Re: [PATCH] perf tests: Fix false TEST_OK result for 'perf test hist'
On Tue, Dec 15, 2015 at 04:18:44AM +, Wang Nan wrote:
> Commit 71d6de64feddd4b45326fba2111b3006d9e0 ('perf test: Fix hist
> testcases when kptr_restrict is on') solves a double free problem when
> 'perf test hist' calling setup_fake_machine(). However, the result is
> still incorrect. For example:
>
> $ ./perf test -v 'filtering hist entries'
> 25: Test filtering hist entries :
> --- start ---
> test child forked, pid 4186
> Cannot create kernel maps
> test child finished with 0
> end
> Test filtering hist entries: Ok
>
> In this case the body of this test is not get executed at all, but the
> result is 'Ok'. I think 'Skip' should be better.
>
> The reason is that *ALL* hists test cases forget to reset err after
> using it to hold an error code.
>
> This patch ensure err is reset.
>
> In case when kptr_restrict prevent normal user get kernel address, this
> test should be skipped, not fail. This patch uses linux/err.h to store
> error code in return value of setup_fake_machine(), and let 'EACCES' to
> indicate this problem.
>
> Signed-off-by: Wang Nan
> Cc: Arnaldo Carvalho de Melo
> Cc: Jiri Olsa
> Cc: Masami Hiramatsu
> Cc: Namhyung Kim
Acked-by: Jiri Olsa
thanks,
jirka
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH] perf tests: Fix false TEST_OK result for 'perf test hist'
On Tue, Dec 15, 2015 at 04:18:44AM +, Wang Nan wrote:
> Commit 71d6de64feddd4b45326fba2111b3006d9e0 ('perf test: Fix hist
> testcases when kptr_restrict is on') solves a double free problem when
> 'perf test hist' calling setup_fake_machine(). However, the result is
> still incorrect. For example:
>
> $ ./perf test -v 'filtering hist entries'
> 25: Test filtering hist entries :
> --- start ---
> test child forked, pid 4186
> Cannot create kernel maps
> test child finished with 0
> end
> Test filtering hist entries: Ok
>
> In this case the body of this test is not get executed at all, but the
> result is 'Ok'. I think 'Skip' should be better.
>
> The reason is that *ALL* hists test cases forget to reset err after
> using it to hold an error code.
>
> This patch ensure err is reset.
>
> In case when kptr_restrict prevent normal user get kernel address, this
> test should be skipped, not fail. This patch uses linux/err.h to store
> error code in return value of setup_fake_machine(), and let 'EACCES' to
> indicate this problem.
>
> Signed-off-by: Wang Nan
> Cc: Arnaldo Carvalho de Melo
> Cc: Jiri Olsa
> Cc: Masami Hiramatsu
> Cc: Namhyung Kim
Acked-by: Jiri Olsa
thanks,
jirka
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH] perf tests: Fix false TEST_OK result for 'perf test hist'
Hi,
On Tue, Dec 15, 2015 at 04:18:44AM +, Wang Nan wrote:
> Commit 71d6de64feddd4b45326fba2111b3006d9e0 ('perf test: Fix hist
> testcases when kptr_restrict is on') solves a double free problem when
> 'perf test hist' calling setup_fake_machine(). However, the result is
> still incorrect. For example:
>
> $ ./perf test -v 'filtering hist entries'
> 25: Test filtering hist entries :
> --- start ---
> test child forked, pid 4186
> Cannot create kernel maps
> test child finished with 0
> end
> Test filtering hist entries: Ok
>
> In this case the body of this test is not get executed at all, but the
> result is 'Ok'. I think 'Skip' should be better.
Right, I missed this part.
But on second thought, I think it should not create real kernel maps.
What setup_fake_machine() does is to create/insert fake
dso/mmap/symbols for testing hist functions. So it has to use the
predefined info not real ones.
>
> The reason is that *ALL* hists test cases forget to reset err after
> using it to hold an error code.
>
> This patch ensure err is reset.
>
> In case when kptr_restrict prevent normal user get kernel address, this
> test should be skipped, not fail. This patch uses linux/err.h to store
> error code in return value of setup_fake_machine(), and let 'EACCES' to
> indicate this problem.
Please just remove the machine__create_kernel_maps() in
setup_fake_machine() and simply return TEST_FAIL if there's an error.
Thanks,
Namhyung
>
> Signed-off-by: Wang Nan
> Cc: Arnaldo Carvalho de Melo
> Cc: Jiri Olsa
> Cc: Masami Hiramatsu
> Cc: Namhyung Kim
> ---
> tools/perf/tests/hists_common.c | 12 +---
> tools/perf/tests/hists_common.h | 1 +
> tools/perf/tests/hists_cumulate.c | 6 +-
> tools/perf/tests/hists_filter.c | 6 +-
> tools/perf/tests/hists_link.c | 6 +-
> tools/perf/tests/hists_output.c | 6 +-
> 6 files changed, 30 insertions(+), 7 deletions(-)
>
> diff --git a/tools/perf/tests/hists_common.c b/tools/perf/tests/hists_common.c
> index bcfd081..2eca1b2 100644
> --- a/tools/perf/tests/hists_common.c
> +++ b/tools/perf/tests/hists_common.c
> @@ -81,15 +81,20 @@ struct machine *setup_fake_machine(struct machines
> *machines)
> {
> struct machine *machine = machines__find(machines, HOST_KERNEL_ID);
> size_t i;
> + int err = -ENOMEM;
>
> if (machine == NULL) {
> pr_debug("Not enough memory for machine setup\n");
> - return NULL;
> + return ERR_PTR(-ENOMEM);
> }
>
> if (machine__create_kernel_maps(machine)) {
> pr_debug("Cannot create kernel maps\n");
> - return NULL;
> + if (symbol_conf.kptr_restrict) {
> + pr_debug("Hint: Check
> /proc/sys/kernel/kptr_restrict.\n");
> + err = -EACCES;
> + }
> + return ERR_PTR(err);
> }
>
> for (i = 0; i < ARRAY_SIZE(fake_threads); i++) {
> @@ -155,7 +160,8 @@ struct machine *setup_fake_machine(struct machines
> *machines)
> out:
> pr_debug("Not enough memory for machine setup\n");
> machine__delete_threads(machine);
> - return NULL;
> + machine__exit(machine);
> + return ERR_PTR(err);
> }
>
> void print_hists_in(struct hists *hists)
> diff --git a/tools/perf/tests/hists_common.h b/tools/perf/tests/hists_common.h
> index 888254e..0252eae 100644
> --- a/tools/perf/tests/hists_common.h
> +++ b/tools/perf/tests/hists_common.h
> @@ -1,5 +1,6 @@
> #ifndef __PERF_TESTS__HISTS_COMMON_H__
> #define __PERF_TESTS__HISTS_COMMON_H__
> +#include
>
> struct machine;
> struct machines;
> diff --git a/tools/perf/tests/hists_cumulate.c
> b/tools/perf/tests/hists_cumulate.c
> index 8292948..c211075 100644
> --- a/tools/perf/tests/hists_cumulate.c
> +++ b/tools/perf/tests/hists_cumulate.c
> @@ -706,13 +706,17 @@ int test__hists_cumulate(int subtest __maybe_unused)
> err = parse_events(evlist, "cpu-clock", NULL);
> if (err)
> goto out;
> + err = TEST_FAIL;
>
> machines__init();
>
> /* setup threads/dso/map/symbols also */
> machine = setup_fake_machine();
> - if (!machine)
> + if (IS_ERR(machine)) {
> + if (PTR_ERR(machine) == -EACCES)
> + err = TEST_SKIP;
> goto out;
> + }
>
> if (verbose > 1)
> machine__fprintf(machine, stderr);
> diff --git a/tools/perf/tests/hists_filter.c b/tools/perf/tests/hists_filter.c
> index ccb5b49..795a04d 100644
> --- a/tools/perf/tests/hists_filter.c
> +++ b/tools/perf/tests/hists_filter.c
> @@ -120,6 +120,7 @@ int test__hists_filter(int subtest __maybe_unused)
> err = parse_events(evlist, "task-clock", NULL);
> if (err)
> goto out;
> + err = TEST_FAIL;
>
> /* default sort order (comm,dso,sym) will be used */
> if (setup_sorting() < 0)
> @@ -129,8 +130,11 @@
[PATCH] perf tests: Fix false TEST_OK result for 'perf test hist'
Commit 71d6de64feddd4b45326fba2111b3006d9e0 ('perf test: Fix hist
testcases when kptr_restrict is on') solves a double free problem when
'perf test hist' calling setup_fake_machine(). However, the result is
still incorrect. For example:
$ ./perf test -v 'filtering hist entries'
25: Test filtering hist entries :
--- start ---
test child forked, pid 4186
Cannot create kernel maps
test child finished with 0
end
Test filtering hist entries: Ok
In this case the body of this test is not get executed at all, but the
result is 'Ok'. I think 'Skip' should be better.
The reason is that *ALL* hists test cases forget to reset err after
using it to hold an error code.
This patch ensure err is reset.
In case when kptr_restrict prevent normal user get kernel address, this
test should be skipped, not fail. This patch uses linux/err.h to store
error code in return value of setup_fake_machine(), and let 'EACCES' to
indicate this problem.
Signed-off-by: Wang Nan
Cc: Arnaldo Carvalho de Melo
Cc: Jiri Olsa
Cc: Masami Hiramatsu
Cc: Namhyung Kim
---
tools/perf/tests/hists_common.c | 12 +---
tools/perf/tests/hists_common.h | 1 +
tools/perf/tests/hists_cumulate.c | 6 +-
tools/perf/tests/hists_filter.c | 6 +-
tools/perf/tests/hists_link.c | 6 +-
tools/perf/tests/hists_output.c | 6 +-
6 files changed, 30 insertions(+), 7 deletions(-)
diff --git a/tools/perf/tests/hists_common.c b/tools/perf/tests/hists_common.c
index bcfd081..2eca1b2 100644
--- a/tools/perf/tests/hists_common.c
+++ b/tools/perf/tests/hists_common.c
@@ -81,15 +81,20 @@ struct machine *setup_fake_machine(struct machines
*machines)
{
struct machine *machine = machines__find(machines, HOST_KERNEL_ID);
size_t i;
+ int err = -ENOMEM;
if (machine == NULL) {
pr_debug("Not enough memory for machine setup\n");
- return NULL;
+ return ERR_PTR(-ENOMEM);
}
if (machine__create_kernel_maps(machine)) {
pr_debug("Cannot create kernel maps\n");
- return NULL;
+ if (symbol_conf.kptr_restrict) {
+ pr_debug("Hint: Check
/proc/sys/kernel/kptr_restrict.\n");
+ err = -EACCES;
+ }
+ return ERR_PTR(err);
}
for (i = 0; i < ARRAY_SIZE(fake_threads); i++) {
@@ -155,7 +160,8 @@ struct machine *setup_fake_machine(struct machines
*machines)
out:
pr_debug("Not enough memory for machine setup\n");
machine__delete_threads(machine);
- return NULL;
+ machine__exit(machine);
+ return ERR_PTR(err);
}
void print_hists_in(struct hists *hists)
diff --git a/tools/perf/tests/hists_common.h b/tools/perf/tests/hists_common.h
index 888254e..0252eae 100644
--- a/tools/perf/tests/hists_common.h
+++ b/tools/perf/tests/hists_common.h
@@ -1,5 +1,6 @@
#ifndef __PERF_TESTS__HISTS_COMMON_H__
#define __PERF_TESTS__HISTS_COMMON_H__
+#include
struct machine;
struct machines;
diff --git a/tools/perf/tests/hists_cumulate.c
b/tools/perf/tests/hists_cumulate.c
index 8292948..c211075 100644
--- a/tools/perf/tests/hists_cumulate.c
+++ b/tools/perf/tests/hists_cumulate.c
@@ -706,13 +706,17 @@ int test__hists_cumulate(int subtest __maybe_unused)
err = parse_events(evlist, "cpu-clock", NULL);
if (err)
goto out;
+ err = TEST_FAIL;
machines__init();
/* setup threads/dso/map/symbols also */
machine = setup_fake_machine();
- if (!machine)
+ if (IS_ERR(machine)) {
+ if (PTR_ERR(machine) == -EACCES)
+ err = TEST_SKIP;
goto out;
+ }
if (verbose > 1)
machine__fprintf(machine, stderr);
diff --git a/tools/perf/tests/hists_filter.c b/tools/perf/tests/hists_filter.c
index ccb5b49..795a04d 100644
--- a/tools/perf/tests/hists_filter.c
+++ b/tools/perf/tests/hists_filter.c
@@ -120,6 +120,7 @@ int test__hists_filter(int subtest __maybe_unused)
err = parse_events(evlist, "task-clock", NULL);
if (err)
goto out;
+ err = TEST_FAIL;
/* default sort order (comm,dso,sym) will be used */
if (setup_sorting() < 0)
@@ -129,8 +130,11 @@ int test__hists_filter(int subtest __maybe_unused)
/* setup threads/dso/map/symbols also */
machine = setup_fake_machine();
- if (!machine)
+ if (IS_ERR(machine)) {
+ if (PTR_ERR(machine) == -EACCES)
+ err = TEST_SKIP;
goto out;
+ }
if (verbose > 1)
machine__fprintf(machine, stderr);
diff --git a/tools/perf/tests/hists_link.c b/tools/perf/tests/hists_link.c
index 6243e2b..f5f37ef9 100644
--- a/tools/perf/tests/hists_link.c
+++ b/tools/perf/tests/hists_link.c
@@ -293,6 +293,7 @@ int test__hists_link(int subtest
Re: [PATCH] perf tests: Fix false TEST_OK result for 'perf test hist'
Hi,
On Tue, Dec 15, 2015 at 04:18:44AM +, Wang Nan wrote:
> Commit 71d6de64feddd4b45326fba2111b3006d9e0 ('perf test: Fix hist
> testcases when kptr_restrict is on') solves a double free problem when
> 'perf test hist' calling setup_fake_machine(). However, the result is
> still incorrect. For example:
>
> $ ./perf test -v 'filtering hist entries'
> 25: Test filtering hist entries :
> --- start ---
> test child forked, pid 4186
> Cannot create kernel maps
> test child finished with 0
> end
> Test filtering hist entries: Ok
>
> In this case the body of this test is not get executed at all, but the
> result is 'Ok'. I think 'Skip' should be better.
Right, I missed this part.
But on second thought, I think it should not create real kernel maps.
What setup_fake_machine() does is to create/insert fake
dso/mmap/symbols for testing hist functions. So it has to use the
predefined info not real ones.
>
> The reason is that *ALL* hists test cases forget to reset err after
> using it to hold an error code.
>
> This patch ensure err is reset.
>
> In case when kptr_restrict prevent normal user get kernel address, this
> test should be skipped, not fail. This patch uses linux/err.h to store
> error code in return value of setup_fake_machine(), and let 'EACCES' to
> indicate this problem.
Please just remove the machine__create_kernel_maps() in
setup_fake_machine() and simply return TEST_FAIL if there's an error.
Thanks,
Namhyung
>
> Signed-off-by: Wang Nan
> Cc: Arnaldo Carvalho de Melo
> Cc: Jiri Olsa
> Cc: Masami Hiramatsu
> Cc: Namhyung Kim
> ---
> tools/perf/tests/hists_common.c | 12 +---
> tools/perf/tests/hists_common.h | 1 +
> tools/perf/tests/hists_cumulate.c | 6 +-
> tools/perf/tests/hists_filter.c | 6 +-
> tools/perf/tests/hists_link.c | 6 +-
> tools/perf/tests/hists_output.c | 6 +-
> 6 files changed, 30 insertions(+), 7 deletions(-)
>
> diff --git a/tools/perf/tests/hists_common.c b/tools/perf/tests/hists_common.c
> index bcfd081..2eca1b2 100644
> --- a/tools/perf/tests/hists_common.c
> +++ b/tools/perf/tests/hists_common.c
> @@ -81,15 +81,20 @@ struct machine *setup_fake_machine(struct machines
> *machines)
> {
> struct machine *machine = machines__find(machines, HOST_KERNEL_ID);
> size_t i;
> + int err = -ENOMEM;
>
> if (machine == NULL) {
> pr_debug("Not enough memory for machine setup\n");
> - return NULL;
> + return ERR_PTR(-ENOMEM);
> }
>
> if (machine__create_kernel_maps(machine)) {
> pr_debug("Cannot create kernel maps\n");
> - return NULL;
> + if (symbol_conf.kptr_restrict) {
> + pr_debug("Hint: Check
> /proc/sys/kernel/kptr_restrict.\n");
> + err = -EACCES;
> + }
> + return ERR_PTR(err);
> }
>
> for (i = 0; i < ARRAY_SIZE(fake_threads); i++) {
> @@ -155,7 +160,8 @@ struct machine *setup_fake_machine(struct machines
> *machines)
> out:
> pr_debug("Not enough memory for machine setup\n");
> machine__delete_threads(machine);
> - return NULL;
> + machine__exit(machine);
> + return ERR_PTR(err);
> }
>
> void print_hists_in(struct hists *hists)
> diff --git a/tools/perf/tests/hists_common.h b/tools/perf/tests/hists_common.h
> index 888254e..0252eae 100644
> --- a/tools/perf/tests/hists_common.h
> +++ b/tools/perf/tests/hists_common.h
> @@ -1,5 +1,6 @@
> #ifndef __PERF_TESTS__HISTS_COMMON_H__
> #define __PERF_TESTS__HISTS_COMMON_H__
> +#include
>
> struct machine;
> struct machines;
> diff --git a/tools/perf/tests/hists_cumulate.c
> b/tools/perf/tests/hists_cumulate.c
> index 8292948..c211075 100644
> --- a/tools/perf/tests/hists_cumulate.c
> +++ b/tools/perf/tests/hists_cumulate.c
> @@ -706,13 +706,17 @@ int test__hists_cumulate(int subtest __maybe_unused)
> err = parse_events(evlist, "cpu-clock", NULL);
> if (err)
> goto out;
> + err = TEST_FAIL;
>
> machines__init();
>
> /* setup threads/dso/map/symbols also */
> machine = setup_fake_machine();
> - if (!machine)
> + if (IS_ERR(machine)) {
> + if (PTR_ERR(machine) == -EACCES)
> + err = TEST_SKIP;
> goto out;
> + }
>
> if (verbose > 1)
> machine__fprintf(machine, stderr);
> diff --git a/tools/perf/tests/hists_filter.c b/tools/perf/tests/hists_filter.c
> index ccb5b49..795a04d 100644
> --- a/tools/perf/tests/hists_filter.c
> +++ b/tools/perf/tests/hists_filter.c
> @@ -120,6 +120,7 @@ int test__hists_filter(int subtest __maybe_unused)
> err = parse_events(evlist, "task-clock", NULL);
> if (err)
> goto out;
> + err = TEST_FAIL;
>
>
[PATCH] perf tests: Fix false TEST_OK result for 'perf test hist'
Commit 71d6de64feddd4b45326fba2111b3006d9e0 ('perf test: Fix hist
testcases when kptr_restrict is on') solves a double free problem when
'perf test hist' calling setup_fake_machine(). However, the result is
still incorrect. For example:
$ ./perf test -v 'filtering hist entries'
25: Test filtering hist entries :
--- start ---
test child forked, pid 4186
Cannot create kernel maps
test child finished with 0
end
Test filtering hist entries: Ok
In this case the body of this test is not get executed at all, but the
result is 'Ok'. I think 'Skip' should be better.
The reason is that *ALL* hists test cases forget to reset err after
using it to hold an error code.
This patch ensure err is reset.
In case when kptr_restrict prevent normal user get kernel address, this
test should be skipped, not fail. This patch uses linux/err.h to store
error code in return value of setup_fake_machine(), and let 'EACCES' to
indicate this problem.
Signed-off-by: Wang Nan
Cc: Arnaldo Carvalho de Melo
Cc: Jiri Olsa
Cc: Masami Hiramatsu
Cc: Namhyung Kim
---
tools/perf/tests/hists_common.c | 12 +---
tools/perf/tests/hists_common.h | 1 +
tools/perf/tests/hists_cumulate.c | 6 +-
tools/perf/tests/hists_filter.c | 6 +-
tools/perf/tests/hists_link.c | 6 +-
tools/perf/tests/hists_output.c | 6 +-
6 files changed, 30 insertions(+), 7 deletions(-)
diff --git a/tools/perf/tests/hists_common.c b/tools/perf/tests/hists_common.c
index bcfd081..2eca1b2 100644
--- a/tools/perf/tests/hists_common.c
+++ b/tools/perf/tests/hists_common.c
@@ -81,15 +81,20 @@ struct machine *setup_fake_machine(struct machines
*machines)
{
struct machine *machine = machines__find(machines, HOST_KERNEL_ID);
size_t i;
+ int err = -ENOMEM;
if (machine == NULL) {
pr_debug("Not enough memory for machine setup\n");
- return NULL;
+ return ERR_PTR(-ENOMEM);
}
if (machine__create_kernel_maps(machine)) {
pr_debug("Cannot create kernel maps\n");
- return NULL;
+ if (symbol_conf.kptr_restrict) {
+ pr_debug("Hint: Check
/proc/sys/kernel/kptr_restrict.\n");
+ err = -EACCES;
+ }
+ return ERR_PTR(err);
}
for (i = 0; i < ARRAY_SIZE(fake_threads); i++) {
@@ -155,7 +160,8 @@ struct machine *setup_fake_machine(struct machines
*machines)
out:
pr_debug("Not enough memory for machine setup\n");
machine__delete_threads(machine);
- return NULL;
+ machine__exit(machine);
+ return ERR_PTR(err);
}
void print_hists_in(struct hists *hists)
diff --git a/tools/perf/tests/hists_common.h b/tools/perf/tests/hists_common.h
index 888254e..0252eae 100644
--- a/tools/perf/tests/hists_common.h
+++ b/tools/perf/tests/hists_common.h
@@ -1,5 +1,6 @@
#ifndef __PERF_TESTS__HISTS_COMMON_H__
#define __PERF_TESTS__HISTS_COMMON_H__
+#include
struct machine;
struct machines;
diff --git a/tools/perf/tests/hists_cumulate.c
b/tools/perf/tests/hists_cumulate.c
index 8292948..c211075 100644
--- a/tools/perf/tests/hists_cumulate.c
+++ b/tools/perf/tests/hists_cumulate.c
@@ -706,13 +706,17 @@ int test__hists_cumulate(int subtest __maybe_unused)
err = parse_events(evlist, "cpu-clock", NULL);
if (err)
goto out;
+ err = TEST_FAIL;
machines__init();
/* setup threads/dso/map/symbols also */
machine = setup_fake_machine();
- if (!machine)
+ if (IS_ERR(machine)) {
+ if (PTR_ERR(machine) == -EACCES)
+ err = TEST_SKIP;
goto out;
+ }
if (verbose > 1)
machine__fprintf(machine, stderr);
diff --git a/tools/perf/tests/hists_filter.c b/tools/perf/tests/hists_filter.c
index ccb5b49..795a04d 100644
--- a/tools/perf/tests/hists_filter.c
+++ b/tools/perf/tests/hists_filter.c
@@ -120,6 +120,7 @@ int test__hists_filter(int subtest __maybe_unused)
err = parse_events(evlist, "task-clock", NULL);
if (err)
goto out;
+ err = TEST_FAIL;
/* default sort order (comm,dso,sym) will be used */
if (setup_sorting() < 0)
@@ -129,8 +130,11 @@ int test__hists_filter(int subtest __maybe_unused)
/* setup threads/dso/map/symbols also */
machine = setup_fake_machine();
- if (!machine)
+ if (IS_ERR(machine)) {
+ if (PTR_ERR(machine) == -EACCES)
+ err = TEST_SKIP;
goto out;
+ }
if (verbose > 1)
machine__fprintf(machine, stderr);
diff --git a/tools/perf/tests/hists_link.c b/tools/perf/tests/hists_link.c
index 6243e2b..f5f37ef9 100644
---
