Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1292262 > unrolled thread

[PATCHSET 00/10] perf tools: Support dynamic sort keys for tracepoints (v2)

Back to article view | Back to linux.kernel

Contents

  [PATCHSET 00/10] perf tools: Support dynamic sort keys for tracepoints (v2) Namhyung Kim <namhyung@kernel.org> - 2015-12-15 16:40 +0100
    [PATCH 02/10] perf hist: Save raw_data/size for tracepoint events Namhyung Kim <namhyung@kernel.org> - 2015-12-15 16:40 +0100
      [PATCH v2.1] perf hist: Save raw_data/size for tracepoint events Namhyung Kim <namhyung@kernel.org> - 2015-12-17 09:20 +0100
        Re: [PATCH v2.1] perf hist: Save raw_data/size for tracepoint events Arnaldo Carvalho de Melo <acme@kernel.org> - 2015-12-17 13:30 +0100
    [PATCH 01/10] perf hist: Pass struct sample to __hists__add_entry() Namhyung Kim <namhyung@kernel.org> - 2015-12-15 16:40 +0100
    [PATCH 03/10] tools lib traceevent: Factor out and export print_event_field[s] Namhyung Kim <namhyung@kernel.org> - 2015-12-15 16:40 +0100
    [PATCH 06/10] perf tools: Try to show pretty printed output for dynamic sort keys Namhyung Kim <namhyung@kernel.org> - 2015-12-15 16:40 +0100
      Re: [PATCH 06/10] perf tools: Try to show pretty printed output for  dynamic sort keys Jiri Olsa <jolsa@redhat.com> - 2015-12-20 15:20 +0100
        Re: [PATCH 06/10] perf tools: Try to show pretty printed output for  dynamic sort keys Namhyung Kim <namhyung@kernel.org> - 2015-12-21 09:40 +0100
    [PATCH 10/10] perf tools: Support shortcuts for events in dynamic sort keys Namhyung Kim <namhyung@kernel.org> - 2015-12-15 16:40 +0100
    [PATCH 07/10] perf tools: Add 'trace' sort key Namhyung Kim <namhyung@kernel.org> - 2015-12-15 16:40 +0100
    [PATCH 08/10] perf tools: Add --raw-trace option Namhyung Kim <namhyung@kernel.org> - 2015-12-15 16:40 +0100
      Re: [PATCH 08/10] perf tools: Add --raw-trace option Jiri Olsa <jolsa@redhat.com> - 2015-12-20 16:00 +0100
        Re: [PATCH 08/10] perf tools: Add --raw-trace option Namhyung Kim <namhyung@kernel.org> - 2015-12-21 09:50 +0100
          Re: [PATCH 08/10] perf tools: Add --raw-trace option Jiri Olsa <jolsa@redhat.com> - 2015-12-22 08:00 +0100
            Re: [PATCH 08/10] perf tools: Add --raw-trace option Namhyung Kim <namhyung@kernel.org> - 2015-12-22 17:30 +0100
    [PATCH 04/10] perf tools: Pass evlist to setup_sorting() Namhyung Kim <namhyung@kernel.org> - 2015-12-15 16:40 +0100
    [PATCH 05/10] perf tools: Add dynamic sort key for tracepoint events Namhyung Kim <namhyung@kernel.org> - 2015-12-15 16:40 +0100
      Re: [PATCH 05/10] perf tools: Add dynamic sort key for tracepoint  events Jiri Olsa <jolsa@redhat.com> - 2015-12-20 15:00 +0100
        Re: [PATCH 05/10] perf tools: Add dynamic sort key for tracepoint  events Namhyung Kim <namhyung@kernel.org> - 2015-12-20 15:10 +0100
    [PATCH 09/10] perf tools: Make 'trace' sort key default for tracepoint events Namhyung Kim <namhyung@kernel.org> - 2015-12-15 16:40 +0100
    Re: [PATCHSET 00/10] perf tools: Support dynamic sort keys for  tracepoints (v2) Arnaldo Carvalho de Melo <acme@kernel.org> - 2015-12-17 01:20 +0100
      Re: [PATCHSET 00/10] perf tools: Support dynamic sort keys for  tracepoints (v2) Namhyung Kim <namhyung@kernel.org> - 2015-12-17 09:00 +0100
        Re: [PATCHSET 00/10] perf tools: Support dynamic sort keys for  tracepoints (v2) Arnaldo Carvalho de Melo <acme@kernel.org> - 2015-12-17 13:30 +0100

Page 2 of 2 — ← Prev page 1 [2]

#1292282 — [PATCH 09/10] perf tools: Make 'trace' sort key default for tracepoint events

When an evlist contains tracepoint events only, use 'trace' sort key as
default.  This will make users more convenient to see trace result.

Suggested-by: Jiri Olsa <jolsa@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/util/sort.c | 22 +++++++++++++++++-----
 1 file changed, 17 insertions(+), 5 deletions(-)

diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c
index da65b07258bb..e661d3dd3ef3 100644
--- a/tools/perf/util/sort.c
+++ b/tools/perf/util/sort.c
@@ -2049,7 +2049,7 @@ static int sort_dimension__add(const char *tok,
 	return -ESRCH;
 }
 
-static const char *get_default_sort_order(void)
+static const char *get_default_sort_order(struct perf_evlist *evlist)
 {
 	const char *default_sort_orders[] = {
 		default_sort_order,
@@ -2058,13 +2058,25 @@ static const char *get_default_sort_order(void)
 		default_top_sort_order,
 		default_diff_sort_order,
 	};
+	bool use_trace = true;
+	struct perf_evsel *evsel;
 
 	BUG_ON(sort__mode >= ARRAY_SIZE(default_sort_orders));
 
+	evlist__for_each(evlist, evsel) {
+		if (evsel->attr.type != PERF_TYPE_TRACEPOINT) {
+			use_trace = false;
+			break;
+		}
+	}
+
+	if (use_trace)
+		return "trace";
+
 	return default_sort_orders[sort__mode];
 }
 
-static int setup_sort_order(void)
+static int setup_sort_order(struct perf_evlist *evlist)
 {
 	char *new_sort_order;
 
@@ -2085,7 +2097,7 @@ static int setup_sort_order(void)
 	 * because it's checked over the rest of the code.
 	 */
 	if (asprintf(&new_sort_order, "%s,%s",
-		     get_default_sort_order(), sort_order + 1) < 0) {
+		     get_default_sort_order(evlist), sort_order + 1) < 0) {
 		error("Not enough memory to set up --sort");
 		return -ENOMEM;
 	}
@@ -2100,7 +2112,7 @@ static int __setup_sorting(struct perf_evlist *evlist)
 	const char *sort_keys;
 	int ret = 0;
 
-	ret = setup_sort_order();
+	ret = setup_sort_order(evlist);
 	if (ret)
 		return ret;
 
@@ -2114,7 +2126,7 @@ static int __setup_sorting(struct perf_evlist *evlist)
 			return 0;
 		}
 
-		sort_keys = get_default_sort_order();
+		sort_keys = get_default_sort_order(evlist);
 	}
 
 	str = strdup(sort_keys);
-- 
2.6.4

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1293417 — Re: [PATCHSET 00/10] perf tools: Support dynamic sort keys for tracepoints (v2)

Em Wed, Dec 16, 2015 at 12:35:33AM +0900, Namhyung Kim escreveu:
> Hello,
> 
> This is an attempt to improve perf to deal with tracepoint events
> better.  The perf tools can handle tracepoint events but perf report
> on them is less useful since they're always sampled in a fixed
> location and not provide event specific info.  We can use perf script
> but I always wishes there's more convenient way to see the result.
> 
>  * changes in v2)
>   - add 'trace' sort key and make it default  (Jiri)
>   - add '--raw-trace' option and '/raw' field modifier  (Jiri)
>   - support event name shortcuts  (David)

Can you take a look if you can reproduce this? Without callchains it works in
all tests I did.

[root@zoo ~]# perf record -g -e kmem:kmalloc -a
^C[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 1.193 MB perf.data (250 samples) ]

[root@zoo ~]# perf report
perf: Segmentation fault
-------- backtrace --------
perf[0x539f1b]
/lib64/libc.so.6(+0x34960)[0x7fc8e3752960]
perf(pevent_read_number+0x78)[0x542e40]
perf(pevent_read_number_field+0x70)[0x542ed3]
/root/.traceevent/plugins/plugin_kmem.so(+0x603)[0x7fc8e1a6c603]
perf(pevent_event_info+0x96)[0x547597]
perf[0x4dce26]
perf[0x4e0e89]
perf(hist_entry_iter__add+0xee)[0x4e125e]
perf[0x4304be]
perf[0x4c2db3]
perf[0x4c3301]
perf[0x4c6089]
perf(perf_session__process_events+0x3f1)[0x4c4b91]
perf(cmd_report+0x120b)[0x4319cb]
perf[0x47c871]
perf(main+0x63f)[0x42242f]
/lib64/libc.so.6(__libc_start_main+0xf0)[0x7fc8e373dfe0]
perf[0x422549]
[0x0]
[root@zoo ~]# ls -la ~/.traceevent/
total 12
drwxr-xr-x.  3 acme acme 4096 May  4  2015 .
drwx------. 61 acme acme 4096 Dec 16 21:12 ..
drwxr-xr-x.  2 acme acme 4096 Nov 23 11:15 plugins
[root@zoo ~]# ls -la ~/.traceevent
lrwxrwxrwx. 1 root root 23 May  4  2015 /root/.traceevent -> /home/acme/.traceevent/
[root@zoo ~]# perf record -e sched:sched_switch -a


^C[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 1.704 MB perf.data (4967 samples) ]

[root@zoo ~]# perf report
[root@zoo ~]# perf record -g -e sched:sched_switch -a
^C[ perf record: Woken up 2 times to write data ]
[ perf record: Captured and wrote 2.062 MB perf.data (4743 samples) ]

[root@zoo ~]# perf report
perf: Segmentation fault
-------- backtrace --------
perf[0x539f1b]
/lib64/libc.so.6(+0x34960)[0x7fecb3223960]
perf(pevent_read_number+0x60)[0x542e28]
perf(pevent_read_number_field+0x70)[0x542ed3]
perf(get_field_val+0x6b)[0x548cd5]
perf(pevent_get_field_val+0x6b)[0x548e77]
/root/.traceevent/plugins/plugin_sched_switch.so(+0x95b)[0x7fecb194195b]
perf(pevent_event_info+0x96)[0x547597]
perf[0x4dce26]
perf[0x4e0e89]
perf(hist_entry_iter__add+0xee)[0x4e125e]
perf[0x4304be]
perf[0x4c2db3]
perf[0x4c3301]
perf[0x4c5fdb]
perf[0x4c3535]
perf(perf_session__process_events+0x3a0)[0x4c4b40]
perf(cmd_report+0x120b)[0x4319cb]
perf[0x47c871]
perf(main+0x63f)[0x42242f]
/lib64/libc.so.6(__libc_start_main+0xf0)[0x7fecb320efe0]
perf[0x422549]
[0x0]
[root@zoo ~]# 

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1293664 — Re: [PATCHSET 00/10] perf tools: Support dynamic sort keys for tracepoints (v2)

Hi Arnaldo,

On Wed, Dec 16, 2015 at 09:17:59PM -0300, Arnaldo Carvalho de Melo wrote:
> Em Wed, Dec 16, 2015 at 12:35:33AM +0900, Namhyung Kim escreveu:
> > Hello,
> > 
> > This is an attempt to improve perf to deal with tracepoint events
> > better.  The perf tools can handle tracepoint events but perf report
> > on them is less useful since they're always sampled in a fixed
> > location and not provide event specific info.  We can use perf script
> > but I always wishes there's more convenient way to see the result.
> > 
> >  * changes in v2)
> >   - add 'trace' sort key and make it default  (Jiri)
> >   - add '--raw-trace' option and '/raw' field modifier  (Jiri)
> >   - support event name shortcuts  (David)
> 
> Can you take a look if you can reproduce this? Without callchains it works in
> all tests I did.

Argh, it was because I forgot to set raw_data/size field for
--children case.  I'll send the fix soon.

Maybe we can disable --children for tracepoint sessions (or if it
doesn't have symbol sort key)?

Thanks,
Namhyung


> 
> [root@zoo ~]# perf record -g -e kmem:kmalloc -a
> ^C[ perf record: Woken up 1 times to write data ]
> [ perf record: Captured and wrote 1.193 MB perf.data (250 samples) ]
> 
> [root@zoo ~]# perf report
> perf: Segmentation fault
> -------- backtrace --------
> perf[0x539f1b]
> /lib64/libc.so.6(+0x34960)[0x7fc8e3752960]
> perf(pevent_read_number+0x78)[0x542e40]
> perf(pevent_read_number_field+0x70)[0x542ed3]
> /root/.traceevent/plugins/plugin_kmem.so(+0x603)[0x7fc8e1a6c603]
> perf(pevent_event_info+0x96)[0x547597]
> perf[0x4dce26]
> perf[0x4e0e89]
> perf(hist_entry_iter__add+0xee)[0x4e125e]
> perf[0x4304be]
> perf[0x4c2db3]
> perf[0x4c3301]
> perf[0x4c6089]
> perf(perf_session__process_events+0x3f1)[0x4c4b91]
> perf(cmd_report+0x120b)[0x4319cb]
> perf[0x47c871]
> perf(main+0x63f)[0x42242f]
> /lib64/libc.so.6(__libc_start_main+0xf0)[0x7fc8e373dfe0]
> perf[0x422549]
> [0x0]
> [root@zoo ~]# ls -la ~/.traceevent/
> total 12
> drwxr-xr-x.  3 acme acme 4096 May  4  2015 .
> drwx------. 61 acme acme 4096 Dec 16 21:12 ..
> drwxr-xr-x.  2 acme acme 4096 Nov 23 11:15 plugins
> [root@zoo ~]# ls -la ~/.traceevent
> lrwxrwxrwx. 1 root root 23 May  4  2015 /root/.traceevent -> /home/acme/.traceevent/
> [root@zoo ~]# perf record -e sched:sched_switch -a
> 
> 
> ^C[ perf record: Woken up 1 times to write data ]
> [ perf record: Captured and wrote 1.704 MB perf.data (4967 samples) ]
> 
> [root@zoo ~]# perf report
> [root@zoo ~]# perf record -g -e sched:sched_switch -a
> ^C[ perf record: Woken up 2 times to write data ]
> [ perf record: Captured and wrote 2.062 MB perf.data (4743 samples) ]
> 
> [root@zoo ~]# perf report
> perf: Segmentation fault
> -------- backtrace --------
> perf[0x539f1b]
> /lib64/libc.so.6(+0x34960)[0x7fecb3223960]
> perf(pevent_read_number+0x60)[0x542e28]
> perf(pevent_read_number_field+0x70)[0x542ed3]
> perf(get_field_val+0x6b)[0x548cd5]
> perf(pevent_get_field_val+0x6b)[0x548e77]
> /root/.traceevent/plugins/plugin_sched_switch.so(+0x95b)[0x7fecb194195b]
> perf(pevent_event_info+0x96)[0x547597]
> perf[0x4dce26]
> perf[0x4e0e89]
> perf(hist_entry_iter__add+0xee)[0x4e125e]
> perf[0x4304be]
> perf[0x4c2db3]
> perf[0x4c3301]
> perf[0x4c5fdb]
> perf[0x4c3535]
> perf(perf_session__process_events+0x3a0)[0x4c4b40]
> perf(cmd_report+0x120b)[0x4319cb]
> perf[0x47c871]
> perf(main+0x63f)[0x42242f]
> /lib64/libc.so.6(__libc_start_main+0xf0)[0x7fecb320efe0]
> perf[0x422549]
> [0x0]
> [root@zoo ~]# 
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1293841 — Re: [PATCHSET 00/10] perf tools: Support dynamic sort keys for tracepoints (v2)

Em Thu, Dec 17, 2015 at 04:56:31PM +0900, Namhyung Kim escreveu:
> Hi Arnaldo,
> 
> On Wed, Dec 16, 2015 at 09:17:59PM -0300, Arnaldo Carvalho de Melo wrote:
> > Em Wed, Dec 16, 2015 at 12:35:33AM +0900, Namhyung Kim escreveu:
> > > Hello,
> > > 
> > > This is an attempt to improve perf to deal with tracepoint events
> > > better.  The perf tools can handle tracepoint events but perf report
> > > on them is less useful since they're always sampled in a fixed
> > > location and not provide event specific info.  We can use perf script
> > > but I always wishes there's more convenient way to see the result.
> > > 
> > >  * changes in v2)
> > >   - add 'trace' sort key and make it default  (Jiri)
> > >   - add '--raw-trace' option and '/raw' field modifier  (Jiri)
> > >   - support event name shortcuts  (David)
> > 
> > Can you take a look if you can reproduce this? Without callchains it works in
> > all tests I did.
> 
> Argh, it was because I forgot to set raw_data/size field for
> --children case.  I'll send the fix soon.
> 
> Maybe we can disable --children for tracepoint sessions (or if it
> doesn't have symbol sort key)?

Looks like a plan, unless we want to have all the callchains that get to
some specific DSO or pid.

- Arnaldo
 
> Thanks,
> Namhyung
> 
> 
> > 
> > [root@zoo ~]# perf record -g -e kmem:kmalloc -a
> > ^C[ perf record: Woken up 1 times to write data ]
> > [ perf record: Captured and wrote 1.193 MB perf.data (250 samples) ]
> > 
> > [root@zoo ~]# perf report
> > perf: Segmentation fault
> > -------- backtrace --------
> > perf[0x539f1b]
> > /lib64/libc.so.6(+0x34960)[0x7fc8e3752960]
> > perf(pevent_read_number+0x78)[0x542e40]
> > perf(pevent_read_number_field+0x70)[0x542ed3]
> > /root/.traceevent/plugins/plugin_kmem.so(+0x603)[0x7fc8e1a6c603]
> > perf(pevent_event_info+0x96)[0x547597]
> > perf[0x4dce26]
> > perf[0x4e0e89]
> > perf(hist_entry_iter__add+0xee)[0x4e125e]
> > perf[0x4304be]
> > perf[0x4c2db3]
> > perf[0x4c3301]
> > perf[0x4c6089]
> > perf(perf_session__process_events+0x3f1)[0x4c4b91]
> > perf(cmd_report+0x120b)[0x4319cb]
> > perf[0x47c871]
> > perf(main+0x63f)[0x42242f]
> > /lib64/libc.so.6(__libc_start_main+0xf0)[0x7fc8e373dfe0]
> > perf[0x422549]
> > [0x0]
> > [root@zoo ~]# ls -la ~/.traceevent/
> > total 12
> > drwxr-xr-x.  3 acme acme 4096 May  4  2015 .
> > drwx------. 61 acme acme 4096 Dec 16 21:12 ..
> > drwxr-xr-x.  2 acme acme 4096 Nov 23 11:15 plugins
> > [root@zoo ~]# ls -la ~/.traceevent
> > lrwxrwxrwx. 1 root root 23 May  4  2015 /root/.traceevent -> /home/acme/.traceevent/
> > [root@zoo ~]# perf record -e sched:sched_switch -a
> > 
> > 
> > ^C[ perf record: Woken up 1 times to write data ]
> > [ perf record: Captured and wrote 1.704 MB perf.data (4967 samples) ]
> > 
> > [root@zoo ~]# perf report
> > [root@zoo ~]# perf record -g -e sched:sched_switch -a
> > ^C[ perf record: Woken up 2 times to write data ]
> > [ perf record: Captured and wrote 2.062 MB perf.data (4743 samples) ]
> > 
> > [root@zoo ~]# perf report
> > perf: Segmentation fault
> > -------- backtrace --------
> > perf[0x539f1b]
> > /lib64/libc.so.6(+0x34960)[0x7fecb3223960]
> > perf(pevent_read_number+0x60)[0x542e28]
> > perf(pevent_read_number_field+0x70)[0x542ed3]
> > perf(get_field_val+0x6b)[0x548cd5]
> > perf(pevent_get_field_val+0x6b)[0x548e77]
> > /root/.traceevent/plugins/plugin_sched_switch.so(+0x95b)[0x7fecb194195b]
> > perf(pevent_event_info+0x96)[0x547597]
> > perf[0x4dce26]
> > perf[0x4e0e89]
> > perf(hist_entry_iter__add+0xee)[0x4e125e]
> > perf[0x4304be]
> > perf[0x4c2db3]
> > perf[0x4c3301]
> > perf[0x4c5fdb]
> > perf[0x4c3535]
> > perf(perf_session__process_events+0x3a0)[0x4c4b40]
> > perf(cmd_report+0x120b)[0x4319cb]
> > perf[0x47c871]
> > perf(main+0x63f)[0x42242f]
> > /lib64/libc.so.6(__libc_start_main+0xf0)[0x7fecb320efe0]
> > perf[0x422549]
> > [0x0]
> > [root@zoo ~]# 
> > 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [standalone]

Page 2 of 2 — ← Prev page 1 [2]

Back to top | Article view | linux.kernel

csiph-web