Re: [PATCH v4 2/7] perf/sdt: Directly record SDT events with 'perf record'

2017-03-08 Thread Masami Hiramatsu 
Hi Ravi,

On Mon,  6 Mar 2017 18:42:59 +0530
Ravi Bangoria  wrote:

> diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
> index bc84a37..e87b19b 100644
> --- a/tools/perf/builtin-record.c
> +++ b/tools/perf/builtin-record.c
> @@ -39,6 +39,7 @@
>  #include "util/trigger.h"
>  #include "util/perf-hooks.h"
>  #include "asm/bug.h"
> +#include "util/probe-file.h"
>  
>  #include 
>  #include 
> @@ -73,6 +74,7 @@ struct record {
>   booltimestamp_filename;
>   struct switch_outputswitch_output;
>   unsigned long long  samples;
> + struct list_headsdt_event_list;
>  };
>  
>  static volatile int auxtrace_record__snapshot_started;
> @@ -1503,6 +1505,26 @@ static struct record record = {
>   },
>  };
>  
> +void sdt_event_list__add(struct list_head *sdt_event_list)
> +{
> + if (list_empty(sdt_event_list))
> + return;
> + list_splice(sdt_event_list, _event_list);
> +}
> +
> +bool is_cmd_record(void)
> +{
> + return (record.evlist != NULL);
> +}

Hmm, exporting this from builtin-record.c makes inverted reference from
builtin-record.o to libperf. I think it is not good way to do.

[...]
> diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
> index 54355d3..1fcc9d13 100644
> --- a/tools/perf/util/parse-events.c
> +++ b/tools/perf/util/parse-events.c
> @@ -1727,12 +1727,66 @@ static void parse_events_print_error(struct 
> parse_events_error *err,
>  
>  #undef MAX_WIDTH
>  
> +/* SDT event needs LIBELF support for creating a probe point */
> +#ifdef HAVE_LIBELF_SUPPORT
> +static int parse_sdt_event(struct perf_evlist *evlist, const char *str,
> +struct parse_events_error *err)
> +{
> + char *ptr = NULL;
> + int ret;
> + struct list_head *sdt_evlist;
> + struct sdt_event_list *sdt_event;
> +
> + if (str[0] == '%')
> + str++;
> +
> + ptr = strdup(str);
> + if (ptr == NULL)
> + return -ENOMEM;
> +
> + sdt_evlist = zalloc(sizeof(*sdt_evlist));
> + if (!sdt_evlist) {
> + free(ptr);
> + pr_debug("Error in sdt_evlist memory allocation\n");
> + return -ENOMEM;
> + }
> + INIT_LIST_HEAD(sdt_evlist);
> +
> + /*
> +  * If there is an error in this call, no need to free
> +  * up sdt_evlist, its already free'ed up in the previous
> +  * call. Free up 'ptr' though.
> +  */
> + ret = add_sdt_event(ptr, sdt_evlist);
> + if (!ret) {
> + list_for_each_entry(sdt_event, sdt_evlist, list) {
> + ret = parse_events(evlist, sdt_event->name, err);
> + if (ret < 0)
> + goto ret;
> + }
> + /* Add it to the record struct */
> + sdt_event_list__add(sdt_evlist);
> + }
> +
> +ret:
> + free(ptr);
> + return ret;
> +}
> +#endif /* HAVE_LIBELF_SUPPORT */
> +
>  int parse_events_option(const struct option *opt, const char *str,
>   int unset __maybe_unused)
>  {
>   struct perf_evlist *evlist = *(struct perf_evlist **)opt->value;
>   struct parse_events_error err = { .idx = 0, };
> - int ret = parse_events(evlist, str, );
> + int ret = 0;
> +
> +#ifdef HAVE_LIBELF_SUPPORT
> + if (is_sdt_event((char *)str) && is_cmd_record())
> + ret = parse_sdt_event(evlist, str, );
> + else
> +#endif

To avoid these ifdefs, could you add a dummy (error return) function
of add_sdt_event() for !HAVE_LIBELF_SUPPORT case in probe-event.h?



> + ret = parse_events(evlist, str, );
>  
>   if (ret)
>   parse_events_print_error(, str);
> diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h
> index c6172cd..0887269 100644
> --- a/tools/perf/util/parse-events.h
> +++ b/tools/perf/util/parse-events.h
> @@ -208,4 +208,6 @@ static inline bool is_sdt_event(char *str)
>   (!strncmp(str, "sdt_", 4) &&
>!!strchr(str, ':') && !strchr(str, '=')));
>  }
> +
> +void sdt_event_list__add(struct list_head *sdt_event_list);
>  #endif /* __PERF_PARSE_EVENTS_H */
> diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
> index 2b1409f..b879076 100644
> --- a/tools/perf/util/probe-event.c
> +++ b/tools/perf/util/probe-event.c
> @@ -1293,7 +1293,7 @@ int parse_line_range_desc(const char *arg, struct 
> line_range *lr)
>   return err;
>  }
>  
> -static int parse_perf_probe_event_name(char **arg, struct perf_probe_event 
> *pev)
> +int parse_perf_probe_event_name(char **arg, struct perf_probe_event *pev)
>  {
>   char *ptr;
>  
> @@ -3125,8 +3125,8 @@ static int find_cached_events(struct perf_probe_event 
> *pev,
>  }
>  
>  /* Try to find probe_trace_event from all probe caches */
> -static int find_cached_events_all(struct perf_probe_event *pev,
> -struct

Re: [PATCH v4 2/7] perf/sdt: Directly record SDT events with 'perf record'

2017-03-08 Thread Masami Hiramatsu 
Hi Ravi,

On Mon,  6 Mar 2017 18:42:59 +0530
Ravi Bangoria  wrote:

> diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
> index bc84a37..e87b19b 100644
> --- a/tools/perf/builtin-record.c
> +++ b/tools/perf/builtin-record.c
> @@ -39,6 +39,7 @@
>  #include "util/trigger.h"
>  #include "util/perf-hooks.h"
>  #include "asm/bug.h"
> +#include "util/probe-file.h"
>  
>  #include 
>  #include 
> @@ -73,6 +74,7 @@ struct record {
>   booltimestamp_filename;
>   struct switch_outputswitch_output;
>   unsigned long long  samples;
> + struct list_headsdt_event_list;
>  };
>  
>  static volatile int auxtrace_record__snapshot_started;
> @@ -1503,6 +1505,26 @@ static struct record record = {
>   },
>  };
>  
> +void sdt_event_list__add(struct list_head *sdt_event_list)
> +{
> + if (list_empty(sdt_event_list))
> + return;
> + list_splice(sdt_event_list, _event_list);
> +}
> +
> +bool is_cmd_record(void)
> +{
> + return (record.evlist != NULL);
> +}

Hmm, exporting this from builtin-record.c makes inverted reference from
builtin-record.o to libperf. I think it is not good way to do.

[...]
> diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
> index 54355d3..1fcc9d13 100644
> --- a/tools/perf/util/parse-events.c
> +++ b/tools/perf/util/parse-events.c
> @@ -1727,12 +1727,66 @@ static void parse_events_print_error(struct 
> parse_events_error *err,
>  
>  #undef MAX_WIDTH
>  
> +/* SDT event needs LIBELF support for creating a probe point */
> +#ifdef HAVE_LIBELF_SUPPORT
> +static int parse_sdt_event(struct perf_evlist *evlist, const char *str,
> +struct parse_events_error *err)
> +{
> + char *ptr = NULL;
> + int ret;
> + struct list_head *sdt_evlist;
> + struct sdt_event_list *sdt_event;
> +
> + if (str[0] == '%')
> + str++;
> +
> + ptr = strdup(str);
> + if (ptr == NULL)
> + return -ENOMEM;
> +
> + sdt_evlist = zalloc(sizeof(*sdt_evlist));
> + if (!sdt_evlist) {
> + free(ptr);
> + pr_debug("Error in sdt_evlist memory allocation\n");
> + return -ENOMEM;
> + }
> + INIT_LIST_HEAD(sdt_evlist);
> +
> + /*
> +  * If there is an error in this call, no need to free
> +  * up sdt_evlist, its already free'ed up in the previous
> +  * call. Free up 'ptr' though.
> +  */
> + ret = add_sdt_event(ptr, sdt_evlist);
> + if (!ret) {
> + list_for_each_entry(sdt_event, sdt_evlist, list) {
> + ret = parse_events(evlist, sdt_event->name, err);
> + if (ret < 0)
> + goto ret;
> + }
> + /* Add it to the record struct */
> + sdt_event_list__add(sdt_evlist);
> + }
> +
> +ret:
> + free(ptr);
> + return ret;
> +}
> +#endif /* HAVE_LIBELF_SUPPORT */
> +
>  int parse_events_option(const struct option *opt, const char *str,
>   int unset __maybe_unused)
>  {
>   struct perf_evlist *evlist = *(struct perf_evlist **)opt->value;
>   struct parse_events_error err = { .idx = 0, };
> - int ret = parse_events(evlist, str, );
> + int ret = 0;
> +
> +#ifdef HAVE_LIBELF_SUPPORT
> + if (is_sdt_event((char *)str) && is_cmd_record())
> + ret = parse_sdt_event(evlist, str, );
> + else
> +#endif

To avoid these ifdefs, could you add a dummy (error return) function
of add_sdt_event() for !HAVE_LIBELF_SUPPORT case in probe-event.h?



> + ret = parse_events(evlist, str, );
>  
>   if (ret)
>   parse_events_print_error(, str);
> diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h
> index c6172cd..0887269 100644
> --- a/tools/perf/util/parse-events.h
> +++ b/tools/perf/util/parse-events.h
> @@ -208,4 +208,6 @@ static inline bool is_sdt_event(char *str)
>   (!strncmp(str, "sdt_", 4) &&
>!!strchr(str, ':') && !strchr(str, '=')));
>  }
> +
> +void sdt_event_list__add(struct list_head *sdt_event_list);
>  #endif /* __PERF_PARSE_EVENTS_H */
> diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
> index 2b1409f..b879076 100644
> --- a/tools/perf/util/probe-event.c
> +++ b/tools/perf/util/probe-event.c
> @@ -1293,7 +1293,7 @@ int parse_line_range_desc(const char *arg, struct 
> line_range *lr)
>   return err;
>  }
>  
> -static int parse_perf_probe_event_name(char **arg, struct perf_probe_event 
> *pev)
> +int parse_perf_probe_event_name(char **arg, struct perf_probe_event *pev)
>  {
>   char *ptr;
>  
> @@ -3125,8 +3125,8 @@ static int find_cached_events(struct perf_probe_event 
> *pev,
>  }
>  
>  /* Try to find probe_trace_event from all probe caches */
> -static int find_cached_events_all(struct perf_probe_event *pev,
> -struct probe_trace_event **tevs)
> +int

[PATCH v4 2/7] perf/sdt: Directly record SDT events with 'perf record'

2017-03-06 Thread Ravi Bangoria 
From: Hemant Kumar 

Add basic support for directly recording SDT events which are present
in the probe cache. Without this patch, we could probe into SDT events
using 'perf probe' and 'perf record'. With this patch, we can probe the
SDT events directly using 'perf record'.

For example :

  $ sudo ./perf list sdt
sdt_libpthread:mutex_entry [SDT event]
sdt_libc:setjmp[SDT event]

  $ sudo ./perf record -a -e sdt_libc:setjmp

  $ sudo ./perf script
 bash   793 [002]   260.382957: sdt_libc:setjmp: (7ff85b6596a1)
reset  1296 [000]   260.511983: sdt_libc:setjmp: (7f26862e06a1)

Recording on SDT events with same provider and marker names is also
supported:

  $ readelf -n /usr/lib64/libpthread-2.24.so | grep -A2 Provider
  Provider: libpthread
  Name: mutex_entry
  Location: 0x9ddb, Base: 0x000139cc, ...
--
  Provider: libpthread
  Name: mutex_entry
  Location: 0xbcbb, Base: 0x000139cc, ...

  $ sudo ./perf record -a -e sdt_libpthread:mutex_entry
Warning : Recording on 2 occurences of sdt_libpthread:mutex_entry

  $ sudo ./perf evlist
sdt_libpthread:mutex_entry_1
sdt_libpthread:mutex_entry

After invoking 'perf record', behind the scenes, it checks whether the
event specified is an SDT event using the string 'sdt_' or flag '%'.
After that, it does a lookup of the probe cache to find out the SDT
event. If its not present, it throws an error. Otherwise, it goes on
and writes the event into the uprobe_events file and starts recording.
It also maintains a list of the event names that were written to
uprobe_events file. At the end of the record session, it removes the
events from the uprobe_events file using the maintained name list.

As mentioned, it always tries to look for sdt event in probe cache and
ignores entries of uprobe_events. Hence, it creates new probe points
for event even if it already exists.

  $ sudo ./perf probe sdt_libpthread:mutex_entry
Added new events:
  sdt_libpthread:mutex_entry (on %mutex_entry in 
/usr/lib64/libpthread-2.24.so)
  sdt_libpthread:mutex_entry_1 (on %mutex_entry in 
/usr/lib64/libpthread-2.24.so)

  $ sudo ./perf record -a -e sdt_libpthread:mutex_entry
Warning : Recording on 2 occurences of sdt_libpthread:mutex_entry

  $ sudo ./perf evlist
sdt_libpthread:mutex_entry_3
sdt_libpthread:mutex_entry_2

As it does not look at uprobe_events file, it can't record those events
whose probe points are created with different name. For ex,

  $ sudo ./perf record -a -e sdt_libpthread:mutex_entry_1
Error: sdt_libpthread:mutex_entry_1 not found in the cache
invalid or unsupported event: 'sdt_libpthread:mutex_entry_1'

Signed-off-by: Hemant Kumar 
Signed-off-by: Ravi Bangoria 
---
 tools/perf/builtin-record.c|  24 
 tools/perf/perf.h  |   1 +
 tools/perf/util/parse-events.c |  56 +-
 tools/perf/util/parse-events.h |   2 +
 tools/perf/util/probe-event.c  |  35 ++-
 tools/perf/util/probe-event.h  |   4 ++
 tools/perf/util/probe-file.c   | 131 +
 tools/perf/util/probe-file.h   |   8 +++
 8 files changed, 257 insertions(+), 4 deletions(-)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index bc84a37..e87b19b 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -39,6 +39,7 @@
 #include "util/trigger.h"
 #include "util/perf-hooks.h"
 #include "asm/bug.h"
+#include "util/probe-file.h"
 
 #include 
 #include 
@@ -73,6 +74,7 @@ struct record {
booltimestamp_filename;
struct switch_outputswitch_output;
unsigned long long  samples;
+   struct list_headsdt_event_list;
 };
 
 static volatile int auxtrace_record__snapshot_started;
@@ -1503,6 +1505,26 @@ static struct record record = {
},
 };
 
+void sdt_event_list__add(struct list_head *sdt_event_list)
+{
+   if (list_empty(sdt_event_list))
+   return;
+   list_splice(sdt_event_list, _event_list);
+}
+
+bool is_cmd_record(void)
+{
+   return (record.evlist != NULL);
+}
+
+static void
+sdt_event_list__remove(struct list_head *sdt_event_list __maybe_unused)
+{
+#ifdef HAVE_LIBELF_SUPPORT
+   return remove_sdt_event_list(sdt_event_list);
+#endif
+}
+
 const char record_callchain_help[] = CALLCHAIN_RECORD_HELP
"\n\t\t\t\tDefault: fp";
 
@@ -1671,6 +1693,7 @@ int cmd_record(int argc, const char **argv, const char 
*prefix __maybe_unused)
if (rec->evlist == NULL)
return -ENOMEM;
 
+   INIT_LIST_HEAD(>sdt_event_list);
err = perf_config(perf_record_config, rec);
if (err)
return err;
@@ -1841,6 +1864,7 @@ int cmd_record(int argc, const char **argv, const char 
*prefix __maybe_unused)

[PATCH v4 2/7] perf/sdt: Directly record SDT events with 'perf record'

2017-03-06 Thread Ravi Bangoria 
From: Hemant Kumar 

Add basic support for directly recording SDT events which are present
in the probe cache. Without this patch, we could probe into SDT events
using 'perf probe' and 'perf record'. With this patch, we can probe the
SDT events directly using 'perf record'.

For example :

  $ sudo ./perf list sdt
sdt_libpthread:mutex_entry [SDT event]
sdt_libc:setjmp[SDT event]

  $ sudo ./perf record -a -e sdt_libc:setjmp

  $ sudo ./perf script
 bash   793 [002]   260.382957: sdt_libc:setjmp: (7ff85b6596a1)
reset  1296 [000]   260.511983: sdt_libc:setjmp: (7f26862e06a1)

Recording on SDT events with same provider and marker names is also
supported:

  $ readelf -n /usr/lib64/libpthread-2.24.so | grep -A2 Provider
  Provider: libpthread
  Name: mutex_entry
  Location: 0x9ddb, Base: 0x000139cc, ...
--
  Provider: libpthread
  Name: mutex_entry
  Location: 0xbcbb, Base: 0x000139cc, ...

  $ sudo ./perf record -a -e sdt_libpthread:mutex_entry
Warning : Recording on 2 occurences of sdt_libpthread:mutex_entry

  $ sudo ./perf evlist
sdt_libpthread:mutex_entry_1
sdt_libpthread:mutex_entry

After invoking 'perf record', behind the scenes, it checks whether the
event specified is an SDT event using the string 'sdt_' or flag '%'.
After that, it does a lookup of the probe cache to find out the SDT
event. If its not present, it throws an error. Otherwise, it goes on
and writes the event into the uprobe_events file and starts recording.
It also maintains a list of the event names that were written to
uprobe_events file. At the end of the record session, it removes the
events from the uprobe_events file using the maintained name list.

As mentioned, it always tries to look for sdt event in probe cache and
ignores entries of uprobe_events. Hence, it creates new probe points
for event even if it already exists.

  $ sudo ./perf probe sdt_libpthread:mutex_entry
Added new events:
  sdt_libpthread:mutex_entry (on %mutex_entry in 
/usr/lib64/libpthread-2.24.so)
  sdt_libpthread:mutex_entry_1 (on %mutex_entry in 
/usr/lib64/libpthread-2.24.so)

  $ sudo ./perf record -a -e sdt_libpthread:mutex_entry
Warning : Recording on 2 occurences of sdt_libpthread:mutex_entry

  $ sudo ./perf evlist
sdt_libpthread:mutex_entry_3
sdt_libpthread:mutex_entry_2

As it does not look at uprobe_events file, it can't record those events
whose probe points are created with different name. For ex,

  $ sudo ./perf record -a -e sdt_libpthread:mutex_entry_1
Error: sdt_libpthread:mutex_entry_1 not found in the cache
invalid or unsupported event: 'sdt_libpthread:mutex_entry_1'

Signed-off-by: Hemant Kumar 
Signed-off-by: Ravi Bangoria 
---
 tools/perf/builtin-record.c|  24 
 tools/perf/perf.h  |   1 +
 tools/perf/util/parse-events.c |  56 +-
 tools/perf/util/parse-events.h |   2 +
 tools/perf/util/probe-event.c  |  35 ++-
 tools/perf/util/probe-event.h  |   4 ++
 tools/perf/util/probe-file.c   | 131 +
 tools/perf/util/probe-file.h   |   8 +++
 8 files changed, 257 insertions(+), 4 deletions(-)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index bc84a37..e87b19b 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -39,6 +39,7 @@
 #include "util/trigger.h"
 #include "util/perf-hooks.h"
 #include "asm/bug.h"
+#include "util/probe-file.h"
 
 #include 
 #include 
@@ -73,6 +74,7 @@ struct record {
booltimestamp_filename;
struct switch_outputswitch_output;
unsigned long long  samples;
+   struct list_headsdt_event_list;
 };
 
 static volatile int auxtrace_record__snapshot_started;
@@ -1503,6 +1505,26 @@ static struct record record = {
},
 };
 
+void sdt_event_list__add(struct list_head *sdt_event_list)
+{
+   if (list_empty(sdt_event_list))
+   return;
+   list_splice(sdt_event_list, _event_list);
+}
+
+bool is_cmd_record(void)
+{
+   return (record.evlist != NULL);
+}
+
+static void
+sdt_event_list__remove(struct list_head *sdt_event_list __maybe_unused)
+{
+#ifdef HAVE_LIBELF_SUPPORT
+   return remove_sdt_event_list(sdt_event_list);
+#endif
+}
+
 const char record_callchain_help[] = CALLCHAIN_RECORD_HELP
"\n\t\t\t\tDefault: fp";
 
@@ -1671,6 +1693,7 @@ int cmd_record(int argc, const char **argv, const char 
*prefix __maybe_unused)
if (rec->evlist == NULL)
return -ENOMEM;
 
+   INIT_LIST_HEAD(>sdt_event_list);
err = perf_config(perf_record_config, rec);
if (err)
return err;
@@ -1841,6 +1864,7 @@ int cmd_record(int argc, const char **argv, const char 
*prefix __maybe_unused)
perf_evlist__delete(rec->evlist);
symbol__exit();