[v1,2/2] perf test: Avoid infinite loop for task exit case

Message ID 20191011091942.29841-2-leo.yan@linaro.org
State New
Headers show
Series
  • [v1,1/2] perf test: Report failure for mmap events
Related show

Commit Message

Leo Yan Oct. 11, 2019, 9:19 a.m.
When execute task exit testing case, Perf tool stucks in this case and
doesn't return back on Arm64 Juno board.

After dig into this issue, since Juno board has Arm's big.LITTLE CPUs,
thus the PMUs are not compatible between the big CPUs and little CPUs.
This leads to PMU event cannot be enabled properly when the traced task
is migrated from one variant's CPU to another variant.  Finally, the
test case runs into infinite loop for cannot read out any event data
after return from polling.

Eventually, we need to work out formal solution to allow PMU events can
be freely migrated from one CPU variant to another, but this is a
difficult task and a different topic.  This patch tries to fix the Perf
test case to avoid infinite loop, when the testing detects 1000 times
retrying for reading empty events, it will directly bail out and return
failure.  This allows the Perf tool can continue its other test cases.

Signed-off-by: Leo Yan <leo.yan@linaro.org>

---
 tools/perf/tests/task-exit.c | 8 ++++++++
 1 file changed, 8 insertions(+)

-- 
2.17.1

Comments

Arnaldo Carvalho de Melo Oct. 11, 2019, 2:38 p.m. | #1
Em Fri, Oct 11, 2019 at 05:19:42PM +0800, Leo Yan escreveu:
> When execute task exit testing case, Perf tool stucks in this case and

> doesn't return back on Arm64 Juno board.

> 

> After dig into this issue, since Juno board has Arm's big.LITTLE CPUs,

> thus the PMUs are not compatible between the big CPUs and little CPUs.

> This leads to PMU event cannot be enabled properly when the traced task

> is migrated from one variant's CPU to another variant.  Finally, the

> test case runs into infinite loop for cannot read out any event data

> after return from polling.


Thanks, I'll go over this after I finish the current pull request,

- Arnaldo
 
> Eventually, we need to work out formal solution to allow PMU events can

> be freely migrated from one CPU variant to another, but this is a

> difficult task and a different topic.  This patch tries to fix the Perf

> test case to avoid infinite loop, when the testing detects 1000 times

> retrying for reading empty events, it will directly bail out and return

> failure.  This allows the Perf tool can continue its other test cases.

> 

> Signed-off-by: Leo Yan <leo.yan@linaro.org>

> ---

>  tools/perf/tests/task-exit.c | 8 ++++++++

>  1 file changed, 8 insertions(+)

> 

> diff --git a/tools/perf/tests/task-exit.c b/tools/perf/tests/task-exit.c

> index ca0a6ca43b13..d85c9f608564 100644

> --- a/tools/perf/tests/task-exit.c

> +++ b/tools/perf/tests/task-exit.c

> @@ -53,6 +53,7 @@ int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unused

>  	struct perf_cpu_map *cpus;

>  	struct perf_thread_map *threads;

>  	struct mmap *md;

> +	int retry_count = 0;

>  

>  	signal(SIGCHLD, sig_handler);

>  

> @@ -132,6 +133,13 @@ int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unused

>  out_init:

>  	if (!exited || !nr_exit) {

>  		evlist__poll(evlist, -1);

> +

> +		if (retry_count++ > 1000) {

> +			pr_debug("Failed after retrying 1000 times\n");

> +			err = -1;

> +			goto out_free_maps;

> +		}

> +

>  		goto retry;

>  	}

>  

> -- 

> 2.17.1


-- 

- Arnaldo
Arnaldo Carvalho de Melo Oct. 14, 2019, 2:11 p.m. | #2
Em Fri, Oct 11, 2019 at 05:19:42PM +0800, Leo Yan escreveu:
> When execute task exit testing case, Perf tool stucks in this case and

> doesn't return back on Arm64 Juno board.

> 

> After dig into this issue, since Juno board has Arm's big.LITTLE CPUs,

> thus the PMUs are not compatible between the big CPUs and little CPUs.

> This leads to PMU event cannot be enabled properly when the traced task

> is migrated from one variant's CPU to another variant.  Finally, the

> test case runs into infinite loop for cannot read out any event data

> after return from polling.

> 

> Eventually, we need to work out formal solution to allow PMU events can

> be freely migrated from one CPU variant to another, but this is a

> difficult task and a different topic.  This patch tries to fix the Perf

> test case to avoid infinite loop, when the testing detects 1000 times

> retrying for reading empty events, it will directly bail out and return

> failure.  This allows the Perf tool can continue its other test cases.


Thanks, applied.

- Arnaldo
 
> Signed-off-by: Leo Yan <leo.yan@linaro.org>

> ---

>  tools/perf/tests/task-exit.c | 8 ++++++++

>  1 file changed, 8 insertions(+)

> 

> diff --git a/tools/perf/tests/task-exit.c b/tools/perf/tests/task-exit.c

> index ca0a6ca43b13..d85c9f608564 100644

> --- a/tools/perf/tests/task-exit.c

> +++ b/tools/perf/tests/task-exit.c

> @@ -53,6 +53,7 @@ int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unused

>  	struct perf_cpu_map *cpus;

>  	struct perf_thread_map *threads;

>  	struct mmap *md;

> +	int retry_count = 0;

>  

>  	signal(SIGCHLD, sig_handler);

>  

> @@ -132,6 +133,13 @@ int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unused

>  out_init:

>  	if (!exited || !nr_exit) {

>  		evlist__poll(evlist, -1);

> +

> +		if (retry_count++ > 1000) {

> +			pr_debug("Failed after retrying 1000 times\n");

> +			err = -1;

> +			goto out_free_maps;

> +		}

> +

>  		goto retry;

>  	}

>  

> -- 

> 2.17.1


-- 

- Arnaldo
Leo Yan Oct. 15, 2019, 3:20 a.m. | #3
On Mon, Oct 14, 2019 at 11:11:36AM -0300, Arnaldo Carvalho de Melo wrote:
> Em Fri, Oct 11, 2019 at 05:19:42PM +0800, Leo Yan escreveu:

> > When execute task exit testing case, Perf tool stucks in this case and

> > doesn't return back on Arm64 Juno board.

> > 

> > After dig into this issue, since Juno board has Arm's big.LITTLE CPUs,

> > thus the PMUs are not compatible between the big CPUs and little CPUs.

> > This leads to PMU event cannot be enabled properly when the traced task

> > is migrated from one variant's CPU to another variant.  Finally, the

> > test case runs into infinite loop for cannot read out any event data

> > after return from polling.

> > 

> > Eventually, we need to work out formal solution to allow PMU events can

> > be freely migrated from one CPU variant to another, but this is a

> > difficult task and a different topic.  This patch tries to fix the Perf

> > test case to avoid infinite loop, when the testing detects 1000 times

> > retrying for reading empty events, it will directly bail out and return

> > failure.  This allows the Perf tool can continue its other test cases.

> 

> Thanks, applied.


Thanks a lot, Arnaldo.

Patch

diff --git a/tools/perf/tests/task-exit.c b/tools/perf/tests/task-exit.c
index ca0a6ca43b13..d85c9f608564 100644
--- a/tools/perf/tests/task-exit.c
+++ b/tools/perf/tests/task-exit.c
@@ -53,6 +53,7 @@  int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unused
 	struct perf_cpu_map *cpus;
 	struct perf_thread_map *threads;
 	struct mmap *md;
+	int retry_count = 0;
 
 	signal(SIGCHLD, sig_handler);
 
@@ -132,6 +133,13 @@  int test__task_exit(struct test *test __maybe_unused, int subtest __maybe_unused
 out_init:
 	if (!exited || !nr_exit) {
 		evlist__poll(evlist, -1);
+
+		if (retry_count++ > 1000) {
+			pr_debug("Failed after retrying 1000 times\n");
+			err = -1;
+			goto out_free_maps;
+		}
+
 		goto retry;
 	}