprintf = { [0] = SCA_FILENAME, /* filename */ }, },
         { .name     = "lsxattr",    .errmsg = true,
           .arg_scnprintf = { [0] = SCA_FILENAME, /* pathname */ }, },
         { .name     = "madvise",    .errmsg = true,
@@ -1098,7 +1116,8 @@ static struct syscall_fmt {
         { .name     = "mknod",      .errmsg = true,
           .arg_scnprintf = { [0] = SCA_FILENAME, /* filename */ }, },
         { .name     = "mknodat",    .errmsg = true,
-          .arg_scnprintf = { [0] = SCA_FDAT, /* fd */ }, },
+          .arg_scnprintf = { [0] = SCA_FDAT, /* fd */
+                             [1] = SCA_FILENAME, /* filename */ }, },
         { .name     = "mlock",      .errmsg = true,
           .arg_scnprintf = { [0] = SCA_HEX, /* addr */ }, },
         { .name     = "mlockall",   .errmsg = true,
@@ -1111,6 +1130,8 @@ static struct syscall_fmt {
         { .name     = "mprotect",   .errmsg = true,
           .arg_scnprintf = { [0] = SCA_HEX, /* start */
                              [2] = SCA_MMAP_PROT, /* prot */ }, },
+        { .name     = "mq_unlink", .errmsg = true,
+          .arg_scnprintf = { [0] = SCA_FILENAME, /* u_name */ }, },
         { .name     = "mremap",     .hexret = true,
           .arg_scnprintf = { [0] = SCA_HEX, /* addr */
                              [3] = SCA_MREMAP_FLAGS, /* flags */
@@ -1162,11 +1183,14 @@ static struct syscall_fmt {
         { .name     = "readv",      .errmsg = true,
           .arg_scnprintf = { [0] = SCA_FD, /* fd */ }, },
         { .name     = "recvfrom",   .errmsg = true,
-          .arg_scnprintf = { [3] = SCA_MSG_FLAGS, /* flags */ }, },
+          .arg_scnprintf = { [0] = SCA_FD, /* fd */
+                             [3] = SCA_MSG_FLAGS, /* flags */ }, },
         { .name     = "recvmmsg",   .errmsg = true,
-          .arg_scnprintf = { [3] = SCA_MSG_FLAGS, /* flags */ }, },
+          .arg_scnprintf = { [0] = SCA_FD, /* fd */
+                             [3] = SCA_MSG_FLAGS, /* flags */ }, },
         { .name     = "recvmsg",    .errmsg = true,
-          .arg_scnprintf = { [2] = SCA_MSG_FLAGS, /* flags */ }, },
+          .arg_scnprintf = { [0] = SCA_FD, /* fd */
+                             [2] = SCA_MSG_FLAGS, /* flags */ }, },
         { .name     = "removexattr", .errmsg = true,
           .arg_scnprintf = { [0] = SCA_FILENAME, /* pathname */ }, },
         { .name     = "renameat",   .errmsg = true,
@@ -1182,11 +1206,14 @@ static struct syscall_fmt {
           .arg_scnprintf = { [2] = SCA_SIGNUM, /* sig */ }, },
         { .name     = "select",     .errmsg = true, .timeout = true, },
         { .name     = "sendmmsg",    .errmsg = true,
-          .arg_scnprintf = { [3] = SCA_MSG_FLAGS, /* flags */ }, },
+          .arg_scnprintf = { [0] = SCA_FD, /* fd */
+                             [3] = SCA_MSG_FLAGS, /* flags */ }, },
         { .name     = "sendmsg",    .errmsg = true,
-          .arg_scnprintf = { [2] = SCA_MSG_FLAGS, /* flags */ }, },
+          .arg_scnprintf = { [0] = SCA_FD, /* fd */
+                             [2] = SCA_MSG_FLAGS, /* flags */ }, },
         { .name     = "sendto",     .errmsg = true,
-          .arg_scnprintf = { [3] = SCA_MSG_FLAGS, /* flags */ }, },
+          .arg_scnprintf = { [0] = SCA_FD, /* fd */
+                             [3] = SCA_MSG_FLAGS, /* flags */ }, },
         { .name     = "setitimer",  .errmsg = true, STRARRAY(0, which, itimers), },
         { .name     = "setrlimit",  .errmsg = true, STRARRAY(0, resource, rlimit_resources), },
         { .name     = "setxattr",   .errmsg = true,
@@ -1201,7 +1228,8 @@ static struct syscall_fmt {
           .arg_scnprintf = { [0] = SCA_STRARRAY, /* family */
                              [1] = SCA_SK_TYPE, /* type */ },
           .arg_parm      = { [0] = &strarray__socket_families, /* family */ }, },
-        { .name     = "stat",       .errmsg = true, .alias = "newstat", },
+        { .name     = "stat",       .errmsg = true, .alias = "newstat",
+          .arg_scnprintf = { [0] = SCA_FILENAME, /* pathname */ }, },
         { .name     = "statfs",     .errmsg = true,
           .arg_scnprintf = { [0] = SCA_FILENAME, /* pathname */ }, },
         { .name     = "swapoff",    .errmsg = true,
@@ -1227,6 +1255,8 @@ static struct syscall_fmt {
                              [1] = SCA_FILENAME, /* filename */ }, },
         { .name     = "utimes",  .errmsg = true,
           .arg_scnprintf = { [0] = SCA_FILENAME, /* filename */ }, },
+        { .name     = "vmsplice",  .errmsg = true,
+          .arg_scnprintf = { [0] = SCA_FD, /* fd */ }, },
         { .name     = "write",      .errmsg = true,
           .arg_scnprintf = { [0] = SCA_FD, /* fd */ }, },
         { .name     = "writev",     .errmsg = true,
diff --git a/tools/perf/tests/hists_cumulate.c b/tools/perf/tests/hists_cumulate.c
index 7d82c8be5e36..7ed737019de7 100644
--- a/tools/perf/tests/hists_cumulate.c
+++ b/tools/perf/tests/hists_cumulate.c
@@ -279,6 +279,7 @@ static int test1(struct perf_evsel *evsel, struct machine *machine)
 
         symbol_conf.use_callchain = false;
         symbol_conf.cumulate_callchain = false;
+        perf_evsel__reset_sample_bit(evsel, CALLCHAIN);
 
         setup_sorting();
         callchain_register_param(&callchain_param);
@@ -425,6 +426,7 @@ static int test2(struct perf_evsel *evsel, struct machine *machine)
 
         symbol_conf.use_callchain = true;
         symbol_conf.cumulate_callchain = false;
+        perf_evsel__set_sample_bit(evsel, CALLCHAIN);
 
         setup_sorting();
         callchain_register_param(&callchain_param);
@@ -482,6 +484,7 @@ static int test3(struct perf_evsel *evsel, struct machine *machine)
 
         symbol_conf.use_callchain = false;
         symbol_conf.cumulate_callchain = true;
+        perf_evsel__reset_sample_bit(evsel, CALLCHAIN);
 
         setup_sorting();
         callchain_register_param(&callchain_param);
@@ -665,6 +668,7 @@ static int test4(struct perf_evsel *evsel, struct machine *machine)
 
         symbol_conf.use_callchain = true;
         symbol_conf.cumulate_callchain = true;
+        perf_evsel__set_sample_bit(evsel, CALLCHAIN);
 
         setup_sorting();
         callchain_register_param(&callchain_param);
diff --git a/tools/perf/tests/parse-events.c b/tools/perf/tests/parse-events.c
index f65bb89e109e..9b6b2b6324a1 100644
--- a/tools/perf/tests/parse-events.c
+++ b/tools/perf/tests/parse-events.c
@@ -479,6 +479,39 @@ static int test__checkevent_pmu_name(struct perf_evlist *evlist)
         return 0;
 }
 
+static int test__checkevent_pmu_partial_time_callgraph(struct perf_evlist *evlist)
+{
+        struct perf_evsel *evsel = perf_evlist__first(evlist);
+
+        /* cpu/config=1,call-graph=fp,time,period=100000/ */
+        TEST_ASSERT_VAL("wrong number of entries", 2 == evlist->nr_entries);
+        TEST_ASSERT_VAL("wrong type", PERF_TYPE_RAW == evsel->attr.type);
+        TEST_ASSERT_VAL("wrong config",  1 == evsel->attr.config);
+        /*
+         * The period, time and callgraph value gets configured
+         * within perf_evlist__config,
+         * while this test executes only parse events method.
+         */
+        TEST_ASSERT_VAL("wrong period",     0 == evsel->attr.sample_period);
+        TEST_ASSERT_VAL("wrong callgraph",  !(PERF_SAMPLE_CALLCHAIN & evsel->attr.sample_type));
+        TEST_ASSERT_VAL("wrong time",  !(PERF_SAMPLE_TIME & evsel->attr.sample_type));
+
+        /* cpu/config=2,call-graph=no,time=0,period=2000/ */
+        evsel = perf_evsel__next(evsel);
+        TEST_ASSERT_VAL("wrong type", PERF_TYPE_RAW == evsel->attr.type);
+        TEST_ASSERT_VAL("wrong config",  2 == evsel->attr.config);
+        /*
+         * The period, time and callgraph value gets configured
+         * within perf_evlist__config,
+         * while this test executes only parse events method.
+         */
+        TEST_ASSERT_VAL("wrong period",     0 == evsel->attr.sample_period);
+        TEST_ASSERT_VAL("wrong callgraph",  !(PERF_SAMPLE_CALLCHAIN & evsel->attr.sample_type));
+        TEST_ASSERT_VAL("wrong time",  !(PERF_SAMPLE_TIME & evsel->attr.sample_type));
+
+        return 0;
+}
+
 static int test__checkevent_pmu_events(struct perf_evlist *evlist)
 {
         struct perf_evsel *evsel = perf_evlist__first(evlist);
@@ -1555,6 +1588,11 @@ static struct evlist_test test__events_pmu[] = {
                 .check = test__checkevent_pmu_name,
                 .id    = 1,
         },
+        {
+                .name  = "cpu/config=1,call-graph=fp,time,period=100000/,cpu/config=2,call-graph=no,time=0,period=2000/",
+                .check = test__checkevent_pmu_partial_time_callgraph,
+                .id    = 2,
+        },
 };
 
 struct terms_test {
diff --git a/tools/perf/ui/browser.c b/tools/perf/ui/browser.c
index 6680fa5cb9dd..c6c7e5189214 100644
--- a/tools/perf/ui/browser.c
+++ b/tools/perf/ui/browser.c
@@ -46,6 +46,21 @@ void ui_browser__gotorc(struct ui_browser *browser, int y, int x)
         SLsmg_gotorc(browser->y + y, browser->x + x);
 }
 
+void ui_browser__write_nstring(struct ui_browser *browser __maybe_unused, const char *msg,
+                               unsigned int width)
+{
+        slsmg_write_nstring(msg, width);
+}
+
+void ui_browser__printf(struct ui_browser *browser __maybe_unused, const char *fmt, ...)
+{
+        va_list args;
+
+        va_start(args, fmt);
+        slsmg_vprintf(fmt, args);
+        va_end(args);
+}
+
 static struct list_head *
 ui_browser__list_head_filter_entries(struct ui_browser *browser,
                                      struct list_head *pos)
@@ -234,7 +249,7 @@ void __ui_browser__show_title(struct ui_browser *browser, const char *title)
 {
         SLsmg_gotorc(0, 0);
         ui_browser__set_color(browser, HE_COLORSET_ROOT);
-        slsmg_write_nstring(title, browser->width + 1);
+        ui_browser__write_nstring(browser, title, browser->width + 1);
 }
 
 void ui_browser__show_title(struct ui_browser *browser, const char *title)
diff --git a/tools/perf/ui/browser.h b/tools/perf/ui/browser.h
index 92ae72113965..52be87125def 100644
--- a/tools/perf/ui/browser.h
+++ b/tools/perf/ui/browser.h
@@ -37,6 +37,9 @@ void ui_browser__refresh_dimensions(struct ui_browser *browser);
 void ui_browser__reset_index(struct ui_browser *browser);
 
 void ui_browser__gotorc(struct ui_browser *browser, int y, int x);
+void ui_browser__write_nstring(struct ui_browser *browser, const char *msg,
+                               unsigned int width);
+void ui_browser__printf(struct ui_browser *browser, const char *fmt, ...);
 void ui_browser__write_graph(struct ui_browser *browser, int graph);
 void __ui_browser__line_arrow(struct ui_browser *browser, unsigned int column,
                               u64 start, u64 end);
diff --git a/tools/perf/ui/browsers/annotate.c b/tools/perf/ui/browsers/annotate.c
index b5fc847f9660..29739b347599 100644
--- a/tools/perf/ui/browsers/annotate.c
+++ b/tools/perf/ui/browsers/annotate.c
@@ -1,7 +1,6 @@
 #include "../../util/util.h"
 #include "../browser.h"
 #include "../helpline.h"
-#include "../libslang.h"
 #include "../ui.h"
 #include "../util.h"
 #include "../../util/annotate.h"
@@ -134,29 +133,31 @@ static void annotate_browser__write(struct ui_browser *browser, void *entry, int
                                 ui_browser__set_percent_color(browser,
                                                         bdl->samples[i].percent,
                                                         current_entry);
-                                if (annotate_browser__opts.show_total_period)
-                                        slsmg_printf("%6" PRIu64 " ",
-                                                     bdl->samples[i].nr);
-                                else
-                                        slsmg_printf("%6.2f ", bdl->samples[i].percent);
+                                if (annotate_browser__opts.show_total_period) {
+                                        ui_browser__printf(browser, "%6" PRIu64 " ",
+                                                           bdl->samples[i].nr);
+                                } else {
+                                        ui_browser__printf(browser, "%6.2f ",
+                                                           bdl->samples[i].percent);
+                                }
                         }
                 } else {
-                        slsmg_write_nstring(" ", 7 * ab->nr_events);
+                        ui_browser__write_nstring(browser, " ", 7 * ab->nr_events);
                 }
         } else {
                 ui_browser__set_percent_color(browser, 0, current_entry);
-                slsmg_write_nstring(" ", 7 * ab->nr_events);
+                ui_browser__write_nstring(browser, " ", 7 * ab->nr_events);
         }
         if (ab->have_cycles) {
                 if (dl->ipc)
-                        slsmg_printf("%*.2f ", IPC_WIDTH - 1, dl->ipc);
+                        ui_browser__printf(browser, "%*.2f ", IPC_WIDTH - 1, dl->ipc);
                 else
-                        slsmg_write_nstring(" ", IPC_WIDTH);
+                        ui_browser__write_nstring(browser, " ", IPC_WIDTH);
                 if (dl->cycles)
-                        slsmg_printf("%*" PRIu64 " ",
-                                     CYCLES_WIDTH - 1, dl->cycles);
+                        ui_browser__printf(browser, "%*" PRIu64 " ",
+                                           CYCLES_WIDTH - 1, dl->cycles);
                 else
-                        slsmg_write_nstring(" ", CYCLES_WIDTH);
+                        ui_browser__write_nstring(browser, " ", CYCLES_WIDTH);
         }
 
         SLsmg_write_char(' ');
@@ -166,7 +167,7 @@ static void annotate_browser__write(struct ui_browser *browser, void *entry, int
                 width += 1;
 
         if (!*dl->line)
-                slsmg_write_nstring(" ", width - pcnt_width);
+                ui_browser__write_nstring(browser, " ", width - pcnt_width);
         else if (dl->offset == -1) {
                 if (dl->line_nr && annotate_browser__opts.show_linenr)
                         printed = scnprintf(bf, sizeof(bf), "%-*d ",
@@ -174,8 +175,8 @@ static void annotate_browser__write(struct ui_browser *browser, void *entry, int
                 else
                         printed = scnprintf(bf, sizeof(bf), "%*s  ",
                                     ab->addr_width, " ");
-                slsmg_write_nstring(bf, printed);
-                slsmg_write_nstring(dl->line, width - printed - pcnt_width + 1);
+                ui_browser__write_nstring(browser, bf, printed);
+                ui_browser__write_nstring(browser, dl->line, width - printed - pcnt_width + 1);
         } else {
                 u64 addr = dl->offset;
                 int color = -1;
@@ -194,7 +195,7 @@ static void annotate_browser__write(struct ui_browser *browser, void *entry, int
                                                             bdl->jump_sources);
                                         prev = annotate_browser__set_jumps_percent_color(ab, bdl->jump_sources,
                                                                                          current_entry);
-                                        slsmg_write_nstring(bf, printed);
+                                        ui_browser__write_nstring(browser, bf, printed);
                                         ui_browser__set_color(browser, prev);
                                 }
 
@@ -208,7 +209,7 @@ static void annotate_browser__write(struct ui_browser *browser, void *entry, int
 
                 if (change_color)
                         color = ui_browser__set_color(browser, HE_COLORSET_ADDR);
-                slsmg_write_nstring(bf, printed);
+                ui_browser__write_nstring(browser, bf, printed);
                 if (change_color)
                         ui_browser__set_color(browser, color);
                 if (dl->ins && dl->ins->ops->scnprintf) {
@@ -222,11 +223,11 @@ static void annotate_browser__write(struct ui_browser *browser, void *entry, int
                                 ui_browser__write_graph(browser, SLSMG_RARROW_CHAR);
                                 SLsmg_write_char(' ');
                         } else {
-                                slsmg_write_nstring(" ", 2);
+                                ui_browser__write_nstring(browser, " ", 2);
                         }
                 } else {
                         if (strcmp(dl->name, "retq")) {
-                                slsmg_write_nstring(" ", 2);
+                                ui_browser__write_nstring(browser, " ", 2);
                         } else {
                                 ui_browser__write_graph(browser, SLSMG_LARROW_CHAR);
                                 SLsmg_write_char(' ');
@@ -234,7 +235,7 @@ static void annotate_browser__write(struct ui_browser *browser, void *entry, int
                 }
 
                 disasm_line__scnprintf(dl, bf, sizeof(bf), !annotate_browser__opts.use_offset);
-                slsmg_write_nstring(bf, width - pcnt_width - 3 - printed);
+                ui_browser__write_nstring(browser, bf, width - pcnt_width - 3 - printed);
         }
 
         if (current_entry)
diff --git a/tools/perf/ui/browsers/header.c b/tools/perf/ui/browsers/header.c
index e8278c558d4a..f10681717f93 100644
--- a/tools/perf/ui/browsers/header.c
+++ b/tools/perf/ui/browsers/header.c
@@ -25,7 +25,7 @@ static void ui_browser__argv_write(struct ui_browser *browser,
         ui_browser__set_color(browser, current_entry ? HE_COLORSET_SELECTED :
                                                        HE_COLORSET_NORMAL);
 
-        slsmg_write_nstring(str, browser->width);
+        ui_browser__write_nstring(browser, str, browser->width);
 }
 
 static int list_menu__run(struct ui_browser *menu)
diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c
index fa67613976a8..10c7ec041039 100644
--- a/tools/perf/ui/browsers/hists.c
+++ b/tools/perf/ui/browsers/hists.c
@@ -1,5 +1,4 @@
 #include <stdio.h>
-#include "../libslang.h"
 #include <stdlib.h>
 #include <string.h>
 #include <linux/rbtree.h>
@@ -540,10 +539,10 @@ static void hist_browser__show_callchain_entry(struct hist_browser *browser,
 
         ui_browser__set_color(&browser->b, color);
         hist_browser__gotorc(browser, row, 0);
-        slsmg_write_nstring(" ", offset);
-        slsmg_printf("%c", folded_sign);
+        ui_browser__write_nstring(&browser->b, " ", offset);
+        ui_browser__printf(&browser->b, "%c", folded_sign);
         ui_browser__write_graph(&browser->b, show_annotated ? SLSMG_RARROW_CHAR : ' ');
-        slsmg_write_nstring(str, width);
+        ui_browser__write_nstring(&browser->b, str, width);
 }
 
 static void hist_browser__fprintf_callchain_entry(struct hist_browser *b __maybe_unused,
@@ -680,7 +679,7 @@ static int __hpp__slsmg_color_printf(struct perf_hpp *hpp, const char *fmt, ...)
         ui_browser__set_percent_color(arg->b, percent, arg->current_entry);
 
         ret = scnprintf(hpp->buf, hpp->size, fmt, len, percent);
-        slsmg_printf("%s", hpp->buf);
+        ui_browser__printf(arg->b, "%s", hpp->buf);
 
         advance_hpp(hpp, ret);
         return ret;
@@ -713,10 +712,11 @@ hist_browser__hpp_color_##_type(struct perf_hpp_fmt *fmt,		\
                                 struct hist_entry *he)                  \
 {                                                                       \
         if (!symbol_conf.cumulate_callchain) {                          \
+                struct hpp_arg *arg = hpp->ptr;                         \
                 int len = fmt->user_len ?: fmt->len;                    \
                 int ret = scnprintf(hpp->buf, hpp->size,                \
                                     "%*s", len, "N/A");                 \
-                slsmg_printf("%s", hpp->buf);                           \
+                ui_browser__printf(arg->b, "%s", hpp->buf);             \
                                                                         \
                 return ret;                                             \
         }                                                               \
@@ -801,12 +801,12 @@ static int hist_browser__show_entry(struct hist_browser *browser,
 
                         if (first) {
                                 if (symbol_conf.use_callchain) {
-                                        slsmg_printf("%c ", folded_sign);
+                                        ui_browser__printf(&browser->b, "%c ", folded_sign);
                                         width -= 2;
                                 }
                                 first = false;
                         } else {
-                                slsmg_printf("  ");
+                                ui_browser__printf(&browser->b, "  ");
                                 width -= 2;
                         }
 
@@ -814,7 +814,7 @@ static int hist_browser__show_entry(struct hist_browser *browser,
                                 width -= fmt->color(fmt, &hpp, entry);
                         } else {
                                 width -= fmt->entry(fmt, &hpp, entry);
-                                slsmg_printf("%s", s);
+                                ui_browser__printf(&browser->b, "%s", s);
                         }
                 }
 
@@ -822,7 +822,7 @@ static int hist_browser__show_entry(struct hist_browser *browser,
                 if (!browser->b.navkeypressed)
                         width += 1;
 
-                slsmg_write_nstring("", width);
+                ui_browser__write_nstring(&browser->b, "", width);
 
                 ++row;
                 ++printed;
@@ -899,7 +899,7 @@ static void hist_browser__show_headers(struct hist_browser *browser)
         hists__scnprintf_headers(headers, sizeof(headers), browser->hists);
         ui_browser__gotorc(&browser->b, 0, 0);
         ui_browser__set_color(&browser->b, HE_COLORSET_ROOT);
-        slsmg_write_nstring(headers, browser->b.width + 1);
+        ui_browser__write_nstring(&browser->b, headers, browser->b.width + 1);
 }
 
 static void ui_browser__hists_init_top(struct ui_browser *browser)
@@ -1267,6 +1267,8 @@ static int hists__browser_title(struct hists *hists,
         const char *ev_name = perf_evsel__name(evsel);
         char buf[512];
         size_t buflen = sizeof(buf);
+        char ref[30] = " show reference callgraph, ";
+        bool enable_ref = false;
 
         if (symbol_conf.filter_relative) {
                 nr_samples = hists->stats.nr_non_filtered_samples;
@@ -1292,10 +1294,13 @@ static int hists__browser_title(struct hists *hists,
                 }
         }
 
+        if (symbol_conf.show_ref_callgraph &&
+            strstr(ev_name, "call-graph=no"))
+                enable_ref = true;
         nr_samples = convert_unit(nr_samples, &unit);
         printed = scnprintf(bf, size,
-                           "Samples: %lu%c of event '%s', Event count (approx.): %" PRIu64,
-                           nr_samples, unit, ev_name, nr_events);
+                           "Samples: %lu%c of event '%s',%sEvent count (approx.): %" PRIu64,
+                           nr_samples, unit, ev_name, enable_ref ? ref : " ", nr_events);
 
 
         if (hists->uid_filter_str)
@@ -1868,6 +1873,7 @@ static int perf_evsel__hists_browse(struct perf_evsel *evsel, int nr_events,
                 case K_RIGHT:
                         /* menu */
                         break;
+                case K_ESC:
                 case K_LEFT: {
                         const void *top;
 
@@ -1877,6 +1883,12 @@ static int perf_evsel__hists_browse(struct perf_evsel *evsel, int nr_events,
                                  */
                                 if (left_exits)
                                         goto out_free_stack;
+
+                                if (key == K_ESC &&
+                                    ui_browser__dialog_yesno(&browser->b,
+                                                             "Do you really want to exit?"))
+                                        goto out_free_stack;
+
                                 continue;
                         }
                         top = pstack__peek(browser->pstack);
@@ -1892,12 +1904,6 @@ static int perf_evsel__hists_browse(struct perf_evsel *evsel, int nr_events,
                                 do_zoom_thread(browser, actions);
                         continue;
                 }
-                case K_ESC:
-                        if (!left_exits &&
-                            !ui_browser__dialog_yesno(&browser->b,
-                                               "Do you really want to exit?"))
-                                continue;
-                        /* Fall thru */
                 case 'q':
                 case CTRL('c'):
                         goto out_free_stack;
@@ -2044,7 +2050,7 @@ static void perf_evsel_menu__write(struct ui_browser *browser,
         nr_events = convert_unit(nr_events, &unit);
         printed = scnprintf(bf, sizeof(bf), "%lu%c%s%s", nr_events,
                            unit, unit == ' ' ? "" : " ", ev_name);
-        slsmg_printf("%s", bf);
+        ui_browser__printf(browser, "%s", bf);
 
         nr_events = hists->stats.nr_events[PERF_RECORD_LOST];
         if (nr_events != 0) {
@@ -2057,7 +2063,7 @@ static void perf_evsel_menu__write(struct ui_browser *browser,
                 warn = bf;
         }
 
-        slsmg_write_nstring(warn, browser->width - printed);
+        ui_browser__write_nstring(browser, warn, browser->width - printed);
 
         if (current_entry)
                 menu->selection = evsel;
@@ -2120,15 +2126,11 @@ browse_hists:
                                 else
                                         pos = perf_evsel__prev(pos);
                                 goto browse_hists;
-                        case K_ESC:
-                                if (!ui_browser__dialog_yesno(&menu->b,
-                                                "Do you really want to exit?"))
-                                        continue;
-                                /* Fall thru */
                         case K_SWITCH_INPUT_DATA:
                         case 'q':
                         case CTRL('c'):
                                 goto out;
+                        case K_ESC:
                         default:
                                 continue;
                         }
diff --git a/tools/perf/ui/browsers/map.c b/tools/perf/ui/browsers/map.c
index b11639f33682..8c154c7d4669 100644
--- a/tools/perf/ui/browsers/map.c
+++ b/tools/perf/ui/browsers/map.c
@@ -1,4 +1,3 @@
-#include "../libslang.h"
 #include <elf.h>
 #include <inttypes.h>
 #include <sys/ttydefaults.h>
@@ -26,13 +25,13 @@ static void map_browser__write(struct ui_browser *browser, void *nd, int row)
         int width;
 
         ui_browser__set_percent_color(browser, 0, current_entry);
-        slsmg_printf("%*" PRIx64 " %*" PRIx64 " %c ",
-                     mb->addrlen, sym->start, mb->addrlen, sym->end,
-                     sym->binding == STB_GLOBAL ? 'g' :
-                     sym->binding == STB_LOCAL  ? 'l' : 'w');
+        ui_browser__printf(browser, "%*" PRIx64 " %*" PRIx64 " %c ",
+                           mb->addrlen, sym->start, mb->addrlen, sym->end,
+                           sym->binding == STB_GLOBAL ? 'g' :
+                                sym->binding == STB_LOCAL  ? 'l' : 'w');
         width = browser->width - ((mb->addrlen * 2) + 4);
         if (width > 0)
-                slsmg_write_nstring(sym->name, width);
+                ui_browser__write_nstring(browser, sym->name, width);
 }
 
 /* FIXME uber-kludgy, see comment on cmd_report... */
diff --git a/tools/perf/ui/browsers/scripts.c b/tools/perf/ui/browsers/scripts.c
index 402d2bd30b09..e13b48d705ef 100644
--- a/tools/perf/ui/browsers/scripts.c
+++ b/tools/perf/ui/browsers/scripts.c
@@ -81,7 +81,7 @@ static void script_browser__write(struct ui_browser *browser,
         ui_browser__set_color(browser, current_entry ? HE_COLORSET_SELECTED :
                                                        HE_COLORSET_NORMAL);
 
-        slsmg_write_nstring(sline->line, browser->width);
+        ui_browser__write_nstring(browser, sline->line, browser->width);
 }
 
 static int script_browser__run(struct perf_script_browser *browser)
diff --git a/tools/perf/ui/libslang.h b/tools/perf/ui/libslang.h
index 4d54b6450f5b..db816695ad97 100644
--- a/tools/perf/ui/libslang.h
+++ b/tools/perf/ui/libslang.h
@@ -14,12 +14,15 @@
 #if SLANG_VERSION < 20104
 #define slsmg_printf(msg, args...) \
         SLsmg_printf((char *)(msg), ##args)
+#define slsmg_vprintf(msg, vargs) \
+        SLsmg_vprintf((char *)(msg), vargs)
 #define slsmg_write_nstring(msg, len) \
         SLsmg_write_nstring((char *)(msg), len)
 #define sltt_set_color(obj, name, fg, bg) \
         SLtt_set_color(obj,(char *)(name), (char *)(fg), (char *)(bg))
 #else
 #define slsmg_printf SLsmg_printf
+#define slsmg_vprintf SLsmg_vprintf
 #define slsmg_write_nstring SLsmg_write_nstring
 #define sltt_set_color SLtt_set_color
 #endif
diff --git a/tools/perf/ui/tui/util.c b/tools/perf/ui/tui/util.c
index bf890f72fe80..d96ad7c8325d 100644
--- a/tools/perf/ui/tui/util.c
+++ b/tools/perf/ui/tui/util.c
@@ -21,7 +21,7 @@ static void ui_browser__argv_write(struct ui_browser *browser,
 
         ui_browser__set_color(browser, current_entry ? HE_COLORSET_SELECTED :
                                                        HE_COLORSET_NORMAL);
-        slsmg_write_nstring(*arg, browser->width);
+        ui_browser__write_nstring(browser, *arg, browser->width);
 }
 
 static int popup_menu__run(struct ui_browser *menu)
diff --git a/tools/perf/util/dwarf-aux.c b/tools/perf/util/dwarf-aux.c
index 57f3ef41c2bc..445f455dd377 100644
--- a/tools/perf/util/dwarf-aux.c
+++ b/tools/perf/util/dwarf-aux.c
@@ -734,15 +734,18 @@ int die_walk_lines(Dwarf_Die *rt_die, line_walk_callback_t callback, void *data)
         Dwarf_Lines *lines;
         Dwarf_Line *line;
         Dwarf_Addr addr;
-        const char *fname;
+        const char *fname, *decf = NULL;
         int lineno, ret = 0;
+        int decl = 0, inl;
         Dwarf_Die die_mem, *cu_die;
         size_t nlines, i;
 
         /* Get the CU die */
-        if (dwarf_tag(rt_die) != DW_TAG_compile_unit)
+        if (dwarf_tag(rt_die) != DW_TAG_compile_unit) {
                 cu_die = dwarf_diecu(rt_die, &die_mem, NULL, NULL);
-        else
+                dwarf_decl_line(rt_die, &decl);
+                decf = dwarf_decl_file(rt_die);
+        } else
                 cu_die = rt_die;
         if (!cu_die) {
                 pr_debug2("Failed to get CU from given DIE.\n");
@@ -773,9 +776,14 @@ int die_walk_lines(Dwarf_Die *rt_die, line_walk_callback_t callback, void *data)
                          * The line is included in given function, and
                          * no inline block includes it.
                          */
-                        if (!dwarf_haspc(rt_die, addr) ||
-                            die_find_inlinefunc(rt_die, addr, &die_mem))
+                        if (!dwarf_haspc(rt_die, addr))
                                 continue;
+                        if (die_find_inlinefunc(rt_die, addr, &die_mem)) {
+                                dwarf_decl_line(&die_mem, &inl);
+                                if (inl != decl ||
+                                    decf != dwarf_decl_file(&die_mem))
+                                        continue;
+                        }
                 /* Get source line */
                 fname = dwarf_linesrc(line, NULL, NULL);
 
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index 04fddddc6b6f..b096ef7a240c 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -588,11 +588,36 @@ perf_evsel__config_callgraph(struct perf_evsel *evsel,
         }
 }
 
-static void apply_config_terms(struct perf_evsel *evsel)
+static void
+perf_evsel__reset_callgraph(struct perf_evsel *evsel,
+                            struct callchain_param *param)
+{
+        struct perf_event_attr *attr = &evsel->attr;
+
+        perf_evsel__reset_sample_bit(evsel, CALLCHAIN);
+        if (param->record_mode == CALLCHAIN_LBR) {
+                perf_evsel__reset_sample_bit(evsel, BRANCH_STACK);
+                attr->branch_sample_type &= ~(PERF_SAMPLE_BRANCH_USER |
+                                              PERF_SAMPLE_BRANCH_CALL_STACK);
+        }
+        if (param->record_mode == CALLCHAIN_DWARF) {
+                perf_evsel__reset_sample_bit(evsel, REGS_USER);
+                perf_evsel__reset_sample_bit(evsel, STACK_USER);
+        }
+}
+
+static void apply_config_terms(struct perf_evsel *evsel,
+                               struct record_opts *opts)
 {
         struct perf_evsel_config_term *term;
         struct list_head *config_terms = &evsel->config_terms;
         struct perf_event_attr *attr = &evsel->attr;
+        struct callchain_param param;
+        u32 dump_size = 0;
+        char *callgraph_buf = NULL;
+
+        /* callgraph default */
+        param.record_mode = callchain_param.record_mode;
 
         list_for_each_entry(term, config_terms, list) {
                 switch (term->type) {
@@ -610,10 +635,48 @@ static void apply_config_terms(struct perf_evsel *evsel)
                         else
                                 perf_evsel__reset_sample_bit(evsel, TIME);
                         break;
+                case PERF_EVSEL__CONFIG_TERM_CALLGRAPH:
+                        callgraph_buf = term->val.callgraph;
+                        break;
+                case PERF_EVSEL__CONFIG_TERM_STACK_USER:
+                        dump_size = term->val.stack_user;
+                        break;
                 default:
                         break;
                 }
         }
+
+        /* User explicitly set per-event callgraph, clear the old setting and reset. */
+        if ((callgraph_buf != NULL) || (dump_size > 0)) {
+
+                /* parse callgraph parameters */
+                if (callgraph_buf != NULL) {
+                        if (!strcmp(callgraph_buf, "no")) {
+                                param.enabled = false;
+                                param.record_mode = CALLCHAIN_NONE;
+                        } else {
+                                param.enabled = true;
+                                if (parse_callchain_record(callgraph_buf, &param)) {
+                                        pr_err("per-event callgraph setting for %s failed. "
+                                               "Apply callgraph global setting for it\n",
+                                               evsel->name);
+                                        return;
+                                }
+                        }
+                }
+                if (dump_size > 0) {
+                        dump_size = round_up(dump_size, sizeof(u64));
+                        param.dump_size = dump_size;
+                }
+
+                /* If global callgraph set, clear it */
+                if (callchain_param.enabled)
+                        perf_evsel__reset_callgraph(evsel, &callchain_param);
+
+                /* set perf-event callgraph */
+                if (param.enabled)
+                        perf_evsel__config_callgraph(evsel, opts, &param);
+        }
 }
 
 /*
@@ -812,7 +875,7 @@ void perf_evsel__config(struct perf_evsel *evsel, struct record_opts *opts)
          * Apply event specific term settings,
          * it overloads any global configuration.
          */
-        apply_config_terms(evsel);
+        apply_config_terms(evsel, opts);
 }
 
 static int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads)
diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h
index fdf2674ab339..93ac6b128149 100644
--- a/tools/perf/util/evsel.h
+++ b/tools/perf/util/evsel.h
@@ -41,6 +41,8 @@ enum {
         PERF_EVSEL__CONFIG_TERM_PERIOD,
         PERF_EVSEL__CONFIG_TERM_FREQ,
         PERF_EVSEL__CONFIG_TERM_TIME,
+        PERF_EVSEL__CONFIG_TERM_CALLGRAPH,
+        PERF_EVSEL__CONFIG_TERM_STACK_USER,
         PERF_EVSEL__CONFIG_TERM_MAX,
 };
 
@@ -51,6 +53,8 @@ struct perf_evsel_config_term {
                 u64     period;
                 u64     freq;
                 bool    time;
+                char    *callgraph;
+                u64     stack_user;
         } val;
 };
 
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index 6bccfae334b1..08b6cd945f1e 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -1109,13 +1109,14 @@ void hists__inc_stats(struct hists *hists, struct hist_entry *h)
 
 static void __hists__insert_output_entry(struct rb_root *entries,
                                          struct hist_entry *he,
-                                         u64 min_callchain_hits)
+                                         u64 min_callchain_hits,
+                                         bool use_callchain)
 {
         struct rb_node **p = &entries->rb_node;
         struct rb_node *parent = NULL;
         struct hist_entry *iter;
 
-        if (symbol_conf.use_callchain)
+        if (use_callchain)
                 callchain_param.sort(&he->sorted_chain, he->callchain,
                                       min_callchain_hits, &callchain_param);
 
@@ -1139,6 +1140,13 @@ void hists__output_resort(struct hists *hists, struct ui_progress *prog)
         struct rb_node *next;
         struct hist_entry *n;
         u64 min_callchain_hits;
+        struct perf_evsel *evsel = hists_to_evsel(hists);
+        bool use_callchain;
+
+        if (evsel && !symbol_conf.show_ref_callgraph)
+                use_callchain = evsel->attr.sample_type & PERF_SAMPLE_CALLCHAIN;
+        else
+                use_callchain = symbol_conf.use_callchain;
 
         min_callchain_hits = hists->stats.total_period * (callchain_param.min_percent / 100);
 
@@ -1157,7 +1165,7 @@ void hists__output_resort(struct hists *hists, struct ui_progress *prog)
                 n = rb_entry(next, struct hist_entry, rb_node_in);
                 next = rb_next(&n->rb_node_in);
 
-                __hists__insert_output_entry(&hists->entries, n, min_callchain_hits);
+                __hists__insert_output_entry(&hists->entries, n, min_callchain_hits, use_callchain);
                 hists__inc_stats(hists, n);
 
                 if (!n->filtered)
diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
index dbf315df4220..d826e6f515db 100644
--- a/tools/perf/util/parse-events.c
+++ b/tools/perf/util/parse-events.c
@@ -614,6 +614,12 @@ do {									   \
                         return -EINVAL;
                 }
                 break;
+        case PARSE_EVENTS__TERM_TYPE_CALLGRAPH:
+                CHECK_TYPE_VAL(STR);
+                break;
+        case PARSE_EVENTS__TERM_TYPE_STACKSIZE:
+                CHECK_TYPE_VAL(NUM);
+                break;
         case PARSE_EVENTS__TERM_TYPE_NAME:
                 CHECK_TYPE_VAL(STR);
                 break;
@@ -668,6 +674,12 @@ do {								\
                 case PARSE_EVENTS__TERM_TYPE_TIME:
                         ADD_CONFIG_TERM(TIME, time, term->val.num);
                         break;
+                case PARSE_EVENTS__TERM_TYPE_CALLGRAPH:
+                        ADD_CONFIG_TERM(CALLGRAPH, callgraph, term->val.str);
+                        break;
+                case PARSE_EVENTS__TERM_TYPE_STACKSIZE:
+                        ADD_CONFIG_TERM(STACK_USER, stack_user, term->val.num);
+                        break;
                 default:
                         break;
                 }
diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h
index ce2d13a16226..a09b0e210997 100644
--- a/tools/perf/util/parse-events.h
+++ b/tools/perf/util/parse-events.h
@@ -65,6 +65,8 @@ enum {
         PARSE_EVENTS__TERM_TYPE_SAMPLE_FREQ,
         PARSE_EVENTS__TERM_TYPE_BRANCH_SAMPLE_TYPE,
         PARSE_EVENTS__TERM_TYPE_TIME,
+        PARSE_EVENTS__TERM_TYPE_CALLGRAPH,
+        PARSE_EVENTS__TERM_TYPE_STACKSIZE,
 };
 
 struct parse_events_term {
diff --git a/tools/perf/util/parse-events.l b/tools/perf/util/parse-events.l
index 4306f5ad75c7..936d566f48d8 100644
--- a/tools/perf/util/parse-events.l
+++ b/tools/perf/util/parse-events.l
@@ -185,6 +185,8 @@ period			{ return term(yyscanner, PARSE_EVENTS__TERM_TYPE_SAMPLE_PERIOD); }
 freq                    { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_SAMPLE_FREQ); }
 branch_type             { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_BRANCH_SAMPLE_TYPE); }
 time                    { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_TIME); }
+call-graph              { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_CALLGRAPH); }
+stack-size              { return term(yyscanner, PARSE_EVENTS__TERM_TYPE_STACKSIZE); }
 ,                       { return ','; }
 "/"                     { BEGIN(INITIAL); return '/'; }
 {name_minus}            { return str(yyscanner, PE_NAME); }
diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c
index d85f11b8cacf..84cad054d6f7 100644
--- a/tools/perf/util/pmu.c
+++ b/tools/perf/util/pmu.c
@@ -634,7 +634,9 @@ static char *formats_error_string(struct list_head *formats)
 {
         struct perf_pmu_format *format;
         char *err, *str;
-        static const char *static_terms = "config,config1,config2,name,period,freq,branch_type,time\n";
+        static const char *static_terms = "config,config1,config2,name,"
+                                          "period,freq,branch_type,time,"
+                                          "call-graph,stack-size\n";
         unsigned i = 0;
 
         if (!asprintf(&str, "valid terms:"))
diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c
index c0c32b050e45..7e3871606df3 100644
--- a/tools/perf/util/sort.c
+++ b/tools/perf/util/sort.c
@@ -330,6 +330,8 @@ static char *get_srcfile(struct hist_entry *e)
 
         sf = get_srcline(map->dso, map__rip_2objdump(map, e->ip),
                          e->ms.sym, true);
+        if (!strcmp(sf, SRCLINE_UNKNOWN))
+                return no_srcfile;
         p = strchr(sf, ':');
         if (p && *sf) {
                 *p = 0;
diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h
index b98ce51af142..a4cde92afbad 100644
--- a/tools/perf/util/symbol.h
+++ b/tools/perf/util/symbol.h
@@ -106,7 +106,8 @@ struct symbol_conf {
                         filter_relative,
                         show_hist_headers,
                         branch_callstack,
-                        has_filter;
+                        has_filter,
+                        show_ref_callgraph;
         const char      *vmlinux_name,
                         *kallsyms_name,
                         *source_prefix,