* [PATCH v1 1/3] perf/core: store context switch out type into Perf trace
2018-03-05 11:35 [PATCH v1 0/3] perf: expose thread context switch out event type to user space Alexey Budankov
@ 2018-03-05 11:36 ` Alexey Budankov
2018-03-05 11:38 ` [PATCH v1 2/3] perf report: extend raw dump (-D) out with switch out event type Alexey Budankov
` (3 subsequent siblings)
4 siblings, 0 replies; 7+ messages in thread
From: Alexey Budankov @ 2018-03-05 11:36 UTC (permalink / raw)
To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo
Cc: Alexander Shishkin, Jiri Olsa, Namhyung Kim, Andi Kleen,
linux-kernel, Dmitri Prokhorov
Store thread context-switch-out event type into Perf trace as a part of
PERF_RECORD_SWITCH[_CPU_WIDE] records.
Introduced types of switch-out events assumed to be
a) preempt: task->state == TASK_RUNNING and b) yield: !preempt;
New yield event type is encoded using special
PERF_RECORD_MISC_SWITCH_OUT_YIELD bit extending PERF_RECORD_MISC_SWITCH_OUT
meaning traditional preemption switch out event:
misc &= PERF_RECORD_MISC_SWITCH_OUT | PERF_RECORD_MISC_SWITCH_OUT_YIELD
Signed-off-by: Alexey Budankov <alexey.budankov@linux.intel.com>
---
include/uapi/linux/perf_event.h | 5 +++++
kernel/events/core.c | 4 +++-
2 files changed, 8 insertions(+), 1 deletion(-)
diff --git a/include/uapi/linux/perf_event.h b/include/uapi/linux/perf_event.h
index 6f873503552d..0339c829cda5 100644
--- a/include/uapi/linux/perf_event.h
+++ b/include/uapi/linux/perf_event.h
@@ -654,6 +654,11 @@ struct perf_event_mmap_page {
* perf_event_attr::precise_ip.
*/
#define PERF_RECORD_MISC_EXACT_IP (1 << 14)
+/*
+ * Indicates that thread explicitly yielded cpu due to
+ * a call of some synchronization API e.g. futex system call
+ */
+#define PERF_RECORD_MISC_SWITCH_OUT_YIELD (1 << 14)
/*
* Reserve the last bit to indicate some extended misc field
*/
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 57898102847f..1faa6dde090c 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -7216,6 +7216,8 @@ static void perf_event_switch(struct task_struct *task,
struct task_struct *next_prev, bool sched_in)
{
struct perf_switch_event switch_event;
+ __u16 switch_type = sched_in ? 0 : PERF_RECORD_MISC_SWITCH_OUT |
+ (task->state == TASK_RUNNING ? 0 : PERF_RECORD_MISC_SWITCH_OUT_YIELD);
/* N.B. caller checks nr_switch_events != 0 */
@@ -7225,7 +7227,7 @@ static void perf_event_switch(struct task_struct *task,
.event_id = {
.header = {
/* .type */
- .misc = sched_in ? 0 : PERF_RECORD_MISC_SWITCH_OUT,
+ .misc = switch_type,
/* .size */
},
/* .next_prev_pid */
^ permalink raw reply related [flat|nested] 7+ messages in thread* [PATCH v1 2/3] perf report: extend raw dump (-D) out with switch out event type
2018-03-05 11:35 [PATCH v1 0/3] perf: expose thread context switch out event type to user space Alexey Budankov
2018-03-05 11:36 ` [PATCH v1 1/3] perf/core: store context switch out type into Perf trace Alexey Budankov
@ 2018-03-05 11:38 ` Alexey Budankov
2018-03-05 11:42 ` [PATCH v1 3/3] perf script: extend misc field deconding " Alexey Budankov
` (2 subsequent siblings)
4 siblings, 0 replies; 7+ messages in thread
From: Alexey Budankov @ 2018-03-05 11:38 UTC (permalink / raw)
To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo
Cc: Alexander Shishkin, Jiri Olsa, Namhyung Kim, Andi Kleen,
linux-kernel, Dmitri Prokhorov
Print additional 'yield' tag for PERF_RECORD_SWITCH[_CPU_WIDE] OUT records when
event header misc field contains PERF_RECORD_MISC_SWITCH_OUT_YIELD bit set
designating synchronization context switch out event:
perf report -D -i system-wide.perf:
...
0x1b9c50 [0x30]: event: 15
.
. ... raw event: size 48 bytes
. 0000: 0f 00 00 00 00 20 30 00 01 1e 00 00 01 1e 00 00 ..... 0.........
. 0010: 00 00 00 00 00 00 00 00 85 ae d4 e3 3e 0e 00 00 ............>...
. 0020: 54 00 00 00 00 00 00 00 05 00 00 00 00 00 00 00 T...............
5 15663273127557 0x1b9c50 [0x30]: PERF_RECORD_SWITCH_CPU_WIDE OUT next pid/tid: 7681/7681
0x2646c0 [0x30]: event: 15
.
. ... raw event: size 48 bytes
. 0000: 0f 00 00 00 00 60 30 00 00 00 00 00 00 00 00 00 .....`0.........
. 0010: 00 1e 00 00 00 1e 00 00 29 1e d5 e3 3e 0e 00 00 ........)...>...
. 0020: 56 00 00 00 00 00 00 00 07 00 00 00 00 00 00 00 V...............
7 15663273156137 0x2646c0 [0x30]: PERF_RECORD_SWITCH_CPU_WIDE OUT yield next pid/tid: 0/0
...
Signed-off-by: Alexey Budankov <alexey.budankov@linux.intel.com>
---
tools/include/uapi/linux/perf_event.h | 5 +++++
tools/perf/util/event.c | 4 +++-
2 files changed, 8 insertions(+), 1 deletion(-)
diff --git a/tools/include/uapi/linux/perf_event.h b/tools/include/uapi/linux/perf_event.h
index 6f873503552d..0339c829cda5 100644
--- a/tools/include/uapi/linux/perf_event.h
+++ b/tools/include/uapi/linux/perf_event.h
@@ -654,6 +654,11 @@ struct perf_event_mmap_page {
* perf_event_attr::precise_ip.
*/
#define PERF_RECORD_MISC_EXACT_IP (1 << 14)
+/*
+ * Indicates that thread explicitly yielded cpu due to
+ * a call of some synchronization API e.g. futex system call
+ */
+#define PERF_RECORD_MISC_SWITCH_OUT_YIELD (1 << 14)
/*
* Reserve the last bit to indicate some extended misc field
*/
diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c
index f0a6cbd033cc..af5a85d56446 100644
--- a/tools/perf/util/event.c
+++ b/tools/perf/util/event.c
@@ -1421,7 +1421,9 @@ size_t perf_event__fprintf_itrace_start(union perf_event *event, FILE *fp)
size_t perf_event__fprintf_switch(union perf_event *event, FILE *fp)
{
bool out = event->header.misc & PERF_RECORD_MISC_SWITCH_OUT;
- const char *in_out = out ? "OUT" : "IN ";
+ const char *in_out = !out ? "IN " :
+ !(event->header.misc & PERF_RECORD_MISC_SWITCH_OUT_YIELD) ?
+ "OUT" : "OUT yield";
if (event->header.type == PERF_RECORD_SWITCH)
return fprintf(fp, " %s\n", in_out);
^ permalink raw reply related [flat|nested] 7+ messages in thread* [PATCH v1 3/3] perf script: extend misc field deconding with switch out event type
2018-03-05 11:35 [PATCH v1 0/3] perf: expose thread context switch out event type to user space Alexey Budankov
2018-03-05 11:36 ` [PATCH v1 1/3] perf/core: store context switch out type into Perf trace Alexey Budankov
2018-03-05 11:38 ` [PATCH v1 2/3] perf report: extend raw dump (-D) out with switch out event type Alexey Budankov
@ 2018-03-05 11:42 ` Alexey Budankov
2018-03-05 15:06 ` [PATCH v1 0/3] perf: expose thread context switch out event type to user space Arnaldo Carvalho de Melo
2018-03-05 16:52 ` Andi Kleen
4 siblings, 0 replies; 7+ messages in thread
From: Alexey Budankov @ 2018-03-05 11:42 UTC (permalink / raw)
To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo
Cc: Alexander Shishkin, Jiri Olsa, Namhyung Kim, Andi Kleen,
linux-kernel, Dmitri Prokhorov
Append 'y' sign to 'S' tag designating the type of context switch out event so
'S' means preemption context switch and 'Sy' means synchronization context
switch. Documentation is extended to cover new presentation changes.
perf script --show-switch-events -F +misc -I -i system-wide.perf:
amplxe-perf 7681 [005] S 15663.273151: PERF_RECORD_SWITCH_CPU_WIDE OUT next pid/tid: 39/39
migration/5 39 [005] 15663.273152: PERF_RECORD_SWITCH_CPU_WIDE IN prev pid/tid: 7681/7681
amplxe-perf 7680 [007] K 15663.273153: 1 context-switch:
aaa488 schedule ([kernel.kallsyms])
1a9f50 __poll_nocancel (inlined)
amplxe-perf 7680 [007] Sy 15663.273156: PERF_RECORD_SWITCH_CPU_WIDE OUT yield next pid/tid: 0/0
migration/5 39 [005] K 15663.273157:
Signed-off-by: Alexey Budankov <alexey.budankov@linux.intel.com>
---
tools/perf/Documentation/perf-script.txt | 17 +++++++++--------
tools/perf/builtin-script.c | 5 ++++-
2 files changed, 13 insertions(+), 9 deletions(-)
diff --git a/tools/perf/Documentation/perf-script.txt b/tools/perf/Documentation/perf-script.txt
index 36ec0257f8d3..80e94cbbf520 100644
--- a/tools/perf/Documentation/perf-script.txt
+++ b/tools/perf/Documentation/perf-script.txt
@@ -228,14 +228,15 @@ OPTIONS
For sample events it's possible to display misc field with -F +misc option,
following letters are displayed for each bit:
- PERF_RECORD_MISC_KERNEL K
- PERF_RECORD_MISC_USER U
- PERF_RECORD_MISC_HYPERVISOR H
- PERF_RECORD_MISC_GUEST_KERNEL G
- PERF_RECORD_MISC_GUEST_USER g
- PERF_RECORD_MISC_MMAP_DATA* M
- PERF_RECORD_MISC_COMM_EXEC E
- PERF_RECORD_MISC_SWITCH_OUT S
+ PERF_RECORD_MISC_KERNEL K
+ PERF_RECORD_MISC_USER U
+ PERF_RECORD_MISC_HYPERVISOR H
+ PERF_RECORD_MISC_GUEST_KERNEL G
+ PERF_RECORD_MISC_GUEST_USER g
+ PERF_RECORD_MISC_MMAP_DATA* M
+ PERF_RECORD_MISC_COMM_EXEC E
+ PERF_RECORD_MISC_SWITCH_OUT S
+ PERF_RECORD_MISC_SWITCH_OUT_YIELD Sy
$ perf script -F +misc ...
sched-messaging 1414 K 28690.636582: 4590 cycles ...
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index cce926aeb0c0..06d7fe73886f 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -657,8 +657,11 @@ static int perf_sample__fprintf_start(struct perf_sample *sample,
break;
case PERF_RECORD_SWITCH:
case PERF_RECORD_SWITCH_CPU_WIDE:
- if (has(SWITCH_OUT))
+ if (has(SWITCH_OUT)) {
ret += fprintf(fp, "S");
+ if (sample->misc & PERF_RECORD_MISC_SWITCH_OUT_YIELD)
+ ret += fprintf(fp, "y");
+ }
default:
break;
}
^ permalink raw reply related [flat|nested] 7+ messages in thread* Re: [PATCH v1 0/3] perf: expose thread context switch out event type to user space
2018-03-05 11:35 [PATCH v1 0/3] perf: expose thread context switch out event type to user space Alexey Budankov
` (2 preceding siblings ...)
2018-03-05 11:42 ` [PATCH v1 3/3] perf script: extend misc field deconding " Alexey Budankov
@ 2018-03-05 15:06 ` Arnaldo Carvalho de Melo
2018-03-05 16:20 ` Alexey Budankov
2018-03-05 16:52 ` Andi Kleen
4 siblings, 1 reply; 7+ messages in thread
From: Arnaldo Carvalho de Melo @ 2018-03-05 15:06 UTC (permalink / raw)
To: Alexey Budankov
Cc: Peter Zijlstra, Ingo Molnar, Alexander Shishkin, Jiri Olsa,
Namhyung Kim, Andi Kleen, linux-kernel, Dmitri Prokhorov
Em Mon, Mar 05, 2018 at 02:35:02PM +0300, Alexey Budankov escreveu:
>
> Here is a series of small patches that implement exposing type of
> context-switch-out event as a part of PERF_RECORD_SWITCH[_CPU_WIDE] record.
>
> Introduced types of context-switch-out events assumed to be:
> a) preempt: task->state == TASK_RUNNING
> misc &= PERF_RECORD_MISC_SWITCH_OUT
>
> b) yield: !preempt - using new bit PERF_RECORD_MISC_SWITCH_OUT_YIELD:
misc &= PERF_RECORD_MISC_SWITCH_OUT|PERF_RECORD_MISC_SWITCH_OUT_YIELD
>
> Perf tool report and script commands output has been extended to decode
> new yield bit and the updated output looks like in the examples below.
I'm just waiting for the current reviewers to be satisfied with this,
but I think this is a great addition and 'perf trace' is another tool
that should jump into this, showing forced context switches together
with syscalls.
- Arnaldo
> The documentation has been updated to mention yield switch out events and its
> decoding symbols in perf script output.
>
> The changes have been manually tested on Fedora 27 with the patched kernel:
> git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git perf/core
>
> perf report -D -i system-wide.perf:
>
> 0x1b9c50 [0x30]: event: 15
> .
> . ... raw event: size 48 bytes
> . 0000: 0f 00 00 00 00 20 30 00 01 1e 00 00 01 1e 00 00 ..... 0.........
> . 0010: 00 00 00 00 00 00 00 00 85 ae d4 e3 3e 0e 00 00 ............>...
> . 0020: 54 00 00 00 00 00 00 00 05 00 00 00 00 00 00 00 T...............
>
> 5 15663273127557 0x1b9c50 [0x30]: PERF_RECORD_SWITCH_CPU_WIDE OUT next pid/tid: 7681/7681
>
> 0x2646c0 [0x30]: event: 15
> .
> . ... raw event: size 48 bytes
> . 0000: 0f 00 00 00 00 60 30 00 00 00 00 00 00 00 00 00 .....`0.........
> . 0010: 00 1e 00 00 00 1e 00 00 29 1e d5 e3 3e 0e 00 00 ........)...>...
> . 0020: 56 00 00 00 00 00 00 00 07 00 00 00 00 00 00 00 V...............
>
> 7 15663273156137 0x2646c0 [0x30]: PERF_RECORD_SWITCH_CPU_WIDE OUT yield next pid/tid: 0/0
>
> perf script --show-switch-events -F +misc -I -i system-wide.perf:
>
> amplxe-perf 7681 [005] S 15663.273151: PERF_RECORD_SWITCH_CPU_WIDE OUT next pid/tid: 39/39
> migration/5 39 [005] 15663.273152: PERF_RECORD_SWITCH_CPU_WIDE IN prev pid/tid: 7681/7681
> amplxe-perf 7680 [007] K 15663.273153: 1 context-switch:
> aaa488 schedule ([kernel.kallsyms])
> 1a9f50 __poll_nocancel (inlined)
>
> amplxe-perf 7680 [007] Sy 15663.273156: PERF_RECORD_SWITCH_CPU_WIDE OUT yield next pid/tid: 0/0
> migration/5 39 [005] K 15663.273157:
>
> ---
> Alexey Budankov (3):
> perf/core: store context switch out type into Perf trace
> perf report: extend raw dump (-D) out with switch out event type
> perf script: extend misc field decoding with switch out event type
>
> include/uapi/linux/perf_event.h | 5 +++++
> kernel/events/core.c | 4 +++-
> tools/include/uapi/linux/perf_event.h | 5 +++++
> tools/perf/Documentation/perf-script.txt | 17 +++++++++--------
> tools/perf/builtin-script.c | 5 ++++-
> tools/perf/util/event.c | 4 +++-
> 6 files changed, 29 insertions(+), 11 deletions(-)
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH v1 0/3] perf: expose thread context switch out event type to user space
2018-03-05 15:06 ` [PATCH v1 0/3] perf: expose thread context switch out event type to user space Arnaldo Carvalho de Melo
@ 2018-03-05 16:20 ` Alexey Budankov
0 siblings, 0 replies; 7+ messages in thread
From: Alexey Budankov @ 2018-03-05 16:20 UTC (permalink / raw)
To: Arnaldo Carvalho de Melo
Cc: Peter Zijlstra, Ingo Molnar, Alexander Shishkin, Jiri Olsa,
Namhyung Kim, Andi Kleen, linux-kernel, Dmitri Prokhorov
Hi Arnaldo,
On 05.03.2018 18:06, Arnaldo Carvalho de Melo wrote:
> Em Mon, Mar 05, 2018 at 02:35:02PM +0300, Alexey Budankov escreveu:
>>
>> Here is a series of small patches that implement exposing type of
>> context-switch-out event as a part of PERF_RECORD_SWITCH[_CPU_WIDE] record.
>>
>> Introduced types of context-switch-out events assumed to be:
>> a) preempt: task->state == TASK_RUNNING
>> misc &= PERF_RECORD_MISC_SWITCH_OUT
>>
>> b) yield: !preempt - using new bit PERF_RECORD_MISC_SWITCH_OUT_YIELD:
> misc &= PERF_RECORD_MISC_SWITCH_OUT|PERF_RECORD_MISC_SWITCH_OUT_YIELD
>>
>> Perf tool report and script commands output has been extended to decode
>> new yield bit and the updated output looks like in the examples below.
>
> I'm just waiting for the current reviewers to be satisfied with this,
> but I think this is a great addition and 'perf trace' is another tool
> that should jump into this, showing forced context switches together
> with syscalls.
It's great to know there is a value in that change for other Perf tools.
Extending perf trace (strace inspired) tool in that respect might makes
sense. I anticipate possible tracing overhead needs to be expected and
probably handled somehow.
But, anyway, yep, per-thread syscall traces enriched with typed context
switch boundaries may be the great extension, in comparison to the
original strace tool.
IMHO, some simple summary metrics like amount of preempt or yield
context switches (per-thread or per-process) could bring even more
value into perf trace tool functionality.
BR,
Alexey
>
> - Arnaldo
>
>> The documentation has been updated to mention yield switch out events and its
>> decoding symbols in perf script output.
>>
>> The changes have been manually tested on Fedora 27 with the patched kernel:
>> git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git perf/core
>>
>> perf report -D -i system-wide.perf:
>>
>> 0x1b9c50 [0x30]: event: 15
>> .
>> . ... raw event: size 48 bytes
>> . 0000: 0f 00 00 00 00 20 30 00 01 1e 00 00 01 1e 00 00 ..... 0.........
>> . 0010: 00 00 00 00 00 00 00 00 85 ae d4 e3 3e 0e 00 00 ............>...
>> . 0020: 54 00 00 00 00 00 00 00 05 00 00 00 00 00 00 00 T...............
>>
>> 5 15663273127557 0x1b9c50 [0x30]: PERF_RECORD_SWITCH_CPU_WIDE OUT next pid/tid: 7681/7681
>>
>> 0x2646c0 [0x30]: event: 15
>> .
>> . ... raw event: size 48 bytes
>> . 0000: 0f 00 00 00 00 60 30 00 00 00 00 00 00 00 00 00 .....`0.........
>> . 0010: 00 1e 00 00 00 1e 00 00 29 1e d5 e3 3e 0e 00 00 ........)...>...
>> . 0020: 56 00 00 00 00 00 00 00 07 00 00 00 00 00 00 00 V...............
>>
>> 7 15663273156137 0x2646c0 [0x30]: PERF_RECORD_SWITCH_CPU_WIDE OUT yield next pid/tid: 0/0
>>
>> perf script --show-switch-events -F +misc -I -i system-wide.perf:
>>
>> amplxe-perf 7681 [005] S 15663.273151: PERF_RECORD_SWITCH_CPU_WIDE OUT next pid/tid: 39/39
>> migration/5 39 [005] 15663.273152: PERF_RECORD_SWITCH_CPU_WIDE IN prev pid/tid: 7681/7681
>> amplxe-perf 7680 [007] K 15663.273153: 1 context-switch:
>> aaa488 schedule ([kernel.kallsyms])
>> 1a9f50 __poll_nocancel (inlined)
>>
>> amplxe-perf 7680 [007] Sy 15663.273156: PERF_RECORD_SWITCH_CPU_WIDE OUT yield next pid/tid: 0/0
>> migration/5 39 [005] K 15663.273157:
>>
>> ---
>> Alexey Budankov (3):
>> perf/core: store context switch out type into Perf trace
>> perf report: extend raw dump (-D) out with switch out event type
>> perf script: extend misc field decoding with switch out event type
>>
>> include/uapi/linux/perf_event.h | 5 +++++
>> kernel/events/core.c | 4 +++-
>> tools/include/uapi/linux/perf_event.h | 5 +++++
>> tools/perf/Documentation/perf-script.txt | 17 +++++++++--------
>> tools/perf/builtin-script.c | 5 ++++-
>> tools/perf/util/event.c | 4 +++-
>> 6 files changed, 29 insertions(+), 11 deletions(-)
>
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v1 0/3] perf: expose thread context switch out event type to user space
2018-03-05 11:35 [PATCH v1 0/3] perf: expose thread context switch out event type to user space Alexey Budankov
` (3 preceding siblings ...)
2018-03-05 15:06 ` [PATCH v1 0/3] perf: expose thread context switch out event type to user space Arnaldo Carvalho de Melo
@ 2018-03-05 16:52 ` Andi Kleen
4 siblings, 0 replies; 7+ messages in thread
From: Andi Kleen @ 2018-03-05 16:52 UTC (permalink / raw)
To: Alexey Budankov
Cc: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo,
Alexander Shishkin, Jiri Olsa, Namhyung Kim, linux-kernel,
Dmitri Prokhorov
Patches look good to me.
Reviewed-by: Andi Kleen <ak@linux.intel.com>
^ permalink raw reply [flat|nested] 7+ messages in thread