* [PATCH] selftests/seccomp: Flush benchmark output regularly
@ 2021-05-27 3:30 Kees Cook
2021-05-27 19:05 ` Will Drewry
0 siblings, 1 reply; 3+ messages in thread
From: Kees Cook @ 2021-05-27 3:30 UTC (permalink / raw)
To: linux-kernel; +Cc: Kees Cook, Andy Lutomirski, Will Drewry
When running the seccomp benchmark under a test runner, it wouldn't
provide any feedback on progress. Force IO flushes during the test.
Signed-off-by: Kees Cook <keescook@chromium.org>
---
tools/testing/selftests/seccomp/seccomp_benchmark.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/tools/testing/selftests/seccomp/seccomp_benchmark.c b/tools/testing/selftests/seccomp/seccomp_benchmark.c
index fcc806585266..7be1e2131f79 100644
--- a/tools/testing/selftests/seccomp/seccomp_benchmark.c
+++ b/tools/testing/selftests/seccomp/seccomp_benchmark.c
@@ -42,6 +42,7 @@ unsigned long long timing(clockid_t clk_id, unsigned long long samples)
finish.tv_sec, finish.tv_nsec,
start.tv_sec, start.tv_nsec,
i, (double)i / 1000000000.0);
+ fflush(NULL);
return i;
}
@@ -54,6 +55,7 @@ unsigned long long calibrate(void)
int seconds = 15;
printf("Calibrating sample size for %d seconds worth of syscalls ...\n", seconds);
+ fflush(NULL);
samples = 0;
pid = getpid();
@@ -157,6 +159,7 @@ int main(int argc, char *argv[])
/* Native call */
native = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples;
printf("getpid native: %llu ns\n", native);
+ fflush(NULL);
ret = prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0);
assert(ret == 0);
@@ -167,6 +170,7 @@ int main(int argc, char *argv[])
bitmap1 = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples;
printf("getpid RET_ALLOW 1 filter (bitmap): %llu ns\n", bitmap1);
+ fflush(NULL);
/* Second filter resulting in a bitmap */
ret = prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &bitmap_prog);
@@ -174,6 +178,7 @@ int main(int argc, char *argv[])
bitmap2 = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples;
printf("getpid RET_ALLOW 2 filters (bitmap): %llu ns\n", bitmap2);
+ fflush(NULL);
/* Third filter, can no longer be converted to bitmap */
ret = prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &prog);
@@ -181,6 +186,7 @@ int main(int argc, char *argv[])
filter1 = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples;
printf("getpid RET_ALLOW 3 filters (full): %llu ns\n", filter1);
+ fflush(NULL);
/* Fourth filter, can not be converted to bitmap because of filter 3 */
ret = prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &bitmap_prog);
@@ -188,6 +194,7 @@ int main(int argc, char *argv[])
filter2 = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples;
printf("getpid RET_ALLOW 4 filters (full): %llu ns\n", filter2);
+ fflush(NULL);
/* Estimations */
#define ESTIMATE(fmt, var, what) do { \
--
2.25.1
^ permalink raw reply related [flat|nested] 3+ messages in thread* Re: [PATCH] selftests/seccomp: Flush benchmark output regularly
2021-05-27 3:30 [PATCH] selftests/seccomp: Flush benchmark output regularly Kees Cook
@ 2021-05-27 19:05 ` Will Drewry
2021-05-27 22:20 ` Kees Cook
0 siblings, 1 reply; 3+ messages in thread
From: Will Drewry @ 2021-05-27 19:05 UTC (permalink / raw)
To: Kees Cook; +Cc: LKML, Andy Lutomirski
On Wed, May 26, 2021 at 10:30 PM Kees Cook <keescook@chromium.org> wrote:
>
> When running the seccomp benchmark under a test runner, it wouldn't
> provide any feedback on progress. Force IO flushes during the test.
Would it make sense to just setbuf(stdout, NULL); in main?
> Signed-off-by: Kees Cook <keescook@chromium.org>
> ---
> tools/testing/selftests/seccomp/seccomp_benchmark.c | 7 +++++++
> 1 file changed, 7 insertions(+)
>
> diff --git a/tools/testing/selftests/seccomp/seccomp_benchmark.c b/tools/testing/selftests/seccomp/seccomp_benchmark.c
> index fcc806585266..7be1e2131f79 100644
> --- a/tools/testing/selftests/seccomp/seccomp_benchmark.c
> +++ b/tools/testing/selftests/seccomp/seccomp_benchmark.c
> @@ -42,6 +42,7 @@ unsigned long long timing(clockid_t clk_id, unsigned long long samples)
> finish.tv_sec, finish.tv_nsec,
> start.tv_sec, start.tv_nsec,
> i, (double)i / 1000000000.0);
> + fflush(NULL);
>
> return i;
> }
> @@ -54,6 +55,7 @@ unsigned long long calibrate(void)
> int seconds = 15;
>
> printf("Calibrating sample size for %d seconds worth of syscalls ...\n", seconds);
> + fflush(NULL);
>
> samples = 0;
> pid = getpid();
> @@ -157,6 +159,7 @@ int main(int argc, char *argv[])
> /* Native call */
> native = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples;
> printf("getpid native: %llu ns\n", native);
> + fflush(NULL);
>
> ret = prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0);
> assert(ret == 0);
> @@ -167,6 +170,7 @@ int main(int argc, char *argv[])
>
> bitmap1 = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples;
> printf("getpid RET_ALLOW 1 filter (bitmap): %llu ns\n", bitmap1);
> + fflush(NULL);
>
> /* Second filter resulting in a bitmap */
> ret = prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &bitmap_prog);
> @@ -174,6 +178,7 @@ int main(int argc, char *argv[])
>
> bitmap2 = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples;
> printf("getpid RET_ALLOW 2 filters (bitmap): %llu ns\n", bitmap2);
> + fflush(NULL);
>
> /* Third filter, can no longer be converted to bitmap */
> ret = prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &prog);
> @@ -181,6 +186,7 @@ int main(int argc, char *argv[])
>
> filter1 = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples;
> printf("getpid RET_ALLOW 3 filters (full): %llu ns\n", filter1);
> + fflush(NULL);
>
> /* Fourth filter, can not be converted to bitmap because of filter 3 */
> ret = prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &bitmap_prog);
> @@ -188,6 +194,7 @@ int main(int argc, char *argv[])
>
> filter2 = timing(CLOCK_PROCESS_CPUTIME_ID, samples) / samples;
> printf("getpid RET_ALLOW 4 filters (full): %llu ns\n", filter2);
> + fflush(NULL);
>
> /* Estimations */
> #define ESTIMATE(fmt, var, what) do { \
> --
> 2.25.1
>
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [PATCH] selftests/seccomp: Flush benchmark output regularly
2021-05-27 19:05 ` Will Drewry
@ 2021-05-27 22:20 ` Kees Cook
0 siblings, 0 replies; 3+ messages in thread
From: Kees Cook @ 2021-05-27 22:20 UTC (permalink / raw)
To: Will Drewry; +Cc: LKML, Andy Lutomirski
On Thu, May 27, 2021 at 02:05:38PM -0500, Will Drewry wrote:
> On Wed, May 26, 2021 at 10:30 PM Kees Cook <keescook@chromium.org> wrote:
> >
> > When running the seccomp benchmark under a test runner, it wouldn't
> > provide any feedback on progress. Force IO flushes during the test.
>
> Would it make sense to just setbuf(stdout, NULL); in main?
*forehead slap* Ugh, yes. That works much better. I will respin. Thank
you for the reminder about setbuf() existing. :)
-Kees
--
Kees Cook
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2021-05-27 22:20 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2021-05-27 3:30 [PATCH] selftests/seccomp: Flush benchmark output regularly Kees Cook
2021-05-27 19:05 ` Will Drewry
2021-05-27 22:20 ` Kees Cook
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox