Merge commit 'v2.6.37-rc1' into for-2.6.37

56 files changed, 1961 insertions, 538 deletions

diff --git a/tools/perf/Documentation/perf-annotate.txt b/tools/perf/Documentation/perf-annotate.txt
index 5164a655c39f..b2c63309a651 100644
--- a/tools/perf/Documentation/perf-annotate.txt
+++ b/tools/perf/Documentation/perf-annotate.txt
@@ -8,7 +8,7 @@ perf-annotate - Read perf.data (created by perf record) and display annotated co
 SYNOPSIS
 --------
 [verse]
-'perf annotate' [-i <file> | --input=file] symbol_name
+'perf annotate' [-i <file> | --input=file] [symbol_name]
 
 DESCRIPTION
 -----------
@@ -24,6 +24,13 @@ OPTIONS
 --input=::
         Input file name. (default: perf.data)
 
+--stdio:: Use the stdio interface.
+
+--tui:: Use the TUI interface Use of --tui requires a tty, if one is not
+        present, as when piping to other commands, the stdio interface is
+        used. This interfaces starts by centering on the line with more
+        samples, TAB/UNTAB cycles thru the lines with more samples.
+
 SEE ALSO
 --------
-linkperf:perf-record[1]
+linkperf:perf-record[1], linkperf:perf-report[1]
diff --git a/tools/perf/Documentation/perf-list.txt b/tools/perf/Documentation/perf-list.txt
index 43e3dd284b90..399751befeed 100644
--- a/tools/perf/Documentation/perf-list.txt
+++ b/tools/perf/Documentation/perf-list.txt
@@ -15,6 +15,23 @@ DESCRIPTION
 This command displays the symbolic event types which can be selected in the
 various perf commands with the -e option.
 
+EVENT MODIFIERS
+---------------
+
+Events can optionally have a modifer by appending a colon and one or
+more modifiers.  Modifiers allow the user to restrict when events are
+counted with 'u' for user-space, 'k' for kernel, 'h' for hypervisor.
+
+The 'p' modifier can be used for specifying how precise the instruction
+address should be. The 'p' modifier is currently only implemented for
+Intel PEBS and can be specified multiple times:
+  0 - SAMPLE_IP can have arbitrary skid
+  1 - SAMPLE_IP must have constant skid
+  2 - SAMPLE_IP requested to have 0 skid
+  3 - SAMPLE_IP must have 0 skid
+
+The PEBS implementation now supports up to 2.
+
 RAW HARDWARE EVENT DESCRIPTOR
 -----------------------------
 Even when an event is not available in a symbolic form within perf right now,
diff --git a/tools/perf/Documentation/perf-probe.txt b/tools/perf/Documentation/perf-probe.txt
index 27d52dae5a43..62de1b7f4e76 100644
--- a/tools/perf/Documentation/perf-probe.txt
+++ b/tools/perf/Documentation/perf-probe.txt
@@ -16,7 +16,9 @@ or
 or
 'perf probe' --list
 or
-'perf probe' --line='FUNC[:RLN[+NUM|:RLN2]]|SRC:ALN[+NUM|:ALN2]'
+'perf probe' [options] --line='FUNC[:RLN[+NUM|:RLN2]]|SRC:ALN[+NUM|:ALN2]'
+or
+'perf probe' [options] --vars='PROBEPOINT'
 
 DESCRIPTION
 -----------
@@ -31,6 +33,11 @@ OPTIONS
 --vmlinux=PATH::
         Specify vmlinux path which has debuginfo (Dwarf binary).
 
+-m::
+--module=MODNAME::
+        Specify module name in which perf-probe searches probe points
+        or lines.
+
 -s::
 --source=PATH::
         Specify path to kernel source.
@@ -57,6 +64,15 @@ OPTIONS
         Show source code lines which can be probed. This needs an argument
         which specifies a range of the source code. (see LINE SYNTAX for detail)
 
+-V::
+--vars=::
+        Show available local variables at given probe point. The argument
+        syntax is same as PROBE SYNTAX, but NO ARGs.
+
+--externs::
+        (Only for --vars) Show external defined variables in addition to local
+        variables.
+
 -f::
 --force::
         Forcibly add events with existing name.
diff --git a/tools/perf/Documentation/perf-record.txt b/tools/perf/Documentation/perf-record.txt
index 3ee27dccfde9..a91f9f9e6e5c 100644
--- a/tools/perf/Documentation/perf-record.txt
+++ b/tools/perf/Documentation/perf-record.txt
@@ -83,6 +83,10 @@ OPTIONS
 --call-graph::
         Do call-graph (stack chain/backtrace) recording.
 
+-q::
+--quiet::
+        Don't print any message, useful for scripting.
+
 -v::
 --verbose::
         Be more verbose (show counter open errors, etc).
diff --git a/tools/perf/Documentation/perf-report.txt b/tools/perf/Documentation/perf-report.txt
index abfabe9147a4..12052c9ed0ba 100644
--- a/tools/perf/Documentation/perf-report.txt
+++ b/tools/perf/Documentation/perf-report.txt
@@ -65,6 +65,13 @@ OPTIONS
                  the tree is considered as a new profiled object. +
         Default: fractal,0.5.
 
+--stdio:: Use the stdio interface.
+
+--tui:: Use the TUI interface, that is integrated with annotate and allows
+        zooming into DSOs or threads, among other features. Use of --tui
+        requires a tty, if one is not present, as when piping to other
+        commands, the stdio interface is used.
+
 SEE ALSO
 --------
 linkperf:perf-stat[1]
diff --git a/tools/perf/Makefile b/tools/perf/Makefile
index 1950e19af1cf..d1db0f676a4b 100644
--- a/tools/perf/Makefile
+++ b/tools/perf/Makefile
@@ -313,6 +313,9 @@ TEST_PROGRAMS =
 
 SCRIPT_SH += perf-archive.sh
 
+grep-libs = $(filter -l%,$(1))
+strip-libs = $(filter-out -l%,$(1))
+
 #
 # No Perl scripts right now:
 #
@@ -588,14 +591,17 @@ endif
 ifdef NO_LIBPERL
         BASIC_CFLAGS += -DNO_LIBPERL
 else
-        PERL_EMBED_LDOPTS = `perl -MExtUtils::Embed -e ldopts 2>/dev/null`
+       PERL_EMBED_LDOPTS = $(shell perl -MExtUtils::Embed -e ldopts 2>/dev/null)
+       PERL_EMBED_LDFLAGS = $(call strip-libs,$(PERL_EMBED_LDOPTS))
+       PERL_EMBED_LIBADD = $(call grep-libs,$(PERL_EMBED_LDOPTS))
         PERL_EMBED_CCOPTS = `perl -MExtUtils::Embed -e ccopts 2>/dev/null`
         FLAGS_PERL_EMBED=$(PERL_EMBED_CCOPTS) $(PERL_EMBED_LDOPTS)
 
         ifneq ($(call try-cc,$(SOURCE_PERL_EMBED),$(FLAGS_PERL_EMBED)),y)
                 BASIC_CFLAGS += -DNO_LIBPERL
         else
-                ALL_LDFLAGS += $(PERL_EMBED_LDOPTS)
+               ALL_LDFLAGS += $(PERL_EMBED_LDFLAGS)
+               EXTLIBS += $(PERL_EMBED_LIBADD)
                 LIB_OBJS += $(OUTPUT)util/scripting-engines/trace-event-perl.o
                 LIB_OBJS += $(OUTPUT)scripts/perl/Perf-Trace-Util/Context.o
         endif
@@ -604,13 +610,16 @@ endif
 ifdef NO_LIBPYTHON
         BASIC_CFLAGS += -DNO_LIBPYTHON
 else
-        PYTHON_EMBED_LDOPTS = `python-config --ldflags 2>/dev/null`
+       PYTHON_EMBED_LDOPTS = $(shell python-config --ldflags 2>/dev/null)
+       PYTHON_EMBED_LDFLAGS = $(call strip-libs,$(PYTHON_EMBED_LDOPTS))
+       PYTHON_EMBED_LIBADD = $(call grep-libs,$(PYTHON_EMBED_LDOPTS))
         PYTHON_EMBED_CCOPTS = `python-config --cflags 2>/dev/null`
         FLAGS_PYTHON_EMBED=$(PYTHON_EMBED_CCOPTS) $(PYTHON_EMBED_LDOPTS)
         ifneq ($(call try-cc,$(SOURCE_PYTHON_EMBED),$(FLAGS_PYTHON_EMBED)),y)
                 BASIC_CFLAGS += -DNO_LIBPYTHON
         else
-                ALL_LDFLAGS += $(PYTHON_EMBED_LDOPTS)
+               ALL_LDFLAGS += $(PYTHON_EMBED_LDFLAGS)
+               EXTLIBS += $(PYTHON_EMBED_LIBADD)
                 LIB_OBJS += $(OUTPUT)util/scripting-engines/trace-event-python.o
                 LIB_OBJS += $(OUTPUT)scripts/python/Perf-Trace-Util/Context.o
         endif
@@ -653,6 +662,15 @@ else
         endif
 endif
 
+
+ifdef NO_STRLCPY
+        BASIC_CFLAGS += -DNO_STRLCPY
+else
+        ifneq ($(call try-cc,$(SOURCE_STRLCPY),),y)
+                BASIC_CFLAGS += -DNO_STRLCPY
+        endif
+endif
+
 ifndef CC_LD_DYNPATH
         ifdef NO_R_TO_GCC_LINKER
                 # Some gcc does not accept and pass -R to the linker to specify
@@ -910,8 +928,8 @@ $(OUTPUT)perf.o: perf.c $(OUTPUT)common-cmds.h $(OUTPUT)PERF-CFLAGS
                 $(ALL_CFLAGS) -c $(filter %.c,$^) -o $@
 
 $(OUTPUT)perf$X: $(OUTPUT)perf.o $(BUILTIN_OBJS) $(PERFLIBS)
-        $(QUIET_LINK)$(CC) $(ALL_CFLAGS) -o $@ $(OUTPUT)perf.o \
-                $(BUILTIN_OBJS) $(ALL_LDFLAGS) $(LIBS)
+        $(QUIET_LINK)$(CC) $(ALL_CFLAGS) $(ALL_LDFLAGS) $(OUTPUT)perf.o \
+               $(BUILTIN_OBJS) $(LIBS) -o $@
 
 $(OUTPUT)builtin-help.o: builtin-help.c $(OUTPUT)common-cmds.h $(OUTPUT)PERF-CFLAGS
         $(QUIET_CC)$(CC) -o $@ -c $(ALL_CFLAGS) \
diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index 1478dc64bf15..6d5604d8df95 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -28,7 +28,7 @@
 
 static char             const *input_name = "perf.data";
 
-static bool             force;
+static bool             force, use_tui, use_stdio;
 
 static bool             full_paths;
 
@@ -321,7 +321,7 @@ static int hist_entry__tty_annotate(struct hist_entry *he)
 
 static void hists__find_annotations(struct hists *self)
 {
-        struct rb_node *first = rb_first(&self->entries), *nd = first;
+        struct rb_node *nd = rb_first(&self->entries), *next;
         int key = KEY_RIGHT;
 
         while (nd) {
@@ -343,20 +343,19 @@ find_next:
 
                 if (use_browser > 0) {
                         key = hist_entry__tui_annotate(he);
-                        if (is_exit_key(key))
-                                break;
                         switch (key) {
                         case KEY_RIGHT:
-                        case '\t':
-                                nd = rb_next(nd);
+                                next = rb_next(nd);
                                 break;
                         case KEY_LEFT:
-                                if (nd == first)
-                                        continue;
-                                nd = rb_prev(nd);
-                        default:
+                                next = rb_prev(nd);
                                 break;
+                        default:
+                                return;
                         }
+
+                        if (next != NULL)
+                                nd = next;
                 } else {
                         hist_entry__tty_annotate(he);
                         nd = rb_next(nd);
@@ -428,6 +427,8 @@ static const struct option options[] = {
                     "be more verbose (show symbol address, etc)"),
         OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace,
                     "dump raw trace in ASCII"),
+        OPT_BOOLEAN(0, "tui", &use_tui, "Use the TUI interface"),
+        OPT_BOOLEAN(0, "stdio", &use_stdio, "Use the stdio interface"),
         OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name,
                    "file", "vmlinux pathname"),
         OPT_BOOLEAN('m', "modules", &symbol_conf.use_modules,
@@ -443,6 +444,11 @@ int cmd_annotate(int argc, const char **argv, const char *prefix __used)
 {
         argc = parse_options(argc, argv, options, annotate_usage, 0);
 
+        if (use_stdio)
+                use_browser = 0;
+        else if (use_tui)
+                use_browser = 1;
+
         setup_browser();
 
         symbol_conf.priv_size = sizeof(struct sym_priv);
diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c
index 199d5e19554f..2e000c068cc5 100644
--- a/tools/perf/builtin-probe.c
+++ b/tools/perf/builtin-probe.c
@@ -50,14 +50,17 @@ static struct {
         bool list_events;
         bool force_add;
         bool show_lines;
+        bool show_vars;
+        bool show_ext_vars;
+        bool mod_events;
         int nevents;
         struct perf_probe_event events[MAX_PROBES];
         struct strlist *dellist;
         struct line_range line_range;
+        const char *target_module;
         int max_probe_points;
 } params;
 
-
 /* Parse an event definition. Note that any error must die. */
 static int parse_probe_event(const char *str)
 {
@@ -92,6 +95,7 @@ static int parse_probe_event_argv(int argc, const char **argv)
         len = 0;
         for (i = 0; i < argc; i++)
                 len += sprintf(&buf[len], "%s ", argv[i]);
+        params.mod_events = true;
         ret = parse_probe_event(buf);
         free(buf);
         return ret;
@@ -100,9 +104,10 @@ static int parse_probe_event_argv(int argc, const char **argv)
 static int opt_add_probe_event(const struct option *opt __used,
                               const char *str, int unset __used)
 {
-        if (str)
+        if (str) {
+                params.mod_events = true;
                 return parse_probe_event(str);
-        else
+        } else
                 return 0;
 }
 
@@ -110,6 +115,7 @@ static int opt_del_probe_event(const struct option *opt __used,
                                const char *str, int unset __used)
 {
         if (str) {
+                params.mod_events = true;
                 if (!params.dellist)
                         params.dellist = strlist__new(true, NULL);
                 strlist__add(params.dellist, str);
@@ -130,6 +136,25 @@ static int opt_show_lines(const struct option *opt __used,
 
         return ret;
 }
+
+static int opt_show_vars(const struct option *opt __used,
+                         const char *str, int unset __used)
+{
+        struct perf_probe_event *pev = &params.events[params.nevents];
+        int ret;
+
+        if (!str)
+                return 0;
+
+        ret = parse_probe_event(str);
+        if (!ret && pev->nargs != 0) {
+                pr_err("  Error: '--vars' doesn't accept arguments.\n");
+                return -EINVAL;
+        }
+        params.show_vars = true;
+
+        return ret;
+}
 #endif
 
 static const char * const probe_usage[] = {
@@ -138,7 +163,8 @@ static const char * const probe_usage[] = {
         "perf probe [<options>] --del '[GROUP:]EVENT' ...",
         "perf probe --list",
 #ifdef DWARF_SUPPORT
-        "perf probe --line 'LINEDESC'",
+        "perf probe [<options>] --line 'LINEDESC'",
+        "perf probe [<options>] --vars 'PROBEPOINT'",
 #endif
         NULL
 };
@@ -180,10 +206,17 @@ static const struct option options[] = {
         OPT_CALLBACK('L', "line", NULL,
                      "FUNC[:RLN[+NUM|-RLN2]]|SRC:ALN[+NUM|-ALN2]",
                      "Show source code lines.", opt_show_lines),
+        OPT_CALLBACK('V', "vars", NULL,
+                     "FUNC[@SRC][+OFF|%return|:RL|;PT]|SRC:AL|SRC;PT",
+                     "Show accessible variables on PROBEDEF", opt_show_vars),
+        OPT_BOOLEAN('\0', "externs", &params.show_ext_vars,
+                    "Show external variables too (with --vars only)"),
         OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name,
                    "file", "vmlinux pathname"),
         OPT_STRING('s', "source", &symbol_conf.source_prefix,
                    "directory", "path to kernel source"),
+        OPT_STRING('m', "module", &params.target_module,
+                   "modname", "target module name"),
 #endif
         OPT__DRY_RUN(&probe_event_dry_run),
         OPT_INTEGER('\0', "max-probes", &params.max_probe_points,
@@ -217,7 +250,7 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used)
                 usage_with_options(probe_usage, options);
 
         if (params.list_events) {
-                if (params.nevents != 0 || params.dellist) {
+                if (params.mod_events) {
                         pr_err("  Error: Don't use --list with --add/--del.\n");
                         usage_with_options(probe_usage, options);
                 }
@@ -225,6 +258,10 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used)
                         pr_err("  Error: Don't use --list with --line.\n");
                         usage_with_options(probe_usage, options);
                 }
+                if (params.show_vars) {
+                        pr_err(" Error: Don't use --list with --vars.\n");
+                        usage_with_options(probe_usage, options);
+                }
                 ret = show_perf_probe_events();
                 if (ret < 0)
                         pr_err("  Error: Failed to show event list. (%d)\n",
@@ -234,17 +271,35 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used)
 
 #ifdef DWARF_SUPPORT
         if (params.show_lines) {
-                if (params.nevents != 0 || params.dellist) {
-                        pr_warning("  Error: Don't use --line with"
-                                   " --add/--del.\n");
+                if (params.mod_events) {
+                        pr_err("  Error: Don't use --line with"
+                               " --add/--del.\n");
+                        usage_with_options(probe_usage, options);
+                }
+                if (params.show_vars) {
+                        pr_err(" Error: Don't use --line with --vars.\n");
                         usage_with_options(probe_usage, options);
                 }
 
-                ret = show_line_range(&params.line_range);
+                ret = show_line_range(&params.line_range, params.target_module);
                 if (ret < 0)
                         pr_err("  Error: Failed to show lines. (%d)\n", ret);
                 return ret;
         }
+        if (params.show_vars) {
+                if (params.mod_events) {
+                        pr_err("  Error: Don't use --vars with"
+                               " --add/--del.\n");
+                        usage_with_options(probe_usage, options);
+                }
+                ret = show_available_vars(params.events, params.nevents,
+                                          params.max_probe_points,
+                                          params.target_module,
+                                          params.show_ext_vars);
+                if (ret < 0)
+                        pr_err("  Error: Failed to show vars. (%d)\n", ret);
+                return ret;
+        }
 #endif
 
         if (params.dellist) {
@@ -258,8 +313,9 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used)
 
         if (params.nevents) {
                 ret = add_perf_probe_events(params.events, params.nevents,
-                                            params.force_add,
-                                            params.max_probe_points);
+                                            params.max_probe_points,
+                                            params.target_module,
+                                            params.force_add);
                 if (ret < 0) {
                         pr_err("  Error: Failed to add events. (%d)\n", ret);
                         return ret;
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index ff77b805de71..4e75583ddd6d 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -353,7 +353,7 @@ try_again:
                 }
 
                 if (read(fd[nr_cpu][counter][thread_index], &read_data, sizeof(read_data)) == -1) {
-                        perror("Unable to read perf file descriptor\n");
+                        perror("Unable to read perf file descriptor");
                         exit(-1);
                 }
 
@@ -626,7 +626,7 @@ static int __cmd_record(int argc, const char **argv)
 
         nr_cpus = read_cpu_map(cpu_list);
         if (nr_cpus < 1) {
-                perror("failed to collect number of CPUs\n");
+                perror("failed to collect number of CPUs");
                 return -1;
         }
 
@@ -761,6 +761,9 @@ static int __cmd_record(int argc, const char **argv)
                 }
         }
 
+        if (quiet)
+                return 0;
+
         fprintf(stderr, "[ perf record: Woken up %ld times to write data ]\n", waking);
 
         /*
@@ -820,6 +823,7 @@ static const struct option options[] = {
                     "do call-graph (stack chain/backtrace) recording"),
         OPT_INCR('v', "verbose", &verbose,
                     "be more verbose (show counter open errors, etc)"),
+        OPT_BOOLEAN('q', "quiet", &quiet, "don't print any message"),
         OPT_BOOLEAN('s', "stat", &inherit_stat,
                     "per thread counts"),
         OPT_BOOLEAN('d', "data", &sample_address,
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index 55fc1f46892a..5de405d45230 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -32,7 +32,7 @@
 
 static char             const *input_name = "perf.data";
 
-static bool             force;
+static bool             force, use_tui, use_stdio;
 static bool             hide_unresolved;
 static bool             dont_use_callchains;
 
@@ -107,7 +107,8 @@ static int perf_session__add_hist_entry(struct perf_session *self,
                 goto out_free_syms;
         err = 0;
         if (symbol_conf.use_callchain) {
-                err = append_chain(he->callchain, data->callchain, syms, data->period);
+                err = callchain_append(he->callchain, data->callchain, syms,
+                                       data->period);
                 if (err)
                         goto out_free_syms;
         }
@@ -450,6 +451,8 @@ static const struct option options[] = {
                     "Show per-thread event counters"),
         OPT_STRING(0, "pretty", &pretty_printing_style, "key",
                    "pretty printing style key: normal raw"),
+        OPT_BOOLEAN(0, "tui", &use_tui, "Use the TUI interface"),
+        OPT_BOOLEAN(0, "stdio", &use_stdio, "Use the stdio interface"),
         OPT_STRING('s', "sort", &sort_order, "key[,key2...]",
                    "sort by key(s): pid, comm, dso, symbol, parent"),
         OPT_BOOLEAN(0, "showcpuutilization", &symbol_conf.show_cpu_utilization,
@@ -482,8 +485,15 @@ int cmd_report(int argc, const char **argv, const char *prefix __used)
 {
         argc = parse_options(argc, argv, options, report_usage, 0);
 
+        if (use_stdio)
+                use_browser = 0;
+        else if (use_tui)
+                use_browser = 1;
+
         if (strcmp(input_name, "-") != 0)
                 setup_browser();
+        else
+                use_browser = 0;
         /*
          * Only in the newt browser we are doing integrated annotation,
          * so don't allocate extra space that won't be used in the stdio
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index 40a6a2992d15..2f8df45c4dcb 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -46,9 +46,6 @@ static struct scripting_ops	*scripting_ops;
 
 static void setup_scripting(void)
 {
-        /* make sure PERF_EXEC_PATH is set for scripts */
-        perf_set_argv_exec_path(perf_exec_path());
-
         setup_perl_scripting();
         setup_python_scripting();
 
@@ -285,7 +282,7 @@ static int parse_scriptname(const struct option *opt __used,
                 script++;
         } else {
                 script = str;
-                ext = strchr(script, '.');
+                ext = strrchr(script, '.');
                 if (!ext) {
                         fprintf(stderr, "invalid script extension");
                         return -1;
@@ -593,6 +590,9 @@ int cmd_trace(int argc, const char **argv, const char *prefix __used)
                 suffix = REPORT_SUFFIX;
         }
 
+        /* make sure PERF_EXEC_PATH is set for scripts */
+        perf_set_argv_exec_path(perf_exec_path());
+
         if (!suffix && argc >= 2 && strncmp(argv[1], "-", strlen("-")) != 0) {
                 char *record_script_path, *report_script_path;
                 int live_pipe[2];
@@ -625,12 +625,13 @@ int cmd_trace(int argc, const char **argv, const char *prefix __used)
                         dup2(live_pipe[1], 1);
                         close(live_pipe[0]);
 
-                        __argv = malloc(5 * sizeof(const char *));
+                        __argv = malloc(6 * sizeof(const char *));
                         __argv[0] = "/bin/sh";
                         __argv[1] = record_script_path;
-                        __argv[2] = "-o";
-                        __argv[3] = "-";
-                        __argv[4] = NULL;
+                        __argv[2] = "-q";
+                        __argv[3] = "-o";
+                        __argv[4] = "-";
+                        __argv[5] = NULL;
 
                         execvp("/bin/sh", (char **)__argv);
                         exit(-1);
diff --git a/tools/perf/feature-tests.mak b/tools/perf/feature-tests.mak
index 7a7b60859053..b253db634f04 100644
--- a/tools/perf/feature-tests.mak
+++ b/tools/perf/feature-tests.mak
@@ -110,6 +110,17 @@ int main(void)
 }
 endef
 
+define SOURCE_STRLCPY
+#include <stdlib.h>
+extern size_t strlcpy(char *dest, const char *src, size_t size);
+
+int main(void)
+{
+        strlcpy(NULL, NULL, 0);
+        return 0;
+}
+endef
+
 # try-cc
 # Usage: option = $(call try-cc, source-to-build, cc-options)
 try-cc = $(shell sh -c                                            \
diff --git a/tools/perf/scripts/perl/bin/failed-syscalls-report b/tools/perf/scripts/perl/bin/failed-syscalls-report
index e3a5e55d54ff..4028d92dc4ae 100644
--- a/tools/perf/scripts/perl/bin/failed-syscalls-report
+++ b/tools/perf/scripts/perl/bin/failed-syscalls-report
@@ -7,4 +7,4 @@ if [ $# -gt 0 ] ; then
         shift
     fi
 fi
-perf trace $@ -s ~/libexec/perf-core/scripts/perl/failed-syscalls.pl $comm
+perf trace $@ -s "$PERF_EXEC_PATH"/scripts/perl/failed-syscalls.pl $comm
diff --git a/tools/perf/scripts/perl/bin/rw-by-file-report b/tools/perf/scripts/perl/bin/rw-by-file-report
index d83070b7eeb5..ba25f4d41fb0 100644
--- a/tools/perf/scripts/perl/bin/rw-by-file-report
+++ b/tools/perf/scripts/perl/bin/rw-by-file-report
@@ -7,7 +7,7 @@ if [ $# -lt 1 ] ; then
 fi
 comm=$1
 shift
-perf trace $@ -s ~/libexec/perf-core/scripts/perl/rw-by-file.pl $comm
+perf trace $@ -s "$PERF_EXEC_PATH"/scripts/perl/rw-by-file.pl $comm
 
 
 
diff --git a/tools/perf/scripts/perl/bin/rw-by-pid-report b/tools/perf/scripts/perl/bin/rw-by-pid-report
index 7ef46983f62f..641a3f5d085c 100644
--- a/tools/perf/scripts/perl/bin/rw-by-pid-report
+++ b/tools/perf/scripts/perl/bin/rw-by-pid-report
@@ -1,6 +1,6 @@
 #!/bin/bash
 # description: system-wide r/w activity
-perf trace $@ -s ~/libexec/perf-core/scripts/perl/rw-by-pid.pl
+perf trace $@ -s "$PERF_EXEC_PATH"/scripts/perl/rw-by-pid.pl
 
 
 
diff --git a/tools/perf/scripts/perl/bin/rwtop-report b/tools/perf/scripts/perl/bin/rwtop-report
index 93e698cd3f38..4918dba77021 100644
--- a/tools/perf/scripts/perl/bin/rwtop-report
+++ b/tools/perf/scripts/perl/bin/rwtop-report
@@ -17,7 +17,7 @@ if [ "$n_args" -gt 0 ] ; then
     interval=$1
     shift
 fi
-perf trace $@ -s ~/libexec/perf-core/scripts/perl/rwtop.pl $interval
+perf trace $@ -s "$PERF_EXEC_PATH"/scripts/perl/rwtop.pl $interval
 
 
 
diff --git a/tools/perf/scripts/perl/bin/wakeup-latency-report b/tools/perf/scripts/perl/bin/wakeup-latency-report
index a0d898f9ca1d..49052ebcb632 100644
--- a/tools/perf/scripts/perl/bin/wakeup-latency-report
+++ b/tools/perf/scripts/perl/bin/wakeup-latency-report
@@ -1,6 +1,6 @@
 #!/bin/bash
 # description: system-wide min/max/avg wakeup latency
-perf trace $@ -s ~/libexec/perf-core/scripts/perl/wakeup-latency.pl
+perf trace $@ -s "$PERF_EXEC_PATH"/scripts/perl/wakeup-latency.pl
 
 
 
diff --git a/tools/perf/scripts/perl/bin/workqueue-stats-report b/tools/perf/scripts/perl/bin/workqueue-stats-report
index 35081132ef97..df0c65f4ca93 100644
--- a/tools/perf/scripts/perl/bin/workqueue-stats-report
+++ b/tools/perf/scripts/perl/bin/workqueue-stats-report
@@ -1,6 +1,6 @@
 #!/bin/bash
 # description: workqueue stats (ins/exe/create/destroy)
-perf trace $@ -s ~/libexec/perf-core/scripts/perl/workqueue-stats.pl
+perf trace $@ -s "$PERF_EXEC_PATH"/scripts/perl/workqueue-stats.pl
 
 
 
diff --git a/tools/perf/scripts/python/Perf-Trace-Util/lib/Perf/Trace/Util.py b/tools/perf/scripts/python/Perf-Trace-Util/lib/Perf/Trace/Util.py
index 9689bc0acd9f..13cc02b5893a 100644
--- a/tools/perf/scripts/python/Perf-Trace-Util/lib/Perf/Trace/Util.py
+++ b/tools/perf/scripts/python/Perf-Trace-Util/lib/Perf/Trace/Util.py
@@ -6,6 +6,14 @@
 # Public License ("GPL") version 2 as published by the Free Software
 # Foundation.
 
+import errno, os
+
+FUTEX_WAIT = 0
+FUTEX_WAKE = 1
+FUTEX_PRIVATE_FLAG = 128
+FUTEX_CLOCK_REALTIME = 256
+FUTEX_CMD_MASK = ~(FUTEX_PRIVATE_FLAG | FUTEX_CLOCK_REALTIME)
+
 NSECS_PER_SEC    = 1000000000
 
 def avg(total, n):
@@ -24,5 +32,55 @@ def nsecs_str(nsecs):
     str = "%5u.%09u" % (nsecs_secs(nsecs), nsecs_nsecs(nsecs)),
     return str
 
+def add_stats(dict, key, value):
+        if not dict.has_key(key):
+                dict[key] = (value, value, value, 1)
+        else:
+                min, max, avg, count = dict[key]
+                if value < min:
+                        min = value
+                if value > max:
+                        max = value
+                avg = (avg + value) / 2
+                dict[key] = (min, max, avg, count + 1)
+
 def clear_term():
     print("\x1b[H\x1b[2J")
+
+audit_package_warned = False
+
+try:
+        import audit
+        machine_to_id = {
+                'x86_64': audit.MACH_86_64,
+                'alpha' : audit.MACH_ALPHA,
+                'ia64'  : audit.MACH_IA64,
+                'ppc'   : audit.MACH_PPC,
+                'ppc64' : audit.MACH_PPC64,
+                's390'  : audit.MACH_S390,
+                's390x' : audit.MACH_S390X,
+                'i386'  : audit.MACH_X86,
+                'i586'  : audit.MACH_X86,
+                'i686'  : audit.MACH_X86,
+        }
+        try:
+                machine_to_id['armeb'] = audit.MACH_ARMEB
+        except:
+                pass
+        machine_id = machine_to_id[os.uname()[4]]
+except:
+        if not audit_package_warned:
+                audit_package_warned = True
+                print "Install the audit-libs-python package to get syscall names"
+
+def syscall_name(id):
+        try:
+                return audit.audit_syscall_to_name(id, machine_id)
+        except:
+                return str(id)
+
+def strerror(nr):
+        try:
+                return errno.errorcode[abs(nr)]
+        except:
+                return "Unknown %d errno" % nr
diff --git a/tools/perf/scripts/python/bin/failed-syscalls-by-pid-report b/tools/perf/scripts/python/bin/failed-syscalls-by-pid-report
index 30293545fcc2..03587021463d 100644
--- a/tools/perf/scripts/python/bin/failed-syscalls-by-pid-report
+++ b/tools/perf/scripts/python/bin/failed-syscalls-by-pid-report
@@ -7,4 +7,4 @@ if [ $# -gt 0 ] ; then
         shift
     fi
 fi
-perf trace $@ -s ~/libexec/perf-core/scripts/python/failed-syscalls-by-pid.py $comm
+perf trace $@ -s "$PERF_EXEC_PATH"/scripts/python/failed-syscalls-by-pid.py $comm
diff --git a/tools/perf/scripts/python/bin/futex-contention-record b/tools/perf/scripts/python/bin/futex-contention-record
new file mode 100644
index 000000000000..5ecbb433caf4
--- /dev/null
+++ b/tools/perf/scripts/python/bin/futex-contention-record
@@ -0,0 +1,2 @@
+#!/bin/bash
+perf record -a -e syscalls:sys_enter_futex -e syscalls:sys_exit_futex $@
diff --git a/tools/perf/scripts/python/bin/futex-contention-report b/tools/perf/scripts/python/bin/futex-contention-report
new file mode 100644
index 000000000000..c8268138fb7e
--- /dev/null
+++ b/tools/perf/scripts/python/bin/futex-contention-report
@@ -0,0 +1,4 @@
+#!/bin/bash
+# description: futext contention measurement
+
+perf trace $@ -s "$PERF_EXEC_PATH"/scripts/python/futex-contention.py
diff --git a/tools/perf/scripts/python/bin/netdev-times-record b/tools/perf/scripts/python/bin/netdev-times-record
new file mode 100644
index 000000000000..d931a828126b
--- /dev/null
+++ b/tools/perf/scripts/python/bin/netdev-times-record
@@ -0,0 +1,8 @@
+#!/bin/bash
+perf record -a -e net:net_dev_xmit -e net:net_dev_queue         \
+                -e net:netif_receive_skb -e net:netif_rx                \
+                -e skb:consume_skb -e skb:kfree_skb                     \
+                -e skb:skb_copy_datagram_iovec -e napi:napi_poll        \
+                -e irq:irq_handler_entry -e irq:irq_handler_exit        \
+                -e irq:softirq_entry -e irq:softirq_exit                \
+                -e irq:softirq_raise $@
diff --git a/tools/perf/scripts/python/bin/netdev-times-report b/tools/perf/scripts/python/bin/netdev-times-report
new file mode 100644
index 000000000000..4ad361b31249
--- /dev/null
+++ b/tools/perf/scripts/python/bin/netdev-times-report
@@ -0,0 +1,5 @@
+#!/bin/bash
+# description: display a process of packet and processing time
+# args: [tx] [rx] [dev=] [debug]
+
+perf trace -s "$PERF_EXEC_PATH"/scripts/python/netdev-times.py $@
diff --git a/tools/perf/scripts/python/bin/sched-migration-report b/tools/perf/scripts/python/bin/sched-migration-report
index 61d05f72e443..df1791f07c24 100644
--- a/tools/perf/scripts/python/bin/sched-migration-report
+++ b/tools/perf/scripts/python/bin/sched-migration-report
@@ -1,3 +1,3 @@
 #!/bin/bash
 # description: sched migration overview
-perf trace $@ -s ~/libexec/perf-core/scripts/python/sched-migration.py
+perf trace $@ -s "$PERF_EXEC_PATH"/scripts/python/sched-migration.py
diff --git a/tools/perf/scripts/python/bin/sctop-report b/tools/perf/scripts/python/bin/sctop-report
index b01c842ae7b4..36b409c05e50 100644
--- a/tools/perf/scripts/python/bin/sctop-report
+++ b/tools/perf/scripts/python/bin/sctop-report
@@ -21,4 +21,4 @@ elif [ "$n_args" -gt 0 ] ; then
     interval=$1
     shift
 fi
-perf trace $@ -s ~/libexec/perf-core/scripts/python/sctop.py $comm $interval
+perf trace $@ -s "$PERF_EXEC_PATH"/scripts/python/sctop.py $comm $interval
diff --git a/tools/perf/scripts/python/bin/syscall-counts-by-pid-report b/tools/perf/scripts/python/bin/syscall-counts-by-pid-report
index 9e9d8ddd72ce..4eb88c9fc83c 100644
--- a/tools/perf/scripts/python/bin/syscall-counts-by-pid-report
+++ b/tools/perf/scripts/python/bin/syscall-counts-by-pid-report
@@ -7,4 +7,4 @@ if [ $# -gt 0 ] ; then
         shift
     fi
 fi
-perf trace $@ -s ~/libexec/perf-core/scripts/python/syscall-counts-by-pid.py $comm
+perf trace $@ -s "$PERF_EXEC_PATH"/scripts/python/syscall-counts-by-pid.py $comm
diff --git a/tools/perf/scripts/python/bin/syscall-counts-report b/tools/perf/scripts/python/bin/syscall-counts-report
index dc076b618796..cb2f9c5cf17e 100644
--- a/tools/perf/scripts/python/bin/syscall-counts-report
+++ b/tools/perf/scripts/python/bin/syscall-counts-report
@@ -7,4 +7,4 @@ if [ $# -gt 0 ] ; then
         shift
     fi
 fi
-perf trace $@ -s ~/libexec/perf-core/scripts/python/syscall-counts.py $comm
+perf trace $@ -s "$PERF_EXEC_PATH"/scripts/python/syscall-counts.py $comm
diff --git a/tools/perf/scripts/python/failed-syscalls-by-pid.py b/tools/perf/scripts/python/failed-syscalls-by-pid.py
index 0ca02278fe69..acd7848717b3 100644
--- a/tools/perf/scripts/python/failed-syscalls-by-pid.py
+++ b/tools/perf/scripts/python/failed-syscalls-by-pid.py
@@ -13,21 +13,26 @@ sys.path.append(os.environ['PERF_EXEC_PATH'] + \
 
 from perf_trace_context import *
 from Core import *
+from Util import *
 
-usage = "perf trace -s syscall-counts-by-pid.py [comm]\n";
+usage = "perf trace -s syscall-counts-by-pid.py [comm|pid]\n";
 
 for_comm = None
+for_pid = None
 
 if len(sys.argv) > 2:
         sys.exit(usage)
 
 if len(sys.argv) > 1:
-        for_comm = sys.argv[1]
+        try:
+                for_pid = int(sys.argv[1])
+        except:
+                for_comm = sys.argv[1]
 
 syscalls = autodict()
 
 def trace_begin():
-        pass
+        print "Press control+C to stop and show the summary"
 
 def trace_end():
         print_error_totals()
@@ -35,9 +40,9 @@ def trace_end():
 def raw_syscalls__sys_exit(event_name, context, common_cpu,
         common_secs, common_nsecs, common_pid, common_comm,
         id, ret):
-        if for_comm is not None:
-                if common_comm != for_comm:
-                        return
+        if (for_comm and common_comm != for_comm) or \
+           (for_pid  and common_pid  != for_pid ):
+                return
 
         if ret < 0:
                 try:
@@ -62,7 +67,7 @@ def print_error_totals():
                     print "\n%s [%d]\n" % (comm, pid),
                     id_keys = syscalls[comm][pid].keys()
                     for id in id_keys:
-                            print "  syscall: %-16d\n" % (id),
+                            print "  syscall: %-16s\n" % syscall_name(id),
                             ret_keys = syscalls[comm][pid][id].keys()
                             for ret, val in sorted(syscalls[comm][pid][id].iteritems(), key = lambda(k, v): (v, k),  reverse = True):
-                                    print "    err = %-20d  %10d\n" % (ret, val),
+                                    print "    err = %-20s  %10d\n" % (strerror(ret), val),
diff --git a/tools/perf/scripts/python/futex-contention.py b/tools/perf/scripts/python/futex-contention.py
new file mode 100644
index 000000000000..11e70a388d41
--- /dev/null
+++ b/tools/perf/scripts/python/futex-contention.py
@@ -0,0 +1,50 @@
+# futex contention
+# (c) 2010, Arnaldo Carvalho de Melo <acme@redhat.com>
+# Licensed under the terms of the GNU GPL License version 2
+#
+# Translation of:
+#
+# http://sourceware.org/systemtap/wiki/WSFutexContention
+#
+# to perf python scripting.
+#
+# Measures futex contention
+
+import os, sys
+sys.path.append(os.environ['PERF_EXEC_PATH'] + '/scripts/python/Perf-Trace-Util/lib/Perf/Trace')
+from Util import *
+
+process_names = {}
+thread_thislock = {}
+thread_blocktime = {}
+
+lock_waits = {} # long-lived stats on (tid,lock) blockage elapsed time
+process_names = {} # long-lived pid-to-execname mapping
+
+def syscalls__sys_enter_futex(event, ctxt, cpu, s, ns, tid, comm,
+                              nr, uaddr, op, val, utime, uaddr2, val3):
+        cmd = op & FUTEX_CMD_MASK
+        if cmd != FUTEX_WAIT:
+                return # we don't care about originators of WAKE events
+
+        process_names[tid] = comm
+        thread_thislock[tid] = uaddr
+        thread_blocktime[tid] = nsecs(s, ns)
+
+def syscalls__sys_exit_futex(event, ctxt, cpu, s, ns, tid, comm,
+                             nr, ret):
+        if thread_blocktime.has_key(tid):
+                elapsed = nsecs(s, ns) - thread_blocktime[tid]
+                add_stats(lock_waits, (tid, thread_thislock[tid]), elapsed)
+                del thread_blocktime[tid]
+                del thread_thislock[tid]
+
+def trace_begin():
+        print "Press control+C to stop and show the summary"
+
+def trace_end():
+        for (tid, lock) in lock_waits:
+                min, max, avg, count = lock_waits[tid, lock]
+                print "%s[%d] lock %x contended %d times, %d avg ns" % \
+                      (process_names[tid], tid, lock, count, avg)
+
diff --git a/tools/perf/scripts/python/netdev-times.py b/tools/perf/scripts/python/netdev-times.py
new file mode 100644
index 000000000000..9aa0a32972e8
--- /dev/null
+++ b/tools/perf/scripts/python/netdev-times.py
@@ -0,0 +1,464 @@
+# Display a process of packets and processed time.
+# It helps us to investigate networking or network device.
+#
+# options
+# tx: show only tx chart
+# rx: show only rx chart
+# dev=: show only thing related to specified device
+# debug: work with debug mode. It shows buffer status.
+
+import os
+import sys
+
+sys.path.append(os.environ['PERF_EXEC_PATH'] + \
+        '/scripts/python/Perf-Trace-Util/lib/Perf/Trace')
+
+from perf_trace_context import *
+from Core import *
+from Util import *
+
+all_event_list = []; # insert all tracepoint event related with this script
+irq_dic = {}; # key is cpu and value is a list which stacks irqs
+              # which raise NET_RX softirq
+net_rx_dic = {}; # key is cpu and value include time of NET_RX softirq-entry
+                 # and a list which stacks receive
+receive_hunk_list = []; # a list which include a sequence of receive events
+rx_skb_list = []; # received packet list for matching
+                       # skb_copy_datagram_iovec
+
+buffer_budget = 65536; # the budget of rx_skb_list, tx_queue_list and
+                       # tx_xmit_list
+of_count_rx_skb_list = 0; # overflow count
+
+tx_queue_list = []; # list of packets which pass through dev_queue_xmit
+of_count_tx_queue_list = 0; # overflow count
+
+tx_xmit_list = [];  # list of packets which pass through dev_hard_start_xmit
+of_count_tx_xmit_list = 0; # overflow count
+
+tx_free_list = [];  # list of packets which is freed
+
+# options
+show_tx = 0;
+show_rx = 0;
+dev = 0; # store a name of device specified by option "dev="
+debug = 0;
+
+# indices of event_info tuple
+EINFO_IDX_NAME=   0
+EINFO_IDX_CONTEXT=1
+EINFO_IDX_CPU=    2
+EINFO_IDX_TIME=   3
+EINFO_IDX_PID=    4
+EINFO_IDX_COMM=   5
+
+# Calculate a time interval(msec) from src(nsec) to dst(nsec)
+def diff_msec(src, dst):
+        return (dst - src) / 1000000.0
+
+# Display a process of transmitting a packet
+def print_transmit(hunk):
+        if dev != 0 and hunk['dev'].find(dev) < 0:
+                return
+        print "%7s %5d %6d.%06dsec %12.3fmsec      %12.3fmsec" % \
+                (hunk['dev'], hunk['len'],
+                nsecs_secs(hunk['queue_t']),
+                nsecs_nsecs(hunk['queue_t'])/1000,
+                diff_msec(hunk['queue_t'], hunk['xmit_t']),
+                diff_msec(hunk['xmit_t'], hunk['free_t']))
+
+# Format for displaying rx packet processing
+PF_IRQ_ENTRY= "  irq_entry(+%.3fmsec irq=%d:%s)"
+PF_SOFT_ENTRY="  softirq_entry(+%.3fmsec)"
+PF_NAPI_POLL= "  napi_poll_exit(+%.3fmsec %s)"
+PF_JOINT=     "         |"
+PF_WJOINT=    "         |            |"
+PF_NET_RECV=  "         |---netif_receive_skb(+%.3fmsec skb=%x len=%d)"
+PF_NET_RX=    "         |---netif_rx(+%.3fmsec skb=%x)"
+PF_CPY_DGRAM= "         |      skb_copy_datagram_iovec(+%.3fmsec %d:%s)"
+PF_KFREE_SKB= "         |      kfree_skb(+%.3fmsec location=%x)"
+PF_CONS_SKB=  "         |      consume_skb(+%.3fmsec)"
+
+# Display a process of received packets and interrputs associated with
+# a NET_RX softirq
+def print_receive(hunk):
+        show_hunk = 0
+        irq_list = hunk['irq_list']
+        cpu = irq_list[0]['cpu']
+        base_t = irq_list[0]['irq_ent_t']
+        # check if this hunk should be showed
+        if dev != 0:
+                for i in range(len(irq_list)):
+                        if irq_list[i]['name'].find(dev) >= 0:
+                                show_hunk = 1
+                                break
+        else:
+                show_hunk = 1
+        if show_hunk == 0:
+                return
+
+        print "%d.%06dsec cpu=%d" % \
+                (nsecs_secs(base_t), nsecs_nsecs(base_t)/1000, cpu)
+        for i in range(len(irq_list)):
+                print PF_IRQ_ENTRY % \
+                        (diff_msec(base_t, irq_list[i]['irq_ent_t']),
+                        irq_list[i]['irq'], irq_list[i]['name'])
+                print PF_JOINT
+                irq_event_list = irq_list[i]['event_list']
+                for j in range(len(irq_event_list)):
+                        irq_event = irq_event_list[j]
+                        if irq_event['event'] == 'netif_rx':
+                                print PF_NET_RX % \
+                                        (diff_msec(base_t, irq_event['time']),
+                                        irq_event['skbaddr'])
+                                print PF_JOINT
+        print PF_SOFT_ENTRY % \
+                diff_msec(base_t, hunk['sirq_ent_t'])
+        print PF_JOINT
+        event_list = hunk['event_list']
+        for i in range(len(event_list)):
+                event = event_list[i]
+                if event['event_name'] == 'napi_poll':
+                        print PF_NAPI_POLL % \
+                            (diff_msec(base_t, event['event_t']), event['dev'])
+                        if i == len(event_list) - 1:
+                                print ""
+                        else:
+                                print PF_JOINT
+                else:
+                        print PF_NET_RECV % \
+                            (diff_msec(base_t, event['event_t']), event['skbaddr'],
+                                event['len'])
+                        if 'comm' in event.keys():
+                                print PF_WJOINT
+                                print PF_CPY_DGRAM % \
+                                        (diff_msec(base_t, event['comm_t']),
+                                        event['pid'], event['comm'])
+                        elif 'handle' in event.keys():
+                                print PF_WJOINT
+                                if event['handle'] == "kfree_skb":
+                                        print PF_KFREE_SKB % \
+                                                (diff_msec(base_t,
+                                                event['comm_t']),
+                                                event['location'])
+                                elif event['handle'] == "consume_skb":
+                                        print PF_CONS_SKB % \
+                                                diff_msec(base_t,
+                                                        event['comm_t'])
+                        print PF_JOINT
+
+def trace_begin():
+        global show_tx
+        global show_rx
+        global dev
+        global debug
+
+        for i in range(len(sys.argv)):
+                if i == 0:
+                        continue
+                arg = sys.argv[i]
+                if arg == 'tx':
+                        show_tx = 1
+                elif arg =='rx':
+                        show_rx = 1
+                elif arg.find('dev=',0, 4) >= 0:
+                        dev = arg[4:]
+                elif arg == 'debug':
+                        debug = 1
+        if show_tx == 0  and show_rx == 0:
+                show_tx = 1
+                show_rx = 1
+
+def trace_end():
+        # order all events in time
+        all_event_list.sort(lambda a,b :cmp(a[EINFO_IDX_TIME],
+                                            b[EINFO_IDX_TIME]))
+        # process all events
+        for i in range(len(all_event_list)):
+                event_info = all_event_list[i]
+                name = event_info[EINFO_IDX_NAME]
+                if name == 'irq__softirq_exit':
+                        handle_irq_softirq_exit(event_info)
+                elif name == 'irq__softirq_entry':
+                        handle_irq_softirq_entry(event_info)
+                elif name == 'irq__softirq_raise':
+                        handle_irq_softirq_raise(event_info)
+                elif name == 'irq__irq_handler_entry':
+                        handle_irq_handler_entry(event_info)
+                elif name == 'irq__irq_handler_exit':
+                        handle_irq_handler_exit(event_info)
+                elif name == 'napi__napi_poll':
+                        handle_napi_poll(event_info)
+                elif name == 'net__netif_receive_skb':
+                        handle_netif_receive_skb(event_info)
+                elif name == 'net__netif_rx':
+                        handle_netif_rx(event_info)
+                elif name == 'skb__skb_copy_datagram_iovec':
+                        handle_skb_copy_datagram_iovec(event_info)
+                elif name == 'net__net_dev_queue':
+                        handle_net_dev_queue(event_info)
+                elif name == 'net__net_dev_xmit':
+                        handle_net_dev_xmit(event_info)
+                elif name == 'skb__kfree_skb':
+                        handle_kfree_skb(event_info)
+                elif name == 'skb__consume_skb':
+                        handle_consume_skb(event_info)
+        # display receive hunks
+        if show_rx:
+                for i in range(len(receive_hunk_list)):
+                        print_receive(receive_hunk_list[i])
+        # display transmit hunks
+        if show_tx:
+                print "   dev    len      Qdisc        " \
+                        "       netdevice             free"
+                for i in range(len(tx_free_list)):
+                        print_transmit(tx_free_list[i])
+        if debug:
+                print "debug buffer status"
+                print "----------------------------"
+                print "xmit Qdisc:remain:%d overflow:%d" % \
+                        (len(tx_queue_list), of_count_tx_queue_list)
+                print "xmit netdevice:remain:%d overflow:%d" % \
+                        (len(tx_xmit_list), of_count_tx_xmit_list)
+                print "receive:remain:%d overflow:%d" % \
+                        (len(rx_skb_list), of_count_rx_skb_list)
+
+# called from perf, when it finds a correspoinding event
+def irq__softirq_entry(name, context, cpu, sec, nsec, pid, comm, vec):
+        if symbol_str("irq__softirq_entry", "vec", vec) != "NET_RX":
+                return
+        event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, vec)
+        all_event_list.append(event_info)
+
+def irq__softirq_exit(name, context, cpu, sec, nsec, pid, comm, vec):
+        if symbol_str("irq__softirq_entry", "vec", vec) != "NET_RX":
+                return
+        event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, vec)
+        all_event_list.append(event_info)
+
+def irq__softirq_raise(name, context, cpu, sec, nsec, pid, comm, vec):
+        if symbol_str("irq__softirq_entry", "vec", vec) != "NET_RX":
+                return
+        event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, vec)
+        all_event_list.append(event_info)
+
+def irq__irq_handler_entry(name, context, cpu, sec, nsec, pid, comm,
+                        irq, irq_name):
+        event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm,
+                        irq, irq_name)
+        all_event_list.append(event_info)
+
+def irq__irq_handler_exit(name, context, cpu, sec, nsec, pid, comm, irq, ret):
+        event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, irq, ret)
+        all_event_list.append(event_info)
+
+def napi__napi_poll(name, context, cpu, sec, nsec, pid, comm, napi, dev_name):
+        event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm,
+                        napi, dev_name)
+        all_event_list.append(event_info)
+
+def net__netif_receive_skb(name, context, cpu, sec, nsec, pid, comm, skbaddr,
+                        skblen, dev_name):
+        event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm,
+                        skbaddr, skblen, dev_name)
+        all_event_list.append(event_info)
+
+def net__netif_rx(name, context, cpu, sec, nsec, pid, comm, skbaddr,
+                        skblen, dev_name):
+        event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm,
+                        skbaddr, skblen, dev_name)
+        all_event_list.append(event_info)
+
+def net__net_dev_queue(name, context, cpu, sec, nsec, pid, comm,
+                        skbaddr, skblen, dev_name):
+        event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm,
+                        skbaddr, skblen, dev_name)
+        all_event_list.append(event_info)
+
+def net__net_dev_xmit(name, context, cpu, sec, nsec, pid, comm,
+                        skbaddr, skblen, rc, dev_name):
+        event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm,
+                        skbaddr, skblen, rc ,dev_name)
+        all_event_list.append(event_info)
+
+def skb__kfree_skb(name, context, cpu, sec, nsec, pid, comm,
+                        skbaddr, protocol, location):
+        event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm,
+                        skbaddr, protocol, location)
+        all_event_list.append(event_info)
+
+def skb__consume_skb(name, context, cpu, sec, nsec, pid, comm, skbaddr):
+        event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm,
+                        skbaddr)
+        all_event_list.append(event_info)
+
+def skb__skb_copy_datagram_iovec(name, context, cpu, sec, nsec, pid, comm,
+        skbaddr, skblen):
+        event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm,
+                        skbaddr, skblen)
+        all_event_list.append(event_info)
+
+def handle_irq_handler_entry(event_info):
+        (name, context, cpu, time, pid, comm, irq, irq_name) = event_info
+        if cpu not in irq_dic.keys():
+                irq_dic[cpu] = []
+        irq_record = {'irq':irq, 'name':irq_name, 'cpu':cpu, 'irq_ent_t':time}
+        irq_dic[cpu].append(irq_record)
+
+def handle_irq_handler_exit(event_info):
+        (name, context, cpu, time, pid, comm, irq, ret) = event_info
+        if cpu not in irq_dic.keys():
+                return
+        irq_record = irq_dic[cpu].pop()
+        if irq != irq_record['irq']:
+                return
+        irq_record.update({'irq_ext_t':time})
+        # if an irq doesn't include NET_RX softirq, drop.
+        if 'event_list' in irq_record.keys():
+                irq_dic[cpu].append(irq_record)
+
+def handle_irq_softirq_raise(event_info):
+        (name, context, cpu, time, pid, comm, vec) = event_info
+        if cpu not in irq_dic.keys() \
+        or len(irq_dic[cpu]) == 0:
+                return
+        irq_record = irq_dic[cpu].pop()
+        if 'event_list' in irq_record.keys():
+                irq_event_list = irq_record['event_list']
+        else:
+                irq_event_list = []
+        irq_event_list.append({'time':time, 'event':'sirq_raise'})
+        irq_record.update({'event_list':irq_event_list})
+        irq_dic[cpu].append(irq_record)
+
+def handle_irq_softirq_entry(event_info):
+        (name, context, cpu, time, pid, comm, vec) = event_info
+        net_rx_dic[cpu] = {'sirq_ent_t':time, 'event_list':[]}
+
+def handle_irq_softirq_exit(event_info):
+        (name, context, cpu, time, pid, comm, vec) = event_info
+        irq_list = []
+        event_list = 0
+        if cpu in irq_dic.keys():
+                irq_list = irq_dic[cpu]
+                del irq_dic[cpu]
+        if cpu in net_rx_dic.keys():
+                sirq_ent_t = net_rx_dic[cpu]['sirq_ent_t']
+                event_list = net_rx_dic[cpu]['event_list']
+                del net_rx_dic[cpu]
+        if irq_list == [] or event_list == 0:
+                return
+        rec_data = {'sirq_ent_t':sirq_ent_t, 'sirq_ext_t':time,
+                    'irq_list':irq_list, 'event_list':event_list}
+        # merge information realted to a NET_RX softirq
+        receive_hunk_list.append(rec_data)
+
+def handle_napi_poll(event_info):
+        (name, context, cpu, time, pid, comm, napi, dev_name) = event_info
+        if cpu in net_rx_dic.keys():
+                event_list = net_rx_dic[cpu]['event_list']
+                rec_data = {'event_name':'napi_poll',
+                                'dev':dev_name, 'event_t':time}
+                event_list.append(rec_data)
+
+def handle_netif_rx(event_info):
+        (name, context, cpu, time, pid, comm,
+                skbaddr, skblen, dev_name) = event_info
+        if cpu not in irq_dic.keys() \
+        or len(irq_dic[cpu]) == 0:
+                return
+        irq_record = irq_dic[cpu].pop()
+        if 'event_list' in irq_record.keys():
+                irq_event_list = irq_record['event_list']
+        else:
+                irq_event_list = []
+        irq_event_list.append({'time':time, 'event':'netif_rx',
+                'skbaddr':skbaddr, 'skblen':skblen, 'dev_name':dev_name})
+        irq_record.update({'event_list':irq_event_list})
+        irq_dic[cpu].append(irq_record)
+
+def handle_netif_receive_skb(event_info):
+        global of_count_rx_skb_list
+
+        (name, context, cpu, time, pid, comm,
+                skbaddr, skblen, dev_name) = event_info
+        if cpu in net_rx_dic.keys():
+                rec_data = {'event_name':'netif_receive_skb',
+                            'event_t':time, 'skbaddr':skbaddr, 'len':skblen}
+                event_list = net_rx_dic[cpu]['event_list']
+                event_list.append(rec_data)
+                rx_skb_list.insert(0, rec_data)
+                if len(rx_skb_list) > buffer_budget:
+                        rx_skb_list.pop()
+                        of_count_rx_skb_list += 1
+
+def handle_net_dev_queue(event_info):
+        global of_count_tx_queue_list
+
+        (name, context, cpu, time, pid, comm,
+                skbaddr, skblen, dev_name) = event_info
+        skb = {'dev':dev_name, 'skbaddr':skbaddr, 'len':skblen, 'queue_t':time}
+        tx_queue_list.insert(0, skb)
+        if len(tx_queue_list) > buffer_budget:
+                tx_queue_list.pop()
+                of_count_tx_queue_list += 1
+
+def handle_net_dev_xmit(event_info):
+        global of_count_tx_xmit_list
+
+        (name, context, cpu, time, pid, comm,
+                skbaddr, skblen, rc, dev_name) = event_info
+        if rc == 0: # NETDEV_TX_OK
+                for i in range(len(tx_queue_list)):
+                        skb = tx_queue_list[i]
+                        if skb['skbaddr'] == skbaddr:
+                                skb['xmit_t'] = time
+                                tx_xmit_list.insert(0, skb)
+                                del tx_queue_list[i]
+                                if len(tx_xmit_list) > buffer_budget:
+                                        tx_xmit_list.pop()
+                                        of_count_tx_xmit_list += 1
+                                return
+
+def handle_kfree_skb(event_info):
+        (name, context, cpu, time, pid, comm,
+                skbaddr, protocol, location) = event_info
+        for i in range(len(tx_queue_list)):
+                skb = tx_queue_list[i]
+                if skb['skbaddr'] == skbaddr:
+                        del tx_queue_list[i]
+                        return
+        for i in range(len(tx_xmit_list)):
+                skb = tx_xmit_list[i]
+                if skb['skbaddr'] == skbaddr:
+                        skb['free_t'] = time
+                        tx_free_list.append(skb)
+                        del tx_xmit_list[i]
+                        return
+        for i in range(len(rx_skb_list)):
+                rec_data = rx_skb_list[i]
+                if rec_data['skbaddr'] == skbaddr:
+                        rec_data.update({'handle':"kfree_skb",
+                                        'comm':comm, 'pid':pid, 'comm_t':time})
+                        del rx_skb_list[i]
+                        return
+
+def handle_consume_skb(event_info):
+        (name, context, cpu, time, pid, comm, skbaddr) = event_info
+        for i in range(len(tx_xmit_list)):
+                skb = tx_xmit_list[i]
+                if skb['skbaddr'] == skbaddr:
+                        skb['free_t'] = time
+                        tx_free_list.append(skb)
+                        del tx_xmit_list[i]
+                        return
+
+def handle_skb_copy_datagram_iovec(event_info):
+        (name, context, cpu, time, pid, comm, skbaddr, skblen) = event_info
+        for i in range(len(rx_skb_list)):
+                rec_data = rx_skb_list[i]
+                if skbaddr == rec_data['skbaddr']:
+                        rec_data.update({'handle':"skb_copy_datagram_iovec",
+                                        'comm':comm, 'pid':pid, 'comm_t':time})
+                        del rx_skb_list[i]
+                        return
diff --git a/tools/perf/scripts/python/sctop.py b/tools/perf/scripts/python/sctop.py
index 6cafad40c296..7a6ec2c7d8ab 100644
--- a/tools/perf/scripts/python/sctop.py
+++ b/tools/perf/scripts/python/sctop.py
@@ -8,10 +8,7 @@
 # will be refreshed every [interval] seconds.  The default interval is
 # 3 seconds.
 
-import thread
-import time
-import os
-import sys
+import os, sys, thread, time
 
 sys.path.append(os.environ['PERF_EXEC_PATH'] + \
         '/scripts/python/Perf-Trace-Util/lib/Perf/Trace')
@@ -20,7 +17,7 @@ from perf_trace_context import *
 from Core import *
 from Util import *
 
-usage = "perf trace -s syscall-counts.py [comm] [interval]\n";
+usage = "perf trace -s sctop.py [comm] [interval]\n";
 
 for_comm = None
 default_interval = 3
@@ -71,7 +68,7 @@ def print_syscall_totals(interval):
                 for id, val in sorted(syscalls.iteritems(), key = lambda(k, v): (v, k), \
                                               reverse = True):
                         try:
-                                print "%-40d  %10d\n" % (id, val),
+                                print "%-40s  %10d\n" % (syscall_name(id), val),
                         except TypeError:
                                 pass
                 syscalls.clear()
diff --git a/tools/perf/scripts/python/syscall-counts-by-pid.py b/tools/perf/scripts/python/syscall-counts-by-pid.py
index af722d6a4b3f..d1ee3ec10cf2 100644
--- a/tools/perf/scripts/python/syscall-counts-by-pid.py
+++ b/tools/perf/scripts/python/syscall-counts-by-pid.py
@@ -5,29 +5,33 @@
 # Displays system-wide system call totals, broken down by syscall.
 # If a [comm] arg is specified, only syscalls called by [comm] are displayed.
 
-import os
-import sys
+import os, sys
 
 sys.path.append(os.environ['PERF_EXEC_PATH'] + \
         '/scripts/python/Perf-Trace-Util/lib/Perf/Trace')
 
 from perf_trace_context import *
 from Core import *
+from Util import syscall_name
 
 usage = "perf trace -s syscall-counts-by-pid.py [comm]\n";
 
 for_comm = None
+for_pid = None
 
 if len(sys.argv) > 2:
         sys.exit(usage)
 
 if len(sys.argv) > 1:
-        for_comm = sys.argv[1]
+        try:
+                for_pid = int(sys.argv[1])
+        except:
+                for_comm = sys.argv[1]
 
 syscalls = autodict()
 
 def trace_begin():
-        pass
+        print "Press control+C to stop and show the summary"
 
 def trace_end():
         print_syscall_totals()
@@ -35,9 +39,10 @@ def trace_end():
 def raw_syscalls__sys_enter(event_name, context, common_cpu,
         common_secs, common_nsecs, common_pid, common_comm,
         id, args):
-        if for_comm is not None:
-                if common_comm != for_comm:
-                        return
+
+        if (for_comm and common_comm != for_comm) or \
+           (for_pid  and common_pid  != for_pid ):
+                return
         try:
                 syscalls[common_comm][common_pid][id] += 1
         except TypeError:
@@ -61,4 +66,4 @@ def print_syscall_totals():
                     id_keys = syscalls[comm][pid].keys()
                     for id, val in sorted(syscalls[comm][pid].iteritems(), \
                                   key = lambda(k, v): (v, k),  reverse = True):
-                            print "  %-38d  %10d\n" % (id, val),
+                            print "  %-38s  %10d\n" % (syscall_name(id), val),
diff --git a/tools/perf/scripts/python/syscall-counts.py b/tools/perf/scripts/python/syscall-counts.py
index f977e85ff049..ea183dc82d29 100644
--- a/tools/perf/scripts/python/syscall-counts.py
+++ b/tools/perf/scripts/python/syscall-counts.py
@@ -13,6 +13,7 @@ sys.path.append(os.environ['PERF_EXEC_PATH'] + \
 
 from perf_trace_context import *
 from Core import *
+from Util import syscall_name
 
 usage = "perf trace -s syscall-counts.py [comm]\n";
 
@@ -27,7 +28,7 @@ if len(sys.argv) > 1:
 syscalls = autodict()
 
 def trace_begin():
-        pass
+        print "Press control+C to stop and show the summary"
 
 def trace_end():
         print_syscall_totals()
@@ -55,4 +56,4 @@ def print_syscall_totals():
 
     for id, val in sorted(syscalls.iteritems(), key = lambda(k, v): (v, k), \
                                   reverse = True):
-            print "%-40d  %10d\n" % (id, val),
+            print "%-40s  %10d\n" % (syscall_name(id), val),
diff --git a/tools/perf/util/cache.h b/tools/perf/util/cache.h
index 27e9ebe4076e..a7729797fd96 100644
--- a/tools/perf/util/cache.h
+++ b/tools/perf/util/cache.h
@@ -82,6 +82,8 @@ extern char *perf_path(const char *fmt, ...) __attribute__((format (printf, 1, 2
 extern char *perf_pathdup(const char *fmt, ...)
         __attribute__((format (printf, 1, 2)));
 
+#ifdef NO_STRLCPY
 extern size_t strlcpy(char *dest, const char *src, size_t size);
+#endif
 
 #endif /* __PERF_CACHE_H */
diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c
index f231f43424d2..e12d539417b2 100644
--- a/tools/perf/util/callchain.c
+++ b/tools/perf/util/callchain.c
@@ -28,6 +28,9 @@ bool ip_callchain__valid(struct ip_callchain *chain, const event_t *event)
 #define chain_for_each_child(child, parent)     \
         list_for_each_entry(child, &parent->children, brothers)
 
+#define chain_for_each_child_safe(child, next, parent)  \
+        list_for_each_entry_safe(child, next, &parent->children, brothers)
+
 static void
 rb_insert_callchain(struct rb_root *root, struct callchain_node *chain,
                     enum chain_mode mode)
@@ -86,10 +89,10 @@ __sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node,
  * sort them by hit
  */
 static void
-sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node,
+sort_chain_flat(struct rb_root *rb_root, struct callchain_root *root,
                 u64 min_hit, struct callchain_param *param __used)
 {
-        __sort_chain_flat(rb_root, node, min_hit);
+        __sort_chain_flat(rb_root, &root->node, min_hit);
 }
 
 static void __sort_chain_graph_abs(struct callchain_node *node,
@@ -108,11 +111,11 @@ static void __sort_chain_graph_abs(struct callchain_node *node,
 }
 
 static void
-sort_chain_graph_abs(struct rb_root *rb_root, struct callchain_node *chain_root,
+sort_chain_graph_abs(struct rb_root *rb_root, struct callchain_root *chain_root,
                      u64 min_hit, struct callchain_param *param __used)
 {
-        __sort_chain_graph_abs(chain_root, min_hit);
-        rb_root->rb_node = chain_root->rb_root.rb_node;
+        __sort_chain_graph_abs(&chain_root->node, min_hit);
+        rb_root->rb_node = chain_root->node.rb_root.rb_node;
 }
 
 static void __sort_chain_graph_rel(struct callchain_node *node,
@@ -133,11 +136,11 @@ static void __sort_chain_graph_rel(struct callchain_node *node,
 }
 
 static void
-sort_chain_graph_rel(struct rb_root *rb_root, struct callchain_node *chain_root,
+sort_chain_graph_rel(struct rb_root *rb_root, struct callchain_root *chain_root,
                      u64 min_hit __used, struct callchain_param *param)
 {
-        __sort_chain_graph_rel(chain_root, param->min_percent / 100.0);
-        rb_root->rb_node = chain_root->rb_root.rb_node;
+        __sort_chain_graph_rel(&chain_root->node, param->min_percent / 100.0);
+        rb_root->rb_node = chain_root->node.rb_root.rb_node;
 }
 
 int register_callchain_param(struct callchain_param *param)
@@ -284,19 +287,18 @@ split_add_child(struct callchain_node *parent, struct resolved_chain *chain,
 }
 
 static int
-__append_chain(struct callchain_node *root, struct resolved_chain *chain,
-               unsigned int start, u64 period);
+append_chain(struct callchain_node *root, struct resolved_chain *chain,
+             unsigned int start, u64 period);
 
 static void
-__append_chain_children(struct callchain_node *root,
-                        struct resolved_chain *chain,
-                        unsigned int start, u64 period)
+append_chain_children(struct callchain_node *root, struct resolved_chain *chain,
+                      unsigned int start, u64 period)
 {
         struct callchain_node *rnode;
 
         /* lookup in childrens */
         chain_for_each_child(rnode, root) {
-                unsigned int ret = __append_chain(rnode, chain, start, period);
+                unsigned int ret = append_chain(rnode, chain, start, period);
 
                 if (!ret)
                         goto inc_children_hit;
@@ -309,8 +311,8 @@ inc_children_hit:
 }
 
 static int
-__append_chain(struct callchain_node *root, struct resolved_chain *chain,
-               unsigned int start, u64 period)
+append_chain(struct callchain_node *root, struct resolved_chain *chain,
+             unsigned int start, u64 period)
 {
         struct callchain_list *cnode;
         unsigned int i = start;
@@ -357,7 +359,7 @@ __append_chain(struct callchain_node *root, struct resolved_chain *chain,
         }
 
         /* We match the node and still have a part remaining */
-        __append_chain_children(root, chain, i, period);
+        append_chain_children(root, chain, i, period);
 
         return 0;
 }
@@ -380,8 +382,8 @@ static void filter_context(struct ip_callchain *old, struct resolved_chain *new,
 }
 
 
-int append_chain(struct callchain_node *root, struct ip_callchain *chain,
-                 struct map_symbol *syms, u64 period)
+int callchain_append(struct callchain_root *root, struct ip_callchain *chain,
+                     struct map_symbol *syms, u64 period)
 {
         struct resolved_chain *filtered;
 
@@ -398,9 +400,65 @@ int append_chain(struct callchain_node *root, struct ip_callchain *chain,
         if (!filtered->nr)
                 goto end;
 
-        __append_chain_children(root, filtered, 0, period);
+        append_chain_children(&root->node, filtered, 0, period);
+
+        if (filtered->nr > root->max_depth)
+                root->max_depth = filtered->nr;
 end:
         free(filtered);
 
         return 0;
 }
+
+static int
+merge_chain_branch(struct callchain_node *dst, struct callchain_node *src,
+                   struct resolved_chain *chain)
+{
+        struct callchain_node *child, *next_child;
+        struct callchain_list *list, *next_list;
+        int old_pos = chain->nr;
+        int err = 0;
+
+        list_for_each_entry_safe(list, next_list, &src->val, list) {
+                chain->ips[chain->nr].ip = list->ip;
+                chain->ips[chain->nr].ms = list->ms;
+                chain->nr++;
+                list_del(&list->list);
+                free(list);
+        }
+
+        if (src->hit)
+                append_chain_children(dst, chain, 0, src->hit);
+
+        chain_for_each_child_safe(child, next_child, src) {
+                err = merge_chain_branch(dst, child, chain);
+                if (err)
+                        break;
+
+                list_del(&child->brothers);
+                free(child);
+        }
+
+        chain->nr = old_pos;
+
+        return err;
+}
+
+int callchain_merge(struct callchain_root *dst, struct callchain_root *src)
+{
+        struct resolved_chain *chain;
+        int err;
+
+        chain = malloc(sizeof(*chain) +
+                       src->max_depth * sizeof(struct resolved_ip));
+        if (!chain)
+                return -ENOMEM;
+
+        chain->nr = 0;
+
+        err = merge_chain_branch(&dst->node, &src->node, chain);
+
+        free(chain);
+
+        return err;
+}
diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h
index 6de4313924fb..c15fb8c24ad2 100644
--- a/tools/perf/util/callchain.h
+++ b/tools/perf/util/callchain.h
@@ -26,9 +26,14 @@ struct callchain_node {
         u64                     children_hit;
 };
 
+struct callchain_root {
+        u64                     max_depth;
+        struct callchain_node   node;
+};
+
 struct callchain_param;
 
-typedef void (*sort_chain_func_t)(struct rb_root *, struct callchain_node *,
+typedef void (*sort_chain_func_t)(struct rb_root *, struct callchain_root *,
                                  u64, struct callchain_param *);
 
 struct callchain_param {
@@ -44,15 +49,16 @@ struct callchain_list {
         struct list_head        list;
 };
 
-static inline void callchain_init(struct callchain_node *node)
+static inline void callchain_init(struct callchain_root *root)
 {
-        INIT_LIST_HEAD(&node->brothers);
-        INIT_LIST_HEAD(&node->children);
-        INIT_LIST_HEAD(&node->val);
+        INIT_LIST_HEAD(&root->node.brothers);
+        INIT_LIST_HEAD(&root->node.children);
+        INIT_LIST_HEAD(&root->node.val);
 
-        node->children_hit = 0;
-        node->parent = NULL;
-        node->hit = 0;
+        root->node.parent = NULL;
+        root->node.hit = 0;
+        root->node.children_hit = 0;
+        root->max_depth = 0;
 }
 
 static inline u64 cumul_hits(struct callchain_node *node)
@@ -61,8 +67,9 @@ static inline u64 cumul_hits(struct callchain_node *node)
 }
 
 int register_callchain_param(struct callchain_param *param);
-int append_chain(struct callchain_node *root, struct ip_callchain *chain,
-                 struct map_symbol *syms, u64 period);
+int callchain_append(struct callchain_root *root, struct ip_callchain *chain,
+                     struct map_symbol *syms, u64 period);
+int callchain_merge(struct callchain_root *dst, struct callchain_root *src);
 
 bool ip_callchain__valid(struct ip_callchain *chain, const event_t *event);
 #endif  /* __PERF_CALLCHAIN_H */
diff --git a/tools/perf/util/debug.c b/tools/perf/util/debug.c
index f9c7e3ad1aa7..c8d81b00089d 100644
--- a/tools/perf/util/debug.c
+++ b/tools/perf/util/debug.c
@@ -12,8 +12,8 @@
 #include "debug.h"
 #include "util.h"
 
-int verbose = 0;
-bool dump_trace = false;
+int verbose;
+bool dump_trace = false, quiet = false;
 
 int eprintf(int level, const char *fmt, ...)
 {
diff --git a/tools/perf/util/debug.h b/tools/perf/util/debug.h
index 7a17ee061bcb..7b514082bbaf 100644
--- a/tools/perf/util/debug.h
+++ b/tools/perf/util/debug.h
@@ -6,7 +6,7 @@
 #include "event.h"
 
 extern int verbose;
-extern bool dump_trace;
+extern bool quiet, dump_trace;
 
 int dump_printf(const char *fmt, ...) __attribute__((format(printf, 1, 2)));
 void trace_event(event_t *event);
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index be22ae6ef055..2022e8740994 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -87,7 +87,7 @@ static void hist_entry__add_cpumode_period(struct hist_entry *self,
 
 static struct hist_entry *hist_entry__new(struct hist_entry *template)
 {
-        size_t callchain_size = symbol_conf.use_callchain ? sizeof(struct callchain_node) : 0;
+        size_t callchain_size = symbol_conf.use_callchain ? sizeof(struct callchain_root) : 0;
         struct hist_entry *self = malloc(sizeof(*self) + callchain_size);
 
         if (self != NULL) {
@@ -226,6 +226,8 @@ static bool collapse__insert_entry(struct rb_root *root, struct hist_entry *he)
 
                 if (!cmp) {
                         iter->period += he->period;
+                        if (symbol_conf.use_callchain)
+                                callchain_merge(iter->callchain, he->callchain);
                         hist_entry__free(he);
                         return false;
                 }
diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h
index 78575796d5f3..b397c0383728 100644
--- a/tools/perf/util/map.h
+++ b/tools/perf/util/map.h
@@ -215,6 +215,16 @@ struct symbol *map_groups__find_function_by_name(struct map_groups *self,
         return map_groups__find_symbol_by_name(self, MAP__FUNCTION, name, mapp, filter);
 }
 
+static inline
+struct symbol *machine__find_kernel_function_by_name(struct machine *self,
+                                                     const char *name,
+                                                     struct map **mapp,
+                                                     symbol_filter_t filter)
+{
+        return map_groups__find_function_by_name(&self->kmaps, name, mapp,
+                                                 filter);
+}
+
 int map_groups__fixup_overlappings(struct map_groups *self, struct map *map,
                                    int verbose, FILE *fp);
 
diff --git a/tools/perf/util/path.c b/tools/perf/util/path.c
index 58a470d036dd..bd7497711424 100644
--- a/tools/perf/util/path.c
+++ b/tools/perf/util/path.c
@@ -22,6 +22,7 @@ static const char *get_perf_dir(void)
         return ".";
 }
 
+#ifdef NO_STRLCPY
 size_t strlcpy(char *dest, const char *src, size_t size)
 {
         size_t ret = strlen(src);
@@ -33,7 +34,7 @@ size_t strlcpy(char *dest, const char *src, size_t size)
         }
         return ret;
 }
-
+#endif
 
 static char *get_pathname(void)
 {
diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index fcc16e4349df..3b6a5297bf16 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -74,10 +74,9 @@ static int e_snprintf(char *str, size_t size, const char *format, ...)
 static char *synthesize_perf_probe_point(struct perf_probe_point *pp);
 static struct machine machine;
 
-/* Initialize symbol maps and path of vmlinux */
+/* Initialize symbol maps and path of vmlinux/modules */
 static int init_vmlinux(void)
 {
-        struct dso *kernel;
         int ret;
 
         symbol_conf.sort_by_name = true;
@@ -91,33 +90,61 @@ static int init_vmlinux(void)
                 goto out;
         }
 
-        ret = machine__init(&machine, "/", 0);
+        ret = machine__init(&machine, "", HOST_KERNEL_ID);
         if (ret < 0)
                 goto out;
 
-        kernel = dso__new_kernel(symbol_conf.vmlinux_name);
-        if (kernel == NULL)
-                die("Failed to create kernel dso.");
-
-        ret = __machine__create_kernel_maps(&machine, kernel);
-        if (ret < 0)
-                pr_debug("Failed to create kernel maps.\n");
-
+        if (machine__create_kernel_maps(&machine) < 0) {
+                pr_debug("machine__create_kernel_maps ");
+                goto out;
+        }
 out:
         if (ret < 0)
                 pr_warning("Failed to init vmlinux path.\n");
         return ret;
 }
 
+static struct symbol *__find_kernel_function_by_name(const char *name,
+                                                     struct map **mapp)
+{
+        return machine__find_kernel_function_by_name(&machine, name, mapp,
+                                                     NULL);
+}
+
+const char *kernel_get_module_path(const char *module)
+{
+        struct dso *dso;
+
+        if (module) {
+                list_for_each_entry(dso, &machine.kernel_dsos, node) {
+                        if (strncmp(dso->short_name + 1, module,
+                                    dso->short_name_len - 2) == 0)
+                                goto found;
+                }
+                pr_debug("Failed to find module %s.\n", module);
+                return NULL;
+        } else {
+                dso = machine.vmlinux_maps[MAP__FUNCTION]->dso;
+                if (dso__load_vmlinux_path(dso,
+                         machine.vmlinux_maps[MAP__FUNCTION], NULL) < 0) {
+                        pr_debug("Failed to load kernel map.\n");
+                        return NULL;
+                }
+        }
+found:
+        return dso->long_name;
+}
+
 #ifdef DWARF_SUPPORT
-static int open_vmlinux(void)
+static int open_vmlinux(const char *module)
 {
-        if (map__load(machine.vmlinux_maps[MAP__FUNCTION], NULL) < 0) {
-                pr_debug("Failed to load kernel map.\n");
-                return -EINVAL;
+        const char *path = kernel_get_module_path(module);
+        if (!path) {
+                pr_err("Failed to find path of %s module", module ?: "kernel");
+                return -ENOENT;
         }
-        pr_debug("Try to open %s\n", machine.vmlinux_maps[MAP__FUNCTION]->dso->long_name);
-        return open(machine.vmlinux_maps[MAP__FUNCTION]->dso->long_name, O_RDONLY);
+        pr_debug("Try to open %s\n", path);
+        return open(path, O_RDONLY);
 }
 
 /*
@@ -125,20 +152,19 @@ static int open_vmlinux(void)
  * Currently only handles kprobes.
  */
 static int kprobe_convert_to_perf_probe(struct probe_trace_point *tp,
-                                       struct perf_probe_point *pp)
+                                        struct perf_probe_point *pp)
 {
         struct symbol *sym;
-        int fd, ret = -ENOENT;
+        struct map *map;
+        u64 addr;
+        int ret = -ENOENT;
 
-        sym = map__find_symbol_by_name(machine.vmlinux_maps[MAP__FUNCTION],
-                                       tp->symbol, NULL);
+        sym = __find_kernel_function_by_name(tp->symbol, &map);
         if (sym) {
-                fd = open_vmlinux();
-                if (fd >= 0) {
-                        ret = find_perf_probe_point(fd,
-                                                 sym->start + tp->offset, pp);
-                        close(fd);
-                }
+                addr = map->unmap_ip(map, sym->start + tp->offset);
+                pr_debug("try to find %s+%ld@%llx\n", tp->symbol,
+                         tp->offset, addr);
+                ret = find_perf_probe_point((unsigned long)addr, pp);
         }
         if (ret <= 0) {
                 pr_debug("Failed to find corresponding probes from "
@@ -156,12 +182,12 @@ static int kprobe_convert_to_perf_probe(struct probe_trace_point *tp,
 /* Try to find perf_probe_event with debuginfo */
 static int try_to_find_probe_trace_events(struct perf_probe_event *pev,
                                            struct probe_trace_event **tevs,
-                                           int max_tevs)
+                                           int max_tevs, const char *module)
 {
         bool need_dwarf = perf_probe_event_need_dwarf(pev);
         int fd, ntevs;
 
-        fd = open_vmlinux();
+        fd = open_vmlinux(module);
         if (fd < 0) {
                 if (need_dwarf) {
                         pr_warning("Failed to open debuginfo file.\n");
@@ -300,7 +326,7 @@ error:
  * Show line-range always requires debuginfo to find source file and
  * line number.
  */
-int show_line_range(struct line_range *lr)
+int show_line_range(struct line_range *lr, const char *module)
 {
         int l = 1;
         struct line_node *ln;
@@ -313,7 +339,7 @@ int show_line_range(struct line_range *lr)
         if (ret < 0)
                 return ret;
 
-        fd = open_vmlinux();
+        fd = open_vmlinux(module);
         if (fd < 0) {
                 pr_warning("Failed to open debuginfo file.\n");
                 return fd;
@@ -378,11 +404,84 @@ end:
         return ret;
 }
 
+static int show_available_vars_at(int fd, struct perf_probe_event *pev,
+                                  int max_vls, bool externs)
+{
+        char *buf;
+        int ret, i;
+        struct str_node *node;
+        struct variable_list *vls = NULL, *vl;
+
+        buf = synthesize_perf_probe_point(&pev->point);
+        if (!buf)
+                return -EINVAL;
+        pr_debug("Searching variables at %s\n", buf);
+
+        ret = find_available_vars_at(fd, pev, &vls, max_vls, externs);
+        if (ret > 0) {
+                /* Some variables were found */
+                fprintf(stdout, "Available variables at %s\n", buf);
+                for (i = 0; i < ret; i++) {
+                        vl = &vls[i];
+                        /*
+                         * A probe point might be converted to
+                         * several trace points.
+                         */
+                        fprintf(stdout, "\t@<%s+%lu>\n", vl->point.symbol,
+                                vl->point.offset);
+                        free(vl->point.symbol);
+                        if (vl->vars) {
+                                strlist__for_each(node, vl->vars)
+                                        fprintf(stdout, "\t\t%s\n", node->s);
+                                strlist__delete(vl->vars);
+                        } else
+                                fprintf(stdout, "(No variables)\n");
+                }
+                free(vls);
+        } else
+                pr_err("Failed to find variables at %s (%d)\n", buf, ret);
+
+        free(buf);
+        return ret;
+}
+
+/* Show available variables on given probe point */
+int show_available_vars(struct perf_probe_event *pevs, int npevs,
+                        int max_vls, const char *module, bool externs)
+{
+        int i, fd, ret = 0;
+
+        ret = init_vmlinux();
+        if (ret < 0)
+                return ret;
+
+        fd = open_vmlinux(module);
+        if (fd < 0) {
+                pr_warning("Failed to open debuginfo file.\n");
+                return fd;
+        }
+
+        setup_pager();
+
+        for (i = 0; i < npevs && ret >= 0; i++)
+                ret = show_available_vars_at(fd, &pevs[i], max_vls, externs);
+
+        close(fd);
+        return ret;
+}
+
 #else   /* !DWARF_SUPPORT */
 
 static int kprobe_convert_to_perf_probe(struct probe_trace_point *tp,
-                                       struct perf_probe_point *pp)
+                                        struct perf_probe_point *pp)
 {
+        struct symbol *sym;
+
+        sym = __find_kernel_function_by_name(tp->symbol, NULL);
+        if (!sym) {
+                pr_err("Failed to find symbol %s in kernel.\n", tp->symbol);
+                return -ENOENT;
+        }
         pp->function = strdup(tp->symbol);
         if (pp->function == NULL)
                 return -ENOMEM;
@@ -394,7 +493,7 @@ static int kprobe_convert_to_perf_probe(struct probe_trace_point *tp,
 
 static int try_to_find_probe_trace_events(struct perf_probe_event *pev,
                                 struct probe_trace_event **tevs __unused,
-                                int max_tevs __unused)
+                                int max_tevs __unused, const char *mod __unused)
 {
         if (perf_probe_event_need_dwarf(pev)) {
                 pr_warning("Debuginfo-analysis is not supported.\n");
@@ -403,12 +502,19 @@ static int try_to_find_probe_trace_events(struct perf_probe_event *pev,
         return 0;
 }
 
-int show_line_range(struct line_range *lr __unused)
+int show_line_range(struct line_range *lr __unused, const char *module __unused)
 {
         pr_warning("Debuginfo-analysis is not supported.\n");
         return -ENOSYS;
 }
 
+int show_available_vars(struct perf_probe_event *pevs __unused,
+                        int npevs __unused, int max_vls __unused,
+                        const char *module __unused, bool externs __unused)
+{
+        pr_warning("Debuginfo-analysis is not supported.\n");
+        return -ENOSYS;
+}
 #endif
 
 int parse_line_range_desc(const char *arg, struct line_range *lr)
@@ -1087,7 +1193,7 @@ error:
 }
 
 static int convert_to_perf_probe_event(struct probe_trace_event *tev,
-                                struct perf_probe_event *pev)
+                                       struct perf_probe_event *pev)
 {
         char buf[64] = "";
         int i, ret;
@@ -1516,14 +1622,14 @@ static int __add_probe_trace_events(struct perf_probe_event *pev,
 
 static int convert_to_probe_trace_events(struct perf_probe_event *pev,
                                           struct probe_trace_event **tevs,
-                                          int max_tevs)
+                                          int max_tevs, const char *module)
 {
         struct symbol *sym;
         int ret = 0, i;
         struct probe_trace_event *tev;
 
         /* Convert perf_probe_event with debuginfo */
-        ret = try_to_find_probe_trace_events(pev, tevs, max_tevs);
+        ret = try_to_find_probe_trace_events(pev, tevs, max_tevs, module);
         if (ret != 0)
                 return ret;
 
@@ -1572,8 +1678,7 @@ static int convert_to_probe_trace_events(struct perf_probe_event *pev,
         }
 
         /* Currently just checking function name from symbol map */
-        sym = map__find_symbol_by_name(machine.vmlinux_maps[MAP__FUNCTION],
-                                       tev->point.symbol, NULL);
+        sym = __find_kernel_function_by_name(tev->point.symbol, NULL);
         if (!sym) {
                 pr_warning("Kernel symbol \'%s\' not found.\n",
                            tev->point.symbol);
@@ -1596,7 +1701,7 @@ struct __event_package {
 };
 
 int add_perf_probe_events(struct perf_probe_event *pevs, int npevs,
-                          bool force_add, int max_tevs)
+                          int max_tevs, const char *module, bool force_add)
 {
         int i, j, ret;
         struct __event_package *pkgs;
@@ -1617,7 +1722,9 @@ int add_perf_probe_events(struct perf_probe_event *pevs, int npevs,
                 pkgs[i].pev = &pevs[i];
                 /* Convert with or without debuginfo */
                 ret  = convert_to_probe_trace_events(pkgs[i].pev,
-                                                      &pkgs[i].tevs, max_tevs);
+                                                     &pkgs[i].tevs,
+                                                     max_tevs,
+                                                     module);
                 if (ret < 0)
                         goto end;
                 pkgs[i].ntevs = ret;
diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h
index 5af39243a25b..5accbedfea37 100644
--- a/tools/perf/util/probe-event.h
+++ b/tools/perf/util/probe-event.h
@@ -90,6 +90,12 @@ struct line_range {
         struct list_head        line_list;      /* Visible lines */
 };
 
+/* List of variables */
+struct variable_list {
+        struct probe_trace_point        point;  /* Actual probepoint */
+        struct strlist                  *vars;  /* Available variables */
+};
+
 /* Command string to events */
 extern int parse_perf_probe_command(const char *cmd,
                                     struct perf_probe_event *pev);
@@ -109,12 +115,18 @@ extern void clear_perf_probe_event(struct perf_probe_event *pev);
 /* Command string to line-range */
 extern int parse_line_range_desc(const char *cmd, struct line_range *lr);
 
+/* Internal use: Return kernel/module path */
+extern const char *kernel_get_module_path(const char *module);
 
 extern int add_perf_probe_events(struct perf_probe_event *pevs, int npevs,
-                                 bool force_add, int max_probe_points);
+                                 int max_probe_points, const char *module,
+                                 bool force_add);
 extern int del_perf_probe_events(struct strlist *dellist);
 extern int show_perf_probe_events(void);
-extern int show_line_range(struct line_range *lr);
+extern int show_line_range(struct line_range *lr, const char *module);
+extern int show_available_vars(struct perf_probe_event *pevs, int npevs,
+                               int max_probe_points, const char *module,
+                               bool externs);
 
 
 /* Maximum index number of event-name postfix */
diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c
index 32b81f707ff5..3991d73d1cff 100644
--- a/tools/perf/util/probe-finder.c
+++ b/tools/perf/util/probe-finder.c
@@ -116,6 +116,101 @@ static void line_list__free(struct list_head *head)
         }
 }
 
+/* Dwarf FL wrappers */
+
+static int __linux_kernel_find_elf(Dwfl_Module *mod,
+                                   void **userdata,
+                                   const char *module_name,
+                                   Dwarf_Addr base,
+                                   char **file_name, Elf **elfp)
+{
+        int fd;
+        const char *path = kernel_get_module_path(module_name);
+
+        if (path) {
+                fd = open(path, O_RDONLY);
+                if (fd >= 0) {
+                        *file_name = strdup(path);
+                        return fd;
+                }
+        }
+        /* If failed, try to call standard method */
+        return dwfl_linux_kernel_find_elf(mod, userdata, module_name, base,
+                                          file_name, elfp);
+}
+
+static char *debuginfo_path;    /* Currently dummy */
+
+static const Dwfl_Callbacks offline_callbacks = {
+        .find_debuginfo = dwfl_standard_find_debuginfo,
+        .debuginfo_path = &debuginfo_path,
+
+        .section_address = dwfl_offline_section_address,
+
+        /* We use this table for core files too.  */
+        .find_elf = dwfl_build_id_find_elf,
+};
+
+static const Dwfl_Callbacks kernel_callbacks = {
+        .find_debuginfo = dwfl_standard_find_debuginfo,
+        .debuginfo_path = &debuginfo_path,
+
+        .find_elf = __linux_kernel_find_elf,
+        .section_address = dwfl_linux_kernel_module_section_address,
+};
+
+/* Get a Dwarf from offline image */
+static Dwarf *dwfl_init_offline_dwarf(int fd, Dwfl **dwflp, Dwarf_Addr *bias)
+{
+        Dwfl_Module *mod;
+        Dwarf *dbg = NULL;
+
+        if (!dwflp)
+                return NULL;
+
+        *dwflp = dwfl_begin(&offline_callbacks);
+        if (!*dwflp)
+                return NULL;
+
+        mod = dwfl_report_offline(*dwflp, "", "", fd);
+        if (!mod)
+                goto error;
+
+        dbg = dwfl_module_getdwarf(mod, bias);
+        if (!dbg) {
+error:
+                dwfl_end(*dwflp);
+                *dwflp = NULL;
+        }
+        return dbg;
+}
+
+/* Get a Dwarf from live kernel image */
+static Dwarf *dwfl_init_live_kernel_dwarf(Dwarf_Addr addr, Dwfl **dwflp,
+                                          Dwarf_Addr *bias)
+{
+        Dwarf *dbg;
+
+        if (!dwflp)
+                return NULL;
+
+        *dwflp = dwfl_begin(&kernel_callbacks);
+        if (!*dwflp)
+                return NULL;
+
+        /* Load the kernel dwarves: Don't care the result here */
+        dwfl_linux_kernel_report_kernel(*dwflp);
+        dwfl_linux_kernel_report_modules(*dwflp);
+
+        dbg = dwfl_addrdwarf(*dwflp, addr, bias);
+        /* Here, check whether we could get a real dwarf */
+        if (!dbg) {
+                dwfl_end(*dwflp);
+                *dwflp = NULL;
+        }
+        return dbg;
+}
+
 /* Dwarf wrappers */
 
 /* Find the realpath of the target file. */
@@ -160,26 +255,44 @@ static bool die_compare_name(Dwarf_Die *dw_die, const char *tname)
         return name ? (strcmp(tname, name) == 0) : false;
 }
 
-/* Get type die, but skip qualifiers and typedef */
-static Dwarf_Die *die_get_real_type(Dwarf_Die *vr_die, Dwarf_Die *die_mem)
+/* Get type die */
+static Dwarf_Die *die_get_type(Dwarf_Die *vr_die, Dwarf_Die *die_mem)
 {
         Dwarf_Attribute attr;
+
+        if (dwarf_attr_integrate(vr_die, DW_AT_type, &attr) &&
+            dwarf_formref_die(&attr, die_mem))
+                return die_mem;
+        else
+                return NULL;
+}
+
+/* Get a type die, but skip qualifiers */
+static Dwarf_Die *__die_get_real_type(Dwarf_Die *vr_die, Dwarf_Die *die_mem)
+{
         int tag;
 
         do {
-                if (dwarf_attr(vr_die, DW_AT_type, &attr) == NULL ||
-                    dwarf_formref_die(&attr, die_mem) == NULL)
-                        return NULL;
-
-                tag = dwarf_tag(die_mem);
-                vr_die = die_mem;
+                vr_die = die_get_type(vr_die, die_mem);
+                if (!vr_die)
+                        break;
+                tag = dwarf_tag(vr_die);
         } while (tag == DW_TAG_const_type ||
                  tag == DW_TAG_restrict_type ||
                  tag == DW_TAG_volatile_type ||
-                 tag == DW_TAG_shared_type ||
-                 tag == DW_TAG_typedef);
+                 tag == DW_TAG_shared_type);
+
+        return vr_die;
+}
 
-        return die_mem;
+/* Get a type die, but skip qualifiers and typedef */
+static Dwarf_Die *die_get_real_type(Dwarf_Die *vr_die, Dwarf_Die *die_mem)
+{
+        do {
+                vr_die = __die_get_real_type(vr_die, die_mem);
+        } while (vr_die && dwarf_tag(vr_die) == DW_TAG_typedef);
+
+        return vr_die;
 }
 
 static bool die_is_signed_type(Dwarf_Die *tp_die)
@@ -320,25 +433,35 @@ static Dwarf_Die *die_find_inlinefunc(Dwarf_Die *sp_die, Dwarf_Addr addr,
         return die_find_child(sp_die, __die_find_inline_cb, &addr, die_mem);
 }
 
+struct __find_variable_param {
+        const char *name;
+        Dwarf_Addr addr;
+};
+
 static int __die_find_variable_cb(Dwarf_Die *die_mem, void *data)
 {
-        const char *name = data;
+        struct __find_variable_param *fvp = data;
         int tag;
 
         tag = dwarf_tag(die_mem);
         if ((tag == DW_TAG_formal_parameter ||
              tag == DW_TAG_variable) &&
-            die_compare_name(die_mem, name))
+            die_compare_name(die_mem, fvp->name))
                 return DIE_FIND_CB_FOUND;
 
-        return DIE_FIND_CB_CONTINUE;
+        if (dwarf_haspc(die_mem, fvp->addr))
+                return DIE_FIND_CB_CONTINUE;
+        else
+                return DIE_FIND_CB_SIBLING;
 }
 
-/* Find a variable called 'name' */
-static Dwarf_Die *die_find_variable(Dwarf_Die *sp_die, const char *name,
-                                    Dwarf_Die *die_mem)
+/* Find a variable called 'name' at given address */
+static Dwarf_Die *die_find_variable_at(Dwarf_Die *sp_die, const char *name,
+                                       Dwarf_Addr addr, Dwarf_Die *die_mem)
 {
-        return die_find_child(sp_die, __die_find_variable_cb, (void *)name,
+        struct __find_variable_param fvp = { .name = name, .addr = addr};
+
+        return die_find_child(sp_die, __die_find_variable_cb, (void *)&fvp,
                               die_mem);
 }
 
@@ -361,6 +484,60 @@ static Dwarf_Die *die_find_member(Dwarf_Die *st_die, const char *name,
                               die_mem);
 }
 
+/* Get the name of given variable DIE */
+static int die_get_typename(Dwarf_Die *vr_die, char *buf, int len)
+{
+        Dwarf_Die type;
+        int tag, ret, ret2;
+        const char *tmp = "";
+
+        if (__die_get_real_type(vr_die, &type) == NULL)
+                return -ENOENT;
+
+        tag = dwarf_tag(&type);
+        if (tag == DW_TAG_array_type || tag == DW_TAG_pointer_type)
+                tmp = "*";
+        else if (tag == DW_TAG_subroutine_type) {
+                /* Function pointer */
+                ret = snprintf(buf, len, "(function_type)");
+                return (ret >= len) ? -E2BIG : ret;
+        } else {
+                if (!dwarf_diename(&type))
+                        return -ENOENT;
+                if (tag == DW_TAG_union_type)
+                        tmp = "union ";
+                else if (tag == DW_TAG_structure_type)
+                        tmp = "struct ";
+                /* Write a base name */
+                ret = snprintf(buf, len, "%s%s", tmp, dwarf_diename(&type));
+                return (ret >= len) ? -E2BIG : ret;
+        }
+        ret = die_get_typename(&type, buf, len);
+        if (ret > 0) {
+                ret2 = snprintf(buf + ret, len - ret, "%s", tmp);
+                ret = (ret2 >= len - ret) ? -E2BIG : ret2 + ret;
+        }
+        return ret;
+}
+
+/* Get the name and type of given variable DIE, stored as "type\tname" */
+static int die_get_varname(Dwarf_Die *vr_die, char *buf, int len)
+{
+        int ret, ret2;
+
+        ret = die_get_typename(vr_die, buf, len);
+        if (ret < 0) {
+                pr_debug("Failed to get type, make it unknown.\n");
+                ret = snprintf(buf, len, "(unknown_type)");
+        }
+        if (ret > 0) {
+                ret2 = snprintf(buf + ret, len - ret, "\t%s",
+                                dwarf_diename(vr_die));
+                ret = (ret2 >= len - ret) ? -E2BIG : ret2 + ret;
+        }
+        return ret;
+}
+
 /*
  * Probe finder related functions
  */
@@ -374,8 +551,13 @@ static struct probe_trace_arg_ref *alloc_trace_arg_ref(long offs)
         return ref;
 }
 
-/* Show a location */
-static int convert_variable_location(Dwarf_Die *vr_die, struct probe_finder *pf)
+/*
+ * Convert a location into trace_arg.
+ * If tvar == NULL, this just checks variable can be converted.
+ */
+static int convert_variable_location(Dwarf_Die *vr_die, Dwarf_Addr addr,
+                                     Dwarf_Op *fb_ops,
+                                     struct probe_trace_arg *tvar)
 {
         Dwarf_Attribute attr;
         Dwarf_Op *op;
@@ -384,20 +566,23 @@ static int convert_variable_location(Dwarf_Die *vr_die, struct probe_finder *pf)
         Dwarf_Word offs = 0;
         bool ref = false;
         const char *regs;
-        struct probe_trace_arg *tvar = pf->tvar;
         int ret;
 
+        if (dwarf_attr(vr_die, DW_AT_external, &attr) != NULL)
+                goto static_var;
+
         /* TODO: handle more than 1 exprs */
         if (dwarf_attr(vr_die, DW_AT_location, &attr) == NULL ||
-            dwarf_getlocation_addr(&attr, pf->addr, &op, &nops, 1) <= 0 ||
+            dwarf_getlocation_addr(&attr, addr, &op, &nops, 1) <= 0 ||
             nops == 0) {
                 /* TODO: Support const_value */
-                pr_err("Failed to find the location of %s at this address.\n"
-                       " Perhaps, it has been optimized out.\n", pf->pvar->var);
                 return -ENOENT;
         }
 
         if (op->atom == DW_OP_addr) {
+static_var:
+                if (!tvar)
+                        return 0;
                 /* Static variables on memory (not stack), make @varname */
                 ret = strlen(dwarf_diename(vr_die));
                 tvar->value = zalloc(ret + 2);
@@ -412,14 +597,11 @@ static int convert_variable_location(Dwarf_Die *vr_die, struct probe_finder *pf)
 
         /* If this is based on frame buffer, set the offset */
         if (op->atom == DW_OP_fbreg) {
-                if (pf->fb_ops == NULL) {
-                        pr_warning("The attribute of frame base is not "
-                                   "supported.\n");
+                if (fb_ops == NULL)
                         return -ENOTSUP;
-                }
                 ref = true;
                 offs = op->number;
-                op = &pf->fb_ops[0];
+                op = &fb_ops[0];
         }
 
         if (op->atom >= DW_OP_breg0 && op->atom <= DW_OP_breg31) {
@@ -435,13 +617,18 @@ static int convert_variable_location(Dwarf_Die *vr_die, struct probe_finder *pf)
         } else if (op->atom == DW_OP_regx) {
                 regn = op->number;
         } else {
-                pr_warning("DW_OP %x is not supported.\n", op->atom);
+                pr_debug("DW_OP %x is not supported.\n", op->atom);
                 return -ENOTSUP;
         }
 
+        if (!tvar)
+                return 0;
+
         regs = get_arch_regstr(regn);
         if (!regs) {
-                pr_warning("Mapping for DWARF register number %u missing on this architecture.", regn);
+                /* This should be a bug in DWARF or this tool */
+                pr_warning("Mapping for DWARF register number %u "
+                           "missing on this architecture.", regn);
                 return -ERANGE;
         }
 
@@ -666,8 +853,14 @@ static int convert_variable(Dwarf_Die *vr_die, struct probe_finder *pf)
         pr_debug("Converting variable %s into trace event.\n",
                  dwarf_diename(vr_die));
 
-        ret = convert_variable_location(vr_die, pf);
-        if (ret == 0 && pf->pvar->field) {
+        ret = convert_variable_location(vr_die, pf->addr, pf->fb_ops,
+                                        pf->tvar);
+        if (ret == -ENOENT)
+                pr_err("Failed to find the location of %s at this address.\n"
+                       " Perhaps, it has been optimized out.\n", pf->pvar->var);
+        else if (ret == -ENOTSUP)
+                pr_err("Sorry, we don't support this variable location yet.\n");
+        else if (pf->pvar->field) {
                 ret = convert_variable_fields(vr_die, pf->pvar->var,
                                               pf->pvar->field, &pf->tvar->ref,
                                               &die_mem);
@@ -722,56 +915,39 @@ static int find_variable(Dwarf_Die *sp_die, struct probe_finder *pf)
         pr_debug("Searching '%s' variable in context.\n",
                  pf->pvar->var);
         /* Search child die for local variables and parameters. */
-        if (die_find_variable(sp_die, pf->pvar->var, &vr_die))
+        if (die_find_variable_at(sp_die, pf->pvar->var, pf->addr, &vr_die))
                 ret = convert_variable(&vr_die, pf);
         else {
                 /* Search upper class */
                 nscopes = dwarf_getscopes_die(sp_die, &scopes);
-                if (nscopes > 0) {
-                        ret = dwarf_getscopevar(scopes, nscopes, pf->pvar->var,
-                                                0, NULL, 0, 0, &vr_die);
-                        if (ret >= 0)
+                while (nscopes-- > 1) {
+                        pr_debug("Searching variables in %s\n",
+                                 dwarf_diename(&scopes[nscopes]));
+                        /* We should check this scope, so give dummy address */
+                        if (die_find_variable_at(&scopes[nscopes],
+                                                 pf->pvar->var, 0,
+                                                 &vr_die)) {
                                 ret = convert_variable(&vr_die, pf);
-                        else
-                                ret = -ENOENT;
+                                goto found;
+                        }
+                }
+                if (scopes)
                         free(scopes);
-                } else
-                        ret = -ENOENT;
+                ret = -ENOENT;
         }
+found:
         if (ret < 0)
                 pr_warning("Failed to find '%s' in this function.\n",
                            pf->pvar->var);
         return ret;
 }
 
-/* Show a probe point to output buffer */
-static int convert_probe_point(Dwarf_Die *sp_die, struct probe_finder *pf)
+/* Convert subprogram DIE to trace point */
+static int convert_to_trace_point(Dwarf_Die *sp_die, Dwarf_Addr paddr,
+                                  bool retprobe, struct probe_trace_point *tp)
 {
-        struct probe_trace_event *tev;
         Dwarf_Addr eaddr;
-        Dwarf_Die die_mem;
         const char *name;
-        int ret, i;
-        Dwarf_Attribute fb_attr;
-        size_t nops;
-
-        if (pf->ntevs == pf->max_tevs) {
-                pr_warning("Too many( > %d) probe point found.\n",
-                           pf->max_tevs);
-                return -ERANGE;
-        }
-        tev = &pf->tevs[pf->ntevs++];
-
-        /* If no real subprogram, find a real one */
-        if (!sp_die || dwarf_tag(sp_die) != DW_TAG_subprogram) {
-                sp_die = die_find_real_subprogram(&pf->cu_die,
-                                                 pf->addr, &die_mem);
-                if (!sp_die) {
-                        pr_warning("Failed to find probe point in any "
-                                   "functions.\n");
-                        return -ENOENT;
-                }
-        }
 
         /* Copy the name of probe point */
         name = dwarf_diename(sp_die);
@@ -781,26 +957,45 @@ static int convert_probe_point(Dwarf_Die *sp_die, struct probe_finder *pf)
                                    dwarf_diename(sp_die));
                         return -ENOENT;
                 }
-                tev->point.symbol = strdup(name);
-                if (tev->point.symbol == NULL)
+                tp->symbol = strdup(name);
+                if (tp->symbol == NULL)
                         return -ENOMEM;
-                tev->point.offset = (unsigned long)(pf->addr - eaddr);
+                tp->offset = (unsigned long)(paddr - eaddr);
         } else
                 /* This function has no name. */
-                tev->point.offset = (unsigned long)pf->addr;
+                tp->offset = (unsigned long)paddr;
 
         /* Return probe must be on the head of a subprogram */
-        if (pf->pev->point.retprobe) {
-                if (tev->point.offset != 0) {
+        if (retprobe) {
+                if (eaddr != paddr) {
                         pr_warning("Return probe must be on the head of"
                                    " a real function\n");
                         return -EINVAL;
                 }
-                tev->point.retprobe = true;
+                tp->retprobe = true;
         }
 
-        pr_debug("Probe point found: %s+%lu\n", tev->point.symbol,
-                 tev->point.offset);
+        return 0;
+}
+
+/* Call probe_finder callback with real subprogram DIE */
+static int call_probe_finder(Dwarf_Die *sp_die, struct probe_finder *pf)
+{
+        Dwarf_Die die_mem;
+        Dwarf_Attribute fb_attr;
+        size_t nops;
+        int ret;
+
+        /* If no real subprogram, find a real one */
+        if (!sp_die || dwarf_tag(sp_die) != DW_TAG_subprogram) {
+                sp_die = die_find_real_subprogram(&pf->cu_die,
+                                                  pf->addr, &die_mem);
+                if (!sp_die) {
+                        pr_warning("Failed to find probe point in any "
+                                   "functions.\n");
+                        return -ENOENT;
+                }
+        }
 
         /* Get the frame base attribute/ops */
         dwarf_attr(sp_die, DW_AT_frame_base, &fb_attr);
@@ -820,22 +1015,13 @@ static int convert_probe_point(Dwarf_Die *sp_die, struct probe_finder *pf)
 #endif
         }
 
-        /* Find each argument */
-        tev->nargs = pf->pev->nargs;
-        tev->args = zalloc(sizeof(struct probe_trace_arg) * tev->nargs);
-        if (tev->args == NULL)
-                return -ENOMEM;
-        for (i = 0; i < pf->pev->nargs; i++) {
-                pf->pvar = &pf->pev->args[i];
-                pf->tvar = &tev->args[i];
-                ret = find_variable(sp_die, pf);
-                if (ret != 0)
-                        return ret;
-        }
+        /* Call finder's callback handler */
+        ret = pf->callback(sp_die, pf);
 
         /* *pf->fb_ops will be cached in libdw. Don't free it. */
         pf->fb_ops = NULL;
-        return 0;
+
+        return ret;
 }
 
 /* Find probe point from its line number */
@@ -871,7 +1057,7 @@ static int find_probe_point_by_line(struct probe_finder *pf)
                          (int)i, lineno, (uintmax_t)addr);
                 pf->addr = addr;
 
-                ret = convert_probe_point(NULL, pf);
+                ret = call_probe_finder(NULL, pf);
                 /* Continuing, because target line might be inlined. */
         }
         return ret;
@@ -984,7 +1170,7 @@ static int find_probe_point_lazy(Dwarf_Die *sp_die, struct probe_finder *pf)
                          (int)i, lineno, (unsigned long long)addr);
                 pf->addr = addr;
 
-                ret = convert_probe_point(sp_die, pf);
+                ret = call_probe_finder(sp_die, pf);
                 /* Continuing, because target line might be inlined. */
         }
         /* TODO: deallocate lines, but how? */
@@ -1019,7 +1205,7 @@ static int probe_point_inline_cb(Dwarf_Die *in_die, void *data)
                 pr_debug("found inline addr: 0x%jx\n",
                          (uintmax_t)pf->addr);
 
-                param->retval = convert_probe_point(in_die, pf);
+                param->retval = call_probe_finder(in_die, pf);
                 if (param->retval < 0)
                         return DWARF_CB_ABORT;
         }
@@ -1057,7 +1243,7 @@ static int probe_point_search_cb(Dwarf_Die *sp_die, void *data)
                         }
                         pf->addr += pp->offset;
                         /* TODO: Check the address in this function */
-                        param->retval = convert_probe_point(sp_die, pf);
+                        param->retval = call_probe_finder(sp_die, pf);
                 }
         } else {
                 struct dwarf_callback_param _param = {.data = (void *)pf,
@@ -1079,90 +1265,276 @@ static int find_probe_point_by_func(struct probe_finder *pf)
         return _param.retval;
 }
 
-/* Find probe_trace_events specified by perf_probe_event from debuginfo */
-int find_probe_trace_events(int fd, struct perf_probe_event *pev,
-                             struct probe_trace_event **tevs, int max_tevs)
+/* Find probe points from debuginfo */
+static int find_probes(int fd, struct probe_finder *pf)
 {
-        struct probe_finder pf = {.pev = pev, .max_tevs = max_tevs};
-        struct perf_probe_point *pp = &pev->point;
+        struct perf_probe_point *pp = &pf->pev->point;
         Dwarf_Off off, noff;
         size_t cuhl;
         Dwarf_Die *diep;
-        Dwarf *dbg;
+        Dwarf *dbg = NULL;
+        Dwfl *dwfl;
+        Dwarf_Addr bias;        /* Currently ignored */
         int ret = 0;
 
-        pf.tevs = zalloc(sizeof(struct probe_trace_event) * max_tevs);
-        if (pf.tevs == NULL)
-                return -ENOMEM;
-        *tevs = pf.tevs;
-        pf.ntevs = 0;
-
-        dbg = dwarf_begin(fd, DWARF_C_READ);
+        dbg = dwfl_init_offline_dwarf(fd, &dwfl, &bias);
         if (!dbg) {
                 pr_warning("No dwarf info found in the vmlinux - "
                         "please rebuild with CONFIG_DEBUG_INFO=y.\n");
-                free(pf.tevs);
-                *tevs = NULL;
                 return -EBADF;
         }
 
 #if _ELFUTILS_PREREQ(0, 142)
         /* Get the call frame information from this dwarf */
-        pf.cfi = dwarf_getcfi(dbg);
+        pf->cfi = dwarf_getcfi(dbg);
 #endif
 
         off = 0;
-        line_list__init(&pf.lcache);
+        line_list__init(&pf->lcache);
         /* Loop on CUs (Compilation Unit) */
         while (!dwarf_nextcu(dbg, off, &noff, &cuhl, NULL, NULL, NULL) &&
                ret >= 0) {
                 /* Get the DIE(Debugging Information Entry) of this CU */
-                diep = dwarf_offdie(dbg, off + cuhl, &pf.cu_die);
+                diep = dwarf_offdie(dbg, off + cuhl, &pf->cu_die);
                 if (!diep)
                         continue;
 
                 /* Check if target file is included. */
                 if (pp->file)
-                        pf.fname = cu_find_realpath(&pf.cu_die, pp->file);
+                        pf->fname = cu_find_realpath(&pf->cu_die, pp->file);
                 else
-                        pf.fname = NULL;
+                        pf->fname = NULL;
 
-                if (!pp->file || pf.fname) {
+                if (!pp->file || pf->fname) {
                         if (pp->function)
-                                ret = find_probe_point_by_func(&pf);
+                                ret = find_probe_point_by_func(pf);
                         else if (pp->lazy_line)
-                                ret = find_probe_point_lazy(NULL, &pf);
+                                ret = find_probe_point_lazy(NULL, pf);
                         else {
-                                pf.lno = pp->line;
-                                ret = find_probe_point_by_line(&pf);
+                                pf->lno = pp->line;
+                                ret = find_probe_point_by_line(pf);
                         }
                 }
                 off = noff;
         }
-        line_list__free(&pf.lcache);
-        dwarf_end(dbg);
+        line_list__free(&pf->lcache);
+        if (dwfl)
+                dwfl_end(dwfl);
 
-        return (ret < 0) ? ret : pf.ntevs;
+        return ret;
+}
+
+/* Add a found probe point into trace event list */
+static int add_probe_trace_event(Dwarf_Die *sp_die, struct probe_finder *pf)
+{
+        struct trace_event_finder *tf =
+                        container_of(pf, struct trace_event_finder, pf);
+        struct probe_trace_event *tev;
+        int ret, i;
+
+        /* Check number of tevs */
+        if (tf->ntevs == tf->max_tevs) {
+                pr_warning("Too many( > %d) probe point found.\n",
+                           tf->max_tevs);
+                return -ERANGE;
+        }
+        tev = &tf->tevs[tf->ntevs++];
+
+        ret = convert_to_trace_point(sp_die, pf->addr, pf->pev->point.retprobe,
+                                     &tev->point);
+        if (ret < 0)
+                return ret;
+
+        pr_debug("Probe point found: %s+%lu\n", tev->point.symbol,
+                 tev->point.offset);
+
+        /* Find each argument */
+        tev->nargs = pf->pev->nargs;
+        tev->args = zalloc(sizeof(struct probe_trace_arg) * tev->nargs);
+        if (tev->args == NULL)
+                return -ENOMEM;
+        for (i = 0; i < pf->pev->nargs; i++) {
+                pf->pvar = &pf->pev->args[i];
+                pf->tvar = &tev->args[i];
+                ret = find_variable(sp_die, pf);
+                if (ret != 0)
+                        return ret;
+        }
+
+        return 0;
+}
+
+/* Find probe_trace_events specified by perf_probe_event from debuginfo */
+int find_probe_trace_events(int fd, struct perf_probe_event *pev,
+                            struct probe_trace_event **tevs, int max_tevs)
+{
+        struct trace_event_finder tf = {
+                        .pf = {.pev = pev, .callback = add_probe_trace_event},
+                        .max_tevs = max_tevs};
+        int ret;
+
+        /* Allocate result tevs array */
+        *tevs = zalloc(sizeof(struct probe_trace_event) * max_tevs);
+        if (*tevs == NULL)
+                return -ENOMEM;
+
+        tf.tevs = *tevs;
+        tf.ntevs = 0;
+
+        ret = find_probes(fd, &tf.pf);
+        if (ret < 0) {
+                free(*tevs);
+                *tevs = NULL;
+                return ret;
+        }
+
+        return (ret < 0) ? ret : tf.ntevs;
+}
+
+#define MAX_VAR_LEN 64
+
+/* Collect available variables in this scope */
+static int collect_variables_cb(Dwarf_Die *die_mem, void *data)
+{
+        struct available_var_finder *af = data;
+        struct variable_list *vl;
+        char buf[MAX_VAR_LEN];
+        int tag, ret;
+
+        vl = &af->vls[af->nvls - 1];
+
+        tag = dwarf_tag(die_mem);
+        if (tag == DW_TAG_formal_parameter ||
+            tag == DW_TAG_variable) {
+                ret = convert_variable_location(die_mem, af->pf.addr,
+                                                af->pf.fb_ops, NULL);
+                if (ret == 0) {
+                        ret = die_get_varname(die_mem, buf, MAX_VAR_LEN);
+                        pr_debug2("Add new var: %s\n", buf);
+                        if (ret > 0)
+                                strlist__add(vl->vars, buf);
+                }
+        }
+
+        if (af->child && dwarf_haspc(die_mem, af->pf.addr))
+                return DIE_FIND_CB_CONTINUE;
+        else
+                return DIE_FIND_CB_SIBLING;
+}
+
+/* Add a found vars into available variables list */
+static int add_available_vars(Dwarf_Die *sp_die, struct probe_finder *pf)
+{
+        struct available_var_finder *af =
+                        container_of(pf, struct available_var_finder, pf);
+        struct variable_list *vl;
+        Dwarf_Die die_mem, *scopes = NULL;
+        int ret, nscopes;
+
+        /* Check number of tevs */
+        if (af->nvls == af->max_vls) {
+                pr_warning("Too many( > %d) probe point found.\n", af->max_vls);
+                return -ERANGE;
+        }
+        vl = &af->vls[af->nvls++];
+
+        ret = convert_to_trace_point(sp_die, pf->addr, pf->pev->point.retprobe,
+                                     &vl->point);
+        if (ret < 0)
+                return ret;
+
+        pr_debug("Probe point found: %s+%lu\n", vl->point.symbol,
+                 vl->point.offset);
+
+        /* Find local variables */
+        vl->vars = strlist__new(true, NULL);
+        if (vl->vars == NULL)
+                return -ENOMEM;
+        af->child = true;
+        die_find_child(sp_die, collect_variables_cb, (void *)af, &die_mem);
+
+        /* Find external variables */
+        if (!af->externs)
+                goto out;
+        /* Don't need to search child DIE for externs. */
+        af->child = false;
+        nscopes = dwarf_getscopes_die(sp_die, &scopes);
+        while (nscopes-- > 1)
+                die_find_child(&scopes[nscopes], collect_variables_cb,
+                               (void *)af, &die_mem);
+        if (scopes)
+                free(scopes);
+
+out:
+        if (strlist__empty(vl->vars)) {
+                strlist__delete(vl->vars);
+                vl->vars = NULL;
+        }
+
+        return ret;
+}
+
+/* Find available variables at given probe point */
+int find_available_vars_at(int fd, struct perf_probe_event *pev,
+                           struct variable_list **vls, int max_vls,
+                           bool externs)
+{
+        struct available_var_finder af = {
+                        .pf = {.pev = pev, .callback = add_available_vars},
+                        .max_vls = max_vls, .externs = externs};
+        int ret;
+
+        /* Allocate result vls array */
+        *vls = zalloc(sizeof(struct variable_list) * max_vls);
+        if (*vls == NULL)
+                return -ENOMEM;
+
+        af.vls = *vls;
+        af.nvls = 0;
+
+        ret = find_probes(fd, &af.pf);
+        if (ret < 0) {
+                /* Free vlist for error */
+                while (af.nvls--) {
+                        if (af.vls[af.nvls].point.symbol)
+                                free(af.vls[af.nvls].point.symbol);
+                        if (af.vls[af.nvls].vars)
+                                strlist__delete(af.vls[af.nvls].vars);
+                }
+                free(af.vls);
+                *vls = NULL;
+                return ret;
+        }
+
+        return (ret < 0) ? ret : af.nvls;
 }
 
 /* Reverse search */
-int find_perf_probe_point(int fd, unsigned long addr,
-                          struct perf_probe_point *ppt)
+int find_perf_probe_point(unsigned long addr, struct perf_probe_point *ppt)
 {
         Dwarf_Die cudie, spdie, indie;
-        Dwarf *dbg;
+        Dwarf *dbg = NULL;
+        Dwfl *dwfl = NULL;
         Dwarf_Line *line;
-        Dwarf_Addr laddr, eaddr;
+        Dwarf_Addr laddr, eaddr, bias = 0;
         const char *tmp;
         int lineno, ret = 0;
         bool found = false;
 
-        dbg = dwarf_begin(fd, DWARF_C_READ);
-        if (!dbg)
-                return -EBADF;
+        /* Open the live linux kernel */
+        dbg = dwfl_init_live_kernel_dwarf(addr, &dwfl, &bias);
+        if (!dbg) {
+                pr_warning("No dwarf info found in the vmlinux - "
+                        "please rebuild with CONFIG_DEBUG_INFO=y.\n");
+                ret = -EINVAL;
+                goto end;
+        }
 
+        /* Adjust address with bias */
+        addr += bias;
         /* Find cu die */
-        if (!dwarf_addrdie(dbg, (Dwarf_Addr)addr, &cudie)) {
+        if (!dwarf_addrdie(dbg, (Dwarf_Addr)addr - bias, &cudie)) {
+                pr_warning("No CU DIE is found at %lx\n", addr);
                 ret = -EINVAL;
                 goto end;
         }
@@ -1225,7 +1597,8 @@ found:
         }
 
 end:
-        dwarf_end(dbg);
+        if (dwfl)
+                dwfl_end(dwfl);
         if (ret >= 0)
                 ret = found ? 1 : 0;
         return ret;
@@ -1358,6 +1731,9 @@ static int line_range_search_cb(Dwarf_Die *sp_die, void *data)
         struct line_finder *lf = param->data;
         struct line_range *lr = lf->lr;
 
+        pr_debug("find (%llx) %s\n",
+                 (unsigned long long)dwarf_dieoffset(sp_die),
+                 dwarf_diename(sp_die));
         if (dwarf_tag(sp_die) == DW_TAG_subprogram &&
             die_compare_name(sp_die, lr->function)) {
                 lf->fname = dwarf_decl_file(sp_die);
@@ -1401,10 +1777,12 @@ int find_line_range(int fd, struct line_range *lr)
         Dwarf_Off off = 0, noff;
         size_t cuhl;
         Dwarf_Die *diep;
-        Dwarf *dbg;
+        Dwarf *dbg = NULL;
+        Dwfl *dwfl;
+        Dwarf_Addr bias;        /* Currently ignored */
         const char *comp_dir;
 
-        dbg = dwarf_begin(fd, DWARF_C_READ);
+        dbg = dwfl_init_offline_dwarf(fd, &dwfl, &bias);
         if (!dbg) {
                 pr_warning("No dwarf info found in the vmlinux - "
                         "please rebuild with CONFIG_DEBUG_INFO=y.\n");
@@ -1450,8 +1828,7 @@ int find_line_range(int fd, struct line_range *lr)
         }
 
         pr_debug("path: %s\n", lr->path);
-        dwarf_end(dbg);
-
+        dwfl_end(dwfl);
         return (ret < 0) ? ret : lf.found;
 }
 
diff --git a/tools/perf/util/probe-finder.h b/tools/perf/util/probe-finder.h
index 4507d519f183..bba69d455699 100644
--- a/tools/perf/util/probe-finder.h
+++ b/tools/perf/util/probe-finder.h
@@ -22,20 +22,27 @@ extern int find_probe_trace_events(int fd, struct perf_probe_event *pev,
                                     int max_tevs);
 
 /* Find a perf_probe_point from debuginfo */
-extern int find_perf_probe_point(int fd, unsigned long addr,
+extern int find_perf_probe_point(unsigned long addr,
                                  struct perf_probe_point *ppt);
 
+/* Find a line range */
 extern int find_line_range(int fd, struct line_range *lr);
 
+/* Find available variables */
+extern int find_available_vars_at(int fd, struct perf_probe_event *pev,
+                                  struct variable_list **vls, int max_points,
+                                  bool externs);
+
 #include <dwarf.h>
 #include <libdw.h>
+#include <libdwfl.h>
 #include <version.h>
 
 struct probe_finder {
         struct perf_probe_event *pev;           /* Target probe event */
-        struct probe_trace_event *tevs;         /* Result trace events */
-        int                     ntevs;          /* Number of trace events */
-        int                     max_tevs;       /* Max number of trace events */
+
+        /* Callback when a probe point is found */
+        int (*callback)(Dwarf_Die *sp_die, struct probe_finder *pf);
 
         /* For function searching */
         int                     lno;            /* Line number */
@@ -53,6 +60,22 @@ struct probe_finder {
         struct probe_trace_arg  *tvar;          /* Current result variable */
 };
 
+struct trace_event_finder {
+        struct probe_finder     pf;
+        struct probe_trace_event *tevs;         /* Found trace events */
+        int                     ntevs;          /* Number of trace events */
+        int                     max_tevs;       /* Max number of trace events */
+};
+
+struct available_var_finder {
+        struct probe_finder     pf;
+        struct variable_list    *vls;           /* Found variable lists */
+        int                     nvls;           /* Number of variable lists */
+        int                     max_vls;        /* Max no. of variable lists */
+        bool                    externs;        /* Find external vars too */
+        bool                    child;          /* Search child scopes */
+};
+
 struct line_finder {
         struct line_range       *lr;            /* Target line range */
 
diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h
index 46e531d09e8b..0b91053a7d11 100644
--- a/tools/perf/util/sort.h
+++ b/tools/perf/util/sort.h
@@ -70,7 +70,7 @@ struct hist_entry {
                 struct hist_entry *pair;
                 struct rb_root    sorted_chain;
         };
-        struct callchain_node   callchain[0];
+        struct callchain_root   callchain[0];
 };
 
 enum sort_type {
diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index b2f5ae97f33d..b39f499e575a 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -388,6 +388,20 @@ size_t dso__fprintf_buildid(struct dso *self, FILE *fp)
         return fprintf(fp, "%s", sbuild_id);
 }
 
+size_t dso__fprintf_symbols_by_name(struct dso *self, enum map_type type, FILE *fp)
+{
+        size_t ret = 0;
+        struct rb_node *nd;
+        struct symbol_name_rb_node *pos;
+
+        for (nd = rb_first(&self->symbol_names[type]); nd; nd = rb_next(nd)) {
+                pos = rb_entry(nd, struct symbol_name_rb_node, rb_node);
+                fprintf(fp, "%s\n", pos->sym.name);
+        }
+
+        return ret;
+}
+
 size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp)
 {
         struct rb_node *nd;
diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h
index ea95c2756f05..038f2201ee09 100644
--- a/tools/perf/util/symbol.h
+++ b/tools/perf/util/symbol.h
@@ -182,6 +182,7 @@ size_t machines__fprintf_dsos(struct rb_root *self, FILE *fp);
 size_t machines__fprintf_dsos_buildid(struct rb_root *self, FILE *fp, bool with_hits);
 
 size_t dso__fprintf_buildid(struct dso *self, FILE *fp);
+size_t dso__fprintf_symbols_by_name(struct dso *self, enum map_type type, FILE *fp);
 size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp);
 
 enum dso_origin {
diff --git a/tools/perf/util/ui/browser.c b/tools/perf/util/ui/browser.c
index 66f2d583d8c4..8bc010edca25 100644
--- a/tools/perf/util/ui/browser.c
+++ b/tools/perf/util/ui/browser.c
@@ -1,16 +1,5 @@
-#define _GNU_SOURCE
-#include <stdio.h>
-#undef _GNU_SOURCE
-/*
- * slang versions <= 2.0.6 have a "#if HAVE_LONG_LONG" that breaks
- * the build if it isn't defined. Use the equivalent one that glibc
- * has on features.h.
- */
-#include <features.h>
-#ifndef HAVE_LONG_LONG
-#define HAVE_LONG_LONG __GLIBC_HAVE_LONG_LONG
-#endif
-#include <slang.h>
+#include "libslang.h"
+#include <linux/compiler.h>
 #include <linux/list.h>
 #include <linux/rbtree.h>
 #include <stdlib.h>
@@ -19,17 +8,9 @@
 #include "helpline.h"
 #include "../color.h"
 #include "../util.h"
+#include <stdio.h>
 
-#if SLANG_VERSION < 20104
-#define sltt_set_color(obj, name, fg, bg) \
-        SLtt_set_color(obj,(char *)name, (char *)fg, (char *)bg)
-#else
-#define sltt_set_color SLtt_set_color
-#endif
-
-newtComponent newt_form__new(void);
-
-int ui_browser__percent_color(double percent, bool current)
+static int ui_browser__percent_color(double percent, bool current)
 {
         if (current)
                 return HE_COLORSET_SELECTED;
@@ -40,6 +21,23 @@ int ui_browser__percent_color(double percent, bool current)
         return HE_COLORSET_NORMAL;
 }
 
+void ui_browser__set_color(struct ui_browser *self __used, int color)
+{
+        SLsmg_set_color(color);
+}
+
+void ui_browser__set_percent_color(struct ui_browser *self,
+                                   double percent, bool current)
+{
+         int color = ui_browser__percent_color(percent, current);
+         ui_browser__set_color(self, color);
+}
+
+void ui_browser__gotorc(struct ui_browser *self, int y, int x)
+{
+        SLsmg_gotorc(self->y + y, self->x + x);
+}
+
 void ui_browser__list_head_seek(struct ui_browser *self, off_t offset, int whence)
 {
         struct list_head *head = self->entries;
@@ -111,7 +109,7 @@ unsigned int ui_browser__rb_tree_refresh(struct ui_browser *self)
         nd = self->top;
 
         while (nd != NULL) {
-                SLsmg_gotorc(self->y + row, self->x);
+                ui_browser__gotorc(self, row, 0);
                 self->write(self, nd, row);
                 if (++row == self->height)
                         break;
@@ -131,13 +129,10 @@ void ui_browser__refresh_dimensions(struct ui_browser *self)
         int cols, rows;
         newtGetScreenSize(&cols, &rows);
 
-        if (self->width > cols - 4)
-                self->width = cols - 4;
-        self->height = rows - 5;
-        if (self->height > self->nr_entries)
-                self->height = self->nr_entries;
-        self->y  = (rows - self->height) / 2;
-        self->x = (cols - self->width) / 2;
+        self->width = cols - 1;
+        self->height = rows - 2;
+        self->y = 1;
+        self->x = 0;
 }
 
 void ui_browser__reset_index(struct ui_browser *self)
@@ -146,34 +141,48 @@ void ui_browser__reset_index(struct ui_browser *self)
         self->seek(self, 0, SEEK_SET);
 }
 
+void ui_browser__add_exit_key(struct ui_browser *self, int key)
+{
+        newtFormAddHotKey(self->form, key);
+}
+
+void ui_browser__add_exit_keys(struct ui_browser *self, int keys[])
+{
+        int i = 0;
+
+        while (keys[i] && i < 64) {
+                ui_browser__add_exit_key(self, keys[i]);
+                ++i;
+        }
+}
+
 int ui_browser__show(struct ui_browser *self, const char *title,
                      const char *helpline, ...)
 {
         va_list ap;
+        int keys[] = { NEWT_KEY_UP, NEWT_KEY_DOWN, NEWT_KEY_PGUP,
+                       NEWT_KEY_PGDN, NEWT_KEY_HOME, NEWT_KEY_END, ' ',
+                       NEWT_KEY_LEFT, NEWT_KEY_ESCAPE, 'q', CTRL('c'), 0 };
 
-        if (self->form != NULL) {
+        if (self->form != NULL)
                 newtFormDestroy(self->form);
-                newtPopWindow();
-        }
+
         ui_browser__refresh_dimensions(self);
-        newtCenteredWindow(self->width, self->height, title);
-        self->form = newt_form__new();
+        self->form = newtForm(NULL, NULL, 0);
         if (self->form == NULL)
                 return -1;
 
-        self->sb = newtVerticalScrollbar(self->width, 0, self->height,
+        self->sb = newtVerticalScrollbar(self->width, 1, self->height,
                                          HE_COLORSET_NORMAL,
                                          HE_COLORSET_SELECTED);
         if (self->sb == NULL)
                 return -1;
 
-        newtFormAddHotKey(self->form, NEWT_KEY_UP);
-        newtFormAddHotKey(self->form, NEWT_KEY_DOWN);
-        newtFormAddHotKey(self->form, NEWT_KEY_PGUP);
-        newtFormAddHotKey(self->form, NEWT_KEY_PGDN);
-        newtFormAddHotKey(self->form, NEWT_KEY_HOME);
-        newtFormAddHotKey(self->form, NEWT_KEY_END);
-        newtFormAddHotKey(self->form, ' ');
+        SLsmg_gotorc(0, 0);
+        ui_browser__set_color(self, NEWT_COLORSET_ROOT);
+        slsmg_write_nstring(title, self->width);
+
+        ui_browser__add_exit_keys(self, keys);
         newtFormAddComponent(self->form, self->sb);
 
         va_start(ap, helpline);
@@ -185,7 +194,6 @@ int ui_browser__show(struct ui_browser *self, const char *title,
 void ui_browser__hide(struct ui_browser *self)
 {
         newtFormDestroy(self->form);
-        newtPopWindow();
         self->form = NULL;
         ui_helpline__pop();
 }
@@ -196,28 +204,28 @@ int ui_browser__refresh(struct ui_browser *self)
 
         newtScrollbarSet(self->sb, self->index, self->nr_entries - 1);
         row = self->refresh(self);
-        SLsmg_set_color(HE_COLORSET_NORMAL);
+        ui_browser__set_color(self, HE_COLORSET_NORMAL);
         SLsmg_fill_region(self->y + row, self->x,
                           self->height - row, self->width, ' ');
 
         return 0;
 }
 
-int ui_browser__run(struct ui_browser *self, struct newtExitStruct *es)
+int ui_browser__run(struct ui_browser *self)
 {
+        struct newtExitStruct es;
+
         if (ui_browser__refresh(self) < 0)
                 return -1;
 
         while (1) {
                 off_t offset;
 
-                newtFormRun(self->form, es);
+                newtFormRun(self->form, &es);
 
-                if (es->reason != NEWT_EXIT_HOTKEY)
+                if (es.reason != NEWT_EXIT_HOTKEY)
                         break;
-                if (is_exit_key(es->u.key))
-                        return es->u.key;
-                switch (es->u.key) {
+                switch (es.u.key) {
                 case NEWT_KEY_DOWN:
                         if (self->index == self->nr_entries - 1)
                                 break;
@@ -274,12 +282,12 @@ int ui_browser__run(struct ui_browser *self, struct newtExitStruct *es)
                         self->seek(self, -offset, SEEK_END);
                         break;
                 default:
-                        return es->u.key;
+                        return es.u.key;
                 }
                 if (ui_browser__refresh(self) < 0)
                         return -1;
         }
-        return 0;
+        return -1;
 }
 
 unsigned int ui_browser__list_head_refresh(struct ui_browser *self)
@@ -294,7 +302,7 @@ unsigned int ui_browser__list_head_refresh(struct ui_browser *self)
         pos = self->top;
 
         list_for_each_from(pos, head) {
-                SLsmg_gotorc(self->y + row, self->x);
+                ui_browser__gotorc(self, row, 0);
                 self->write(self, pos, row);
                 if (++row == self->height)
                         break;
diff --git a/tools/perf/util/ui/browser.h b/tools/perf/util/ui/browser.h
index 0b9f829214f7..0dc7e4da36f5 100644
--- a/tools/perf/util/ui/browser.h
+++ b/tools/perf/util/ui/browser.h
@@ -25,16 +25,21 @@ struct ui_browser {
 };
 
 
-int ui_browser__percent_color(double percent, bool current);
+void ui_browser__set_color(struct ui_browser *self, int color);
+void ui_browser__set_percent_color(struct ui_browser *self,
+                                   double percent, bool current);
 bool ui_browser__is_current_entry(struct ui_browser *self, unsigned row);
 void ui_browser__refresh_dimensions(struct ui_browser *self);
 void ui_browser__reset_index(struct ui_browser *self);
 
+void ui_browser__gotorc(struct ui_browser *self, int y, int x);
+void ui_browser__add_exit_key(struct ui_browser *self, int key);
+void ui_browser__add_exit_keys(struct ui_browser *self, int keys[]);
 int ui_browser__show(struct ui_browser *self, const char *title,
                      const char *helpline, ...);
 void ui_browser__hide(struct ui_browser *self);
 int ui_browser__refresh(struct ui_browser *self);
-int ui_browser__run(struct ui_browser *self, struct newtExitStruct *es);
+int ui_browser__run(struct ui_browser *self);
 
 void ui_browser__rb_tree_seek(struct ui_browser *self, off_t offset, int whence);
 unsigned int ui_browser__rb_tree_refresh(struct ui_browser *self);
diff --git a/tools/perf/util/ui/browsers/annotate.c b/tools/perf/util/ui/browsers/annotate.c
index a90273e63f4f..82b78f99251b 100644
--- a/tools/perf/util/ui/browsers/annotate.c
+++ b/tools/perf/util/ui/browsers/annotate.c
@@ -40,14 +40,12 @@ static void annotate_browser__write(struct ui_browser *self, void *entry, int ro
 
         if (ol->offset != -1) {
                 struct objdump_line_rb_node *olrb = objdump_line__rb(ol);
-                int color = ui_browser__percent_color(olrb->percent, current_entry);
-                SLsmg_set_color(color);
+                ui_browser__set_percent_color(self, olrb->percent, current_entry);
                 slsmg_printf(" %7.2f ", olrb->percent);
                 if (!current_entry)
-                        SLsmg_set_color(HE_COLORSET_CODE);
+                        ui_browser__set_color(self, HE_COLORSET_CODE);
         } else {
-                int color = ui_browser__percent_color(0, current_entry);
-                SLsmg_set_color(color);
+                ui_browser__set_percent_color(self, 0, current_entry);
                 slsmg_write_nstring(" ", 9);
         }
 
@@ -135,32 +133,31 @@ static void annotate_browser__set_top(struct annotate_browser *self,
         self->curr_hot = nd;
 }
 
-static int annotate_browser__run(struct annotate_browser *self,
-                                 struct newtExitStruct *es)
+static int annotate_browser__run(struct annotate_browser *self)
 {
         struct rb_node *nd;
         struct hist_entry *he = self->b.priv;
+        int key;
 
         if (ui_browser__show(&self->b, he->ms.sym->name,
-                             "<- or ESC: exit, TAB/shift+TAB: cycle thru samples") < 0)
+                             "<-, -> or ESC: exit, TAB/shift+TAB: cycle thru samples") < 0)
                 return -1;
-
-        newtFormAddHotKey(self->b.form, NEWT_KEY_LEFT);
-        newtFormAddHotKey(self->b.form, NEWT_KEY_RIGHT);
+        /*
+         * To allow builtin-annotate to cycle thru multiple symbols by
+         * examining the exit key for this function.
+         */
+        ui_browser__add_exit_key(&self->b, NEWT_KEY_RIGHT);
 
         nd = self->curr_hot;
         if (nd) {
-                newtFormAddHotKey(self->b.form, NEWT_KEY_TAB);
-                newtFormAddHotKey(self->b.form, NEWT_KEY_UNTAB);
+                int tabs[] = { NEWT_KEY_TAB, NEWT_KEY_UNTAB, 0 };
+                ui_browser__add_exit_keys(&self->b, tabs);
         }
 
         while (1) {
-                ui_browser__run(&self->b, es);
-
-                if (es->reason != NEWT_EXIT_HOTKEY)
-                        break;
+                key = ui_browser__run(&self->b);
 
-                switch (es->u.key) {
+                switch (key) {
                 case NEWT_KEY_TAB:
                         nd = rb_prev(nd);
                         if (nd == NULL)
@@ -179,12 +176,11 @@ static int annotate_browser__run(struct annotate_browser *self,
         }
 out:
         ui_browser__hide(&self->b);
-        return es->u.key;
+        return key;
 }
 
 int hist_entry__tui_annotate(struct hist_entry *self)
 {
-        struct newtExitStruct es;
         struct objdump_line *pos, *n;
         struct objdump_line_rb_node *rbpos;
         LIST_HEAD(head);
@@ -232,7 +228,7 @@ int hist_entry__tui_annotate(struct hist_entry *self)
                 annotate_browser__set_top(&browser, browser.curr_hot);
 
         browser.b.width += 18; /* Percentage */
-        ret = annotate_browser__run(&browser, &es);
+        ret = annotate_browser__run(&browser);
         list_for_each_entry_safe(pos, n, &head, node) {
                 list_del(&pos->node);
                 objdump_line__free(pos);
diff --git a/tools/perf/util/ui/browsers/hists.c b/tools/perf/util/ui/browsers/hists.c
index 6866aa4c41e0..ebda8c3fde9e 100644
--- a/tools/perf/util/ui/browsers/hists.c
+++ b/tools/perf/util/ui/browsers/hists.c
@@ -58,6 +58,11 @@ static char callchain_list__folded(const struct callchain_list *self)
         return map_symbol__folded(&self->ms);
 }
 
+static void map_symbol__set_folding(struct map_symbol *self, bool unfold)
+{
+        self->unfolded = unfold ? self->has_children : false;
+}
+
 static int callchain_node__count_rows_rb_tree(struct callchain_node *self)
 {
         int n = 0;
@@ -129,16 +134,16 @@ static void callchain_node__init_have_children_rb_tree(struct callchain_node *se
         for (nd = rb_first(&self->rb_root); nd; nd = rb_next(nd)) {
                 struct callchain_node *child = rb_entry(nd, struct callchain_node, rb_node);
                 struct callchain_list *chain;
-                int first = true;
+                bool first = true;
 
                 list_for_each_entry(chain, &child->val, list) {
                         if (first) {
                                 first = false;
                                 chain->ms.has_children = chain->list.next != &child->val ||
-                                                         rb_first(&child->rb_root) != NULL;
+                                                         !RB_EMPTY_ROOT(&child->rb_root);
                         } else
                                 chain->ms.has_children = chain->list.next == &child->val &&
-                                                         rb_first(&child->rb_root) != NULL;
+                                                         !RB_EMPTY_ROOT(&child->rb_root);
                 }
 
                 callchain_node__init_have_children_rb_tree(child);
@@ -150,7 +155,7 @@ static void callchain_node__init_have_children(struct callchain_node *self)
         struct callchain_list *chain;
 
         list_for_each_entry(chain, &self->val, list)
-                chain->ms.has_children = rb_first(&self->rb_root) != NULL;
+                chain->ms.has_children = !RB_EMPTY_ROOT(&self->rb_root);
 
         callchain_node__init_have_children_rb_tree(self);
 }
@@ -168,6 +173,7 @@ static void callchain__init_have_children(struct rb_root *self)
 static void hist_entry__init_have_children(struct hist_entry *self)
 {
         if (!self->init_have_children) {
+                self->ms.has_children = !RB_EMPTY_ROOT(&self->sorted_chain);
                 callchain__init_have_children(&self->sorted_chain);
                 self->init_have_children = true;
         }
@@ -195,43 +201,114 @@ static bool hist_browser__toggle_fold(struct hist_browser *self)
         return false;
 }
 
-static int hist_browser__run(struct hist_browser *self, const char *title,
-                             struct newtExitStruct *es)
+static int callchain_node__set_folding_rb_tree(struct callchain_node *self, bool unfold)
+{
+        int n = 0;
+        struct rb_node *nd;
+
+        for (nd = rb_first(&self->rb_root); nd; nd = rb_next(nd)) {
+                struct callchain_node *child = rb_entry(nd, struct callchain_node, rb_node);
+                struct callchain_list *chain;
+                bool has_children = false;
+
+                list_for_each_entry(chain, &child->val, list) {
+                        ++n;
+                        map_symbol__set_folding(&chain->ms, unfold);
+                        has_children = chain->ms.has_children;
+                }
+
+                if (has_children)
+                        n += callchain_node__set_folding_rb_tree(child, unfold);
+        }
+
+        return n;
+}
+
+static int callchain_node__set_folding(struct callchain_node *node, bool unfold)
+{
+        struct callchain_list *chain;
+        bool has_children = false;
+        int n = 0;
+
+        list_for_each_entry(chain, &node->val, list) {
+                ++n;
+                map_symbol__set_folding(&chain->ms, unfold);
+                has_children = chain->ms.has_children;
+        }
+
+        if (has_children)
+                n += callchain_node__set_folding_rb_tree(node, unfold);
+
+        return n;
+}
+
+static int callchain__set_folding(struct rb_root *chain, bool unfold)
+{
+        struct rb_node *nd;
+        int n = 0;
+
+        for (nd = rb_first(chain); nd; nd = rb_next(nd)) {
+                struct callchain_node *node = rb_entry(nd, struct callchain_node, rb_node);
+                n += callchain_node__set_folding(node, unfold);
+        }
+
+        return n;
+}
+
+static void hist_entry__set_folding(struct hist_entry *self, bool unfold)
+{
+        hist_entry__init_have_children(self);
+        map_symbol__set_folding(&self->ms, unfold);
+
+        if (self->ms.has_children) {
+                int n = callchain__set_folding(&self->sorted_chain, unfold);
+                self->nr_rows = unfold ? n : 0;
+        } else
+                self->nr_rows = 0;
+}
+
+static void hists__set_folding(struct hists *self, bool unfold)
+{
+        struct rb_node *nd;
+
+        self->nr_entries = 0;
+
+        for (nd = rb_first(&self->entries); nd; nd = rb_next(nd)) {
+                struct hist_entry *he = rb_entry(nd, struct hist_entry, rb_node);
+                hist_entry__set_folding(he, unfold);
+                self->nr_entries += 1 + he->nr_rows;
+        }
+}
+
+static void hist_browser__set_folding(struct hist_browser *self, bool unfold)
+{
+        hists__set_folding(self->hists, unfold);
+        self->b.nr_entries = self->hists->nr_entries;
+        /* Go to the start, we may be way after valid entries after a collapse */
+        ui_browser__reset_index(&self->b);
+}
+
+static int hist_browser__run(struct hist_browser *self, const char *title)
 {
-        char str[256], unit;
-        unsigned long nr_events = self->hists->stats.nr_events[PERF_RECORD_SAMPLE];
+        int key;
+        int exit_keys[] = { 'a', '?', 'h', 'C', 'd', 'D', 'E', 't',
+                            NEWT_KEY_ENTER, NEWT_KEY_RIGHT, NEWT_KEY_LEFT, 0, };
 
         self->b.entries = &self->hists->entries;
         self->b.nr_entries = self->hists->nr_entries;
 
         hist_browser__refresh_dimensions(self);
 
-        nr_events = convert_unit(nr_events, &unit);
-        snprintf(str, sizeof(str), "Events: %lu%c                            ",
-                 nr_events, unit);
-        newtDrawRootText(0, 0, str);
-
         if (ui_browser__show(&self->b, title,
                              "Press '?' for help on key bindings") < 0)
                 return -1;
 
-        newtFormAddHotKey(self->b.form, 'a');
-        newtFormAddHotKey(self->b.form, '?');
-        newtFormAddHotKey(self->b.form, 'h');
-        newtFormAddHotKey(self->b.form, 'd');
-        newtFormAddHotKey(self->b.form, 'D');
-        newtFormAddHotKey(self->b.form, 't');
-
-        newtFormAddHotKey(self->b.form, NEWT_KEY_LEFT);
-        newtFormAddHotKey(self->b.form, NEWT_KEY_RIGHT);
-        newtFormAddHotKey(self->b.form, NEWT_KEY_ENTER);
+        ui_browser__add_exit_keys(&self->b, exit_keys);
 
         while (1) {
-                ui_browser__run(&self->b, es);
+                key = ui_browser__run(&self->b);
 
-                if (es->reason != NEWT_EXIT_HOTKEY)
-                        break;
-                switch (es->u.key) {
+                switch (key) {
                 case 'D': { /* Debug */
                         static int seq;
                         struct hist_entry *h = rb_entry(self->b.top,
@@ -245,18 +322,26 @@ static int hist_browser__run(struct hist_browser *self, const char *title,
                                            self->b.top_idx,
                                            h->row_offset, h->nr_rows);
                 }
-                        continue;
+                        break;
+                case 'C':
+                        /* Collapse the whole world. */
+                        hist_browser__set_folding(self, false);
+                        break;
+                case 'E':
+                        /* Expand the whole world. */
+                        hist_browser__set_folding(self, true);
+                        break;
                 case NEWT_KEY_ENTER:
                         if (hist_browser__toggle_fold(self))
                                 break;
                         /* fall thru */
                 default:
-                        return 0;
+                        goto out;
                 }
         }
-
+out:
         ui_browser__hide(&self->b);
-        return 0;
+        return key;
 }
 
 static char *callchain_list__sym_name(struct callchain_list *self,
@@ -306,15 +391,10 @@ static int hist_browser__show_callchain_node_rb_tree(struct hist_browser *self,
                         int color;
                         bool was_first = first;
 
-                        if (first) {
+                        if (first)
                                 first = false;
-                                chain->ms.has_children = chain->list.next != &child->val ||
-                                                         rb_first(&child->rb_root) != NULL;
-                        } else {
+                        else
                                 extra_offset = LEVEL_OFFSET_STEP;
-                                chain->ms.has_children = chain->list.next == &child->val &&
-                                                         rb_first(&child->rb_root) != NULL;
-                        }
 
                         folded_sign = callchain_list__folded(chain);
                         if (*row_offset != 0) {
@@ -341,8 +421,8 @@ static int hist_browser__show_callchain_node_rb_tree(struct hist_browser *self,
                                 *is_current_entry = true;
                         }
 
-                        SLsmg_set_color(color);
-                        SLsmg_gotorc(self->b.y + row, self->b.x);
+                        ui_browser__set_color(&self->b, color);
+                        ui_browser__gotorc(&self->b, row, 0);
                         slsmg_write_nstring(" ", offset + extra_offset);
                         slsmg_printf("%c ", folded_sign);
                         slsmg_write_nstring(str, width);
@@ -384,12 +464,7 @@ static int hist_browser__show_callchain_node(struct hist_browser *self,
         list_for_each_entry(chain, &node->val, list) {
                 char ipstr[BITS_PER_LONG / 4 + 1], *s;
                 int color;
-                /*
-                 * FIXME: This should be moved to somewhere else,
-                 * probably when the callchain is created, so as not to
-                 * traverse it all over again
-                 */
-                chain->ms.has_children = rb_first(&node->rb_root) != NULL;
+
                 folded_sign = callchain_list__folded(chain);
 
                 if (*row_offset != 0) {
@@ -405,8 +480,8 @@ static int hist_browser__show_callchain_node(struct hist_browser *self,
                 }
 
                 s = callchain_list__sym_name(chain, ipstr, sizeof(ipstr));
-                SLsmg_gotorc(self->b.y + row, self->b.x);
-                SLsmg_set_color(color);
+                ui_browser__gotorc(&self->b, row, 0);
+                ui_browser__set_color(&self->b, color);
                 slsmg_write_nstring(" ", offset);
                 slsmg_printf("%c ", folded_sign);
                 slsmg_write_nstring(s, width - 2);
@@ -465,7 +540,7 @@ static int hist_browser__show_entry(struct hist_browser *self,
         }
 
         if (symbol_conf.use_callchain) {
-                entry->ms.has_children = !RB_EMPTY_ROOT(&entry->sorted_chain);
+                hist_entry__init_have_children(entry);
                 folded_sign = hist_entry__folded(entry);
         }
 
@@ -484,8 +559,8 @@ static int hist_browser__show_entry(struct hist_browser *self,
                                 color = HE_COLORSET_NORMAL;
                 }
 
-                SLsmg_set_color(color);
-                SLsmg_gotorc(self->b.y + row, self->b.x);
+                ui_browser__set_color(&self->b, color);
+                ui_browser__gotorc(&self->b, row, 0);
                 if (symbol_conf.use_callchain) {
                         slsmg_printf("%c ", folded_sign);
                         width -= 2;
@@ -687,8 +762,6 @@ static struct hist_browser *hist_browser__new(struct hists *hists)
 
 static void hist_browser__delete(struct hist_browser *self)
 {
-        newtFormDestroy(self->b.form);
-        newtPopWindow();
         free(self);
 }
 
@@ -702,21 +775,26 @@ static struct thread *hist_browser__selected_thread(struct hist_browser *self)
         return self->he_selection->thread;
 }
 
-static int hist_browser__title(char *bf, size_t size, const char *ev_name,
-                               const struct dso *dso, const struct thread *thread)
+static int hists__browser_title(struct hists *self, char *bf, size_t size,
+                                const char *ev_name, const struct dso *dso,
+                                const struct thread *thread)
 {
-        int printed = 0;
+        char unit;
+        int printed;
+        unsigned long nr_events = self->stats.nr_events[PERF_RECORD_SAMPLE];
+
+        nr_events = convert_unit(nr_events, &unit);
+        printed = snprintf(bf, size, "Events: %lu%c %s", nr_events, unit, ev_name);
 
         if (thread)
                 printed += snprintf(bf + printed, size - printed,
-                                    "Thread: %s(%d)",
-                                    (thread->comm_set ?  thread->comm : ""),
+                                    ", Thread: %s(%d)",
+                                    (thread->comm_set ? thread->comm : ""),
                                     thread->pid);
         if (dso)
                 printed += snprintf(bf + printed, size - printed,
-                                    "%sDSO: %s", thread ? " " : "",
-                                    dso->short_name);
-        return printed ?: snprintf(bf, size, "Event: %s", ev_name);
+                                    ", DSO: %s", dso->short_name);
+        return printed;
 }
 
 int hists__browse(struct hists *self, const char *helpline, const char *ev_name)
@@ -725,7 +803,6 @@ int hists__browse(struct hists *self, const char *helpline, const char *ev_name)
         struct pstack *fstack;
         const struct thread *thread_filter = NULL;
         const struct dso *dso_filter = NULL;
-        struct newtExitStruct es;
         char msg[160];
         int key = -1;
 
@@ -738,9 +815,8 @@ int hists__browse(struct hists *self, const char *helpline, const char *ev_name)
 
         ui_helpline__push(helpline);
 
-        hist_browser__title(msg, sizeof(msg), ev_name,
-                            dso_filter, thread_filter);
-
+        hists__browser_title(self, msg, sizeof(msg), ev_name,
+                             dso_filter, thread_filter);
         while (1) {
                 const struct thread *thread;
                 const struct dso *dso;
@@ -749,70 +825,63 @@ int hists__browse(struct hists *self, const char *helpline, const char *ev_name)
                     annotate = -2, zoom_dso = -2, zoom_thread = -2,
                     browse_map = -2;
 
-                if (hist_browser__run(browser, msg, &es))
-                        break;
+                key = hist_browser__run(browser, msg);
 
                 thread = hist_browser__selected_thread(browser);
                 dso = browser->selection->map ? browser->selection->map->dso : NULL;
 
-                if (es.reason == NEWT_EXIT_HOTKEY) {
-                        key = es.u.key;
-
-                        switch (key) {
-                        case NEWT_KEY_F1:
-                                goto do_help;
-                        case NEWT_KEY_TAB:
-                        case NEWT_KEY_UNTAB:
-                                /*
-                                 * Exit the browser, let hists__browser_tree
-                                 * go to the next or previous
-                                 */
-                                goto out_free_stack;
-                        default:;
-                        }
-
-                        switch (key) {
-                        case 'a':
-                                if (browser->selection->map == NULL ||
-                                    browser->selection->map->dso->annotate_warned)
-                                        continue;
-                                goto do_annotate;
-                        case 'd':
-                                goto zoom_dso;
-                        case 't':
-                                goto zoom_thread;
-                        case 'h':
-                        case '?':
-do_help:
-                                ui__help_window("->        Zoom into DSO/Threads & Annotate current symbol\n"
-                                                "<-        Zoom out\n"
-                                                "a         Annotate current symbol\n"
-                                                "h/?/F1    Show this window\n"
-                                                "d         Zoom into current DSO\n"
-                                                "t         Zoom into current Thread\n"
-                                                "q/CTRL+C  Exit browser");
+                switch (key) {
+                case NEWT_KEY_TAB:
+                case NEWT_KEY_UNTAB:
+                        /*
+                         * Exit the browser, let hists__browser_tree
+                         * go to the next or previous
+                         */
+                        goto out_free_stack;
+                case 'a':
+                        if (browser->selection->map == NULL &&
+                            browser->selection->map->dso->annotate_warned)
                                 continue;
-                        default:;
-                        }
-                        if (is_exit_key(key)) {
-                                if (key == NEWT_KEY_ESCAPE &&
-                                    !ui__dialog_yesno("Do you really want to exit?"))
-                                        continue;
-                                break;
-                        }
-
-                        if (es.u.key == NEWT_KEY_LEFT) {
-                                const void *top;
+                        goto do_annotate;
+                case 'd':
+                        goto zoom_dso;
+                case 't':
+                        goto zoom_thread;
+                case NEWT_KEY_F1:
+                case 'h':
+                case '?':
+                        ui__help_window("->        Zoom into DSO/Threads & Annotate current symbol\n"
+                                        "<-        Zoom out\n"
+                                        "a         Annotate current symbol\n"
+                                        "h/?/F1    Show this window\n"
+                                        "C         Collapse all callchains\n"
+                                        "E         Expand all callchains\n"
+                                        "d         Zoom into current DSO\n"
+                                        "t         Zoom into current Thread\n"
+                                        "q/CTRL+C  Exit browser");
+                        continue;
+                case NEWT_KEY_ENTER:
+                case NEWT_KEY_RIGHT:
+                        /* menu */
+                        break;
+                case NEWT_KEY_LEFT: {
+                        const void *top;
 
-                                if (pstack__empty(fstack))
-                                        continue;
-                                top = pstack__pop(fstack);
-                                if (top == &dso_filter)
-                                        goto zoom_out_dso;
-                                if (top == &thread_filter)
-                                        goto zoom_out_thread;
+                        if (pstack__empty(fstack))
                                 continue;
-                        }
+                        top = pstack__pop(fstack);
+                        if (top == &dso_filter)
+                                goto zoom_out_dso;
+                        if (top == &thread_filter)
+                                goto zoom_out_thread;
+                        continue;
+                }
+                case NEWT_KEY_ESCAPE:
+                        if (!ui__dialog_yesno("Do you really want to exit?"))
+                                continue;
+                        /* Fall thru */
+                default:
+                        goto out_free_stack;
                 }
 
                 if (browser->selection->sym != NULL &&
@@ -885,8 +954,8 @@ zoom_out_dso:
                                 pstack__push(fstack, &dso_filter);
                         }
                         hists__filter_by_dso(self, dso_filter);
-                        hist_browser__title(msg, sizeof(msg), ev_name,
-                                            dso_filter, thread_filter);
+                        hists__browser_title(self, msg, sizeof(msg), ev_name,
+                                             dso_filter, thread_filter);
                         hist_browser__reset(browser);
                 } else if (choice == zoom_thread) {
 zoom_thread:
@@ -903,8 +972,8 @@ zoom_out_thread:
                                 pstack__push(fstack, &thread_filter);
                         }
                         hists__filter_by_thread(self, thread_filter);
-                        hist_browser__title(msg, sizeof(msg), ev_name,
-                                            dso_filter, thread_filter);
+                        hists__browser_title(self, msg, sizeof(msg), ev_name,
+                                             dso_filter, thread_filter);
                         hist_browser__reset(browser);
                 }
         }
@@ -925,10 +994,6 @@ int hists__tui_browse_tree(struct rb_root *self, const char *help)
                 const char *ev_name = __event_name(hists->type, hists->config);
 
                 key = hists__browse(hists, help, ev_name);
-
-                if (is_exit_key(key))
-                        break;
-
                 switch (key) {
                 case NEWT_KEY_TAB:
                         next = rb_next(nd);
@@ -940,7 +1005,7 @@ int hists__tui_browse_tree(struct rb_root *self, const char *help)
                                 continue;
                         nd = rb_prev(nd);
                 default:
-                        break;
+                        return key;
                 }
         }
 
diff --git a/tools/perf/util/ui/browsers/map.c b/tools/perf/util/ui/browsers/map.c
index 142b825b42bf..e35437dfa5b4 100644
--- a/tools/perf/util/ui/browsers/map.c
+++ b/tools/perf/util/ui/browsers/map.c
@@ -1,6 +1,5 @@
 #include "../libslang.h"
 #include <elf.h>
-#include <newt.h>
 #include <sys/ttydefaults.h>
 #include <ctype.h>
 #include <string.h>
@@ -47,7 +46,6 @@ out_free_form:
 struct map_browser {
         struct ui_browser b;
         struct map        *map;
-        u16               namelen;
         u8                addrlen;
 };
 
@@ -56,14 +54,16 @@ static void map_browser__write(struct ui_browser *self, void *nd, int row)
         struct symbol *sym = rb_entry(nd, struct symbol, rb_node);
         struct map_browser *mb = container_of(self, struct map_browser, b);
         bool current_entry = ui_browser__is_current_entry(self, row);
-        int color = ui_browser__percent_color(0, current_entry);
+        int width;
 
-        SLsmg_set_color(color);
+        ui_browser__set_percent_color(self, 0, current_entry);
         slsmg_printf("%*llx %*llx %c ",
                      mb->addrlen, sym->start, mb->addrlen, sym->end,
                      sym->binding == STB_GLOBAL ? 'g' :
                      sym->binding == STB_LOCAL  ? 'l' : 'w');
-        slsmg_write_nstring(sym->name, mb->namelen);
+        width = self->width - ((mb->addrlen * 2) + 4);
+        if (width > 0)
+                slsmg_write_nstring(sym->name, width);
 }
 
 /* FIXME uber-kludgy, see comment on cmd_report... */
@@ -98,31 +98,29 @@ static int map_browser__search(struct map_browser *self)
         return 0;
 }
 
-static int map_browser__run(struct map_browser *self, struct newtExitStruct *es)
+static int map_browser__run(struct map_browser *self)
 {
+        int key;
+
         if (ui_browser__show(&self->b, self->map->dso->long_name,
                              "Press <- or ESC to exit, %s / to search",
                              verbose ? "" : "restart with -v to use") < 0)
                 return -1;
 
-        newtFormAddHotKey(self->b.form, NEWT_KEY_LEFT);
-        newtFormAddHotKey(self->b.form, NEWT_KEY_ENTER);
         if (verbose)
-                newtFormAddHotKey(self->b.form, '/');
+                ui_browser__add_exit_key(&self->b, '/');
 
         while (1) {
-                ui_browser__run(&self->b, es);
+                key = ui_browser__run(&self->b);
 
-                if (es->reason != NEWT_EXIT_HOTKEY)
-                        break;
-                if (verbose && es->u.key == '/')
+                if (verbose && key == '/')
                         map_browser__search(self);
                 else
                         break;
         }
 
         ui_browser__hide(&self->b);
-        return 0;
+        return key;
 }
 
 int map__browse(struct map *self)
@@ -136,7 +134,6 @@ int map__browse(struct map *self)
                 },
                 .map = self,
         };
-        struct newtExitStruct es;
         struct rb_node *nd;
         char tmp[BITS_PER_LONG / 4];
         u64 maxaddr = 0;
@@ -144,8 +141,6 @@ int map__browse(struct map *self)
         for (nd = rb_first(mb.b.entries); nd; nd = rb_next(nd)) {
                 struct symbol *pos = rb_entry(nd, struct symbol, rb_node);
 
-                if (mb.namelen < pos->namelen)
-                        mb.namelen = pos->namelen;
                 if (maxaddr < pos->end)
                         maxaddr = pos->end;
                 if (verbose) {
@@ -156,6 +151,5 @@ int map__browse(struct map *self)
         }
 
         mb.addrlen = snprintf(tmp, sizeof(tmp), "%llx", maxaddr);
-        mb.b.width += mb.addrlen * 2 + 4 + mb.namelen;
-        return map_browser__run(&mb, &es);
+        return map_browser__run(&mb);
 }
diff --git a/tools/perf/util/ui/util.c b/tools/perf/util/ui/util.c
index 04600e26ceea..9706d9d40279 100644
--- a/tools/perf/util/ui/util.c
+++ b/tools/perf/util/ui/util.c
@@ -11,8 +11,6 @@
 #include "helpline.h"
 #include "util.h"
 
-newtComponent newt_form__new(void);
-
 static void newt_form__set_exit_keys(newtComponent self)
 {
         newtFormAddHotKey(self, NEWT_KEY_LEFT);
@@ -22,7 +20,7 @@ static void newt_form__set_exit_keys(newtComponent self)
         newtFormAddHotKey(self, CTRL('c'));
 }
 
-newtComponent newt_form__new(void)
+static newtComponent newt_form__new(void)
 {
         newtComponent self = newtForm(NULL, NULL, 0);
         if (self)
diff --git a/tools/perf/util/util.h b/tools/perf/util/util.h
index f380fed74359..7562707ddd1c 100644
--- a/tools/perf/util/util.h
+++ b/tools/perf/util/util.h
@@ -266,19 +266,6 @@ bool strglobmatch(const char *str, const char *pat);
 bool strlazymatch(const char *str, const char *pat);
 unsigned long convert_unit(unsigned long value, char *unit);
 
-#ifndef ESC
-#define ESC 27
-#endif
-
-static inline bool is_exit_key(int key)
-{
-        char up;
-        if (key == CTRL('c') || key == ESC)
-                return true;
-        up = toupper(key);
-        return up == 'Q';
-}
-
 #define _STR(x) #x
 #define STR(x) _STR(x)