* [LTP] [PATCH] clocks/invaliddates.c: relax acceptable delta
@ 2020-06-18 10:24 Jan Stancek
2020-06-18 20:37 ` Petr Vorel
` (2 more replies)
0 siblings, 3 replies; 6+ messages in thread
From: Jan Stancek @ 2020-06-18 10:24 UTC (permalink / raw)
To: ltp
Test allows just 5ms delta for PASS, and test randomly fails in
environments with shared resources and increased steal time.
Relax the condition and also print deltas if test fails.
Remove DEBUG ifdefs and main parameters to avoid unused variable
warning.
Signed-off-by: Jan Stancek <jstancek@redhat.com>
---
.../functional/timers/clocks/invaliddates.c | 25 +++++++------------
1 file changed, 9 insertions(+), 16 deletions(-)
diff --git a/testcases/open_posix_testsuite/functional/timers/clocks/invaliddates.c b/testcases/open_posix_testsuite/functional/timers/clocks/invaliddates.c
index face334fd250..d4116b1e9bc0 100644
--- a/testcases/open_posix_testsuite/functional/timers/clocks/invaliddates.c
+++ b/testcases/open_posix_testsuite/functional/timers/clocks/invaliddates.c
@@ -18,8 +18,7 @@
#define NUMTESTS 5
-#define ACCEPTABLESECDELTA 0
-#define ACCEPTABLENSECDELTA 5000000
+#define ACCEPTABLESECDELTA 2
static int testtimes[NUMTESTS][2] = {
{INT32_MAX, 999999999}, // large number
@@ -29,7 +28,7 @@ static int testtimes[NUMTESTS][2] = {
{1049623200, 999999999}, // daylight savings 2003
};
-int main(int argc, char *argv[])
+int main(void)
{
struct timespec tpset, tpget, tsreset;
int secdelta, nsecdelta;
@@ -44,9 +43,6 @@ int main(int argc, char *argv[])
for (i = 0; i < NUMTESTS; i++) {
tpset.tv_sec = testtimes[i][0];
tpset.tv_nsec = testtimes[i][1];
-#ifdef DEBUG
- printf("Test: %ds %dns\n", testtimes[i][0], testtimes[i][1]);
-#endif
if (clock_settime(CLOCK_REALTIME, &tpset) == 0) {
if (clock_gettime(CLOCK_REALTIME, &tpget) == -1) {
printf("Error in clock_gettime()\n");
@@ -58,16 +54,13 @@ int main(int argc, char *argv[])
nsecdelta = nsecdelta + 1000000000;
secdelta = secdelta - 1;
}
-#ifdef DEBUG
- printf("Delta: %ds %dns\n", secdelta, nsecdelta);
-#endif
- if ((secdelta > ACCEPTABLESECDELTA) || (secdelta < 0)) {
- printf("clock does not appear to be set\n");
- failure = 1;
- }
- if ((nsecdelta > ACCEPTABLENSECDELTA) ||
- (nsecdelta < 0)) {
- printf("clock does not appear to be set\n");
+
+ if ((secdelta > ACCEPTABLESECDELTA)
+ || (secdelta < 0)) {
+ printf("FAIL: test(%d,%d), secdelta: %d,"
+ " nsecdelta: %d\n",
+ testtimes[i][0], testtimes[i][1],
+ secdelta, nsecdelta);
failure = 1;
}
} else {
--
2.18.1
^ permalink raw reply related [flat|nested] 6+ messages in thread* [LTP] [PATCH] clocks/invaliddates.c: relax acceptable delta
2020-06-18 10:24 [LTP] [PATCH] clocks/invaliddates.c: relax acceptable delta Jan Stancek
@ 2020-06-18 20:37 ` Petr Vorel
2020-06-19 1:31 ` Li Wang
2020-06-19 2:25 ` [LTP] 答复: " Xu, Yang
2 siblings, 0 replies; 6+ messages in thread
From: Petr Vorel @ 2020-06-18 20:37 UTC (permalink / raw)
To: ltp
Hi Jan,
> Test allows just 5ms delta for PASS, and test randomly fails in
> environments with shared resources and increased steal time.
Good.
> Relax the condition and also print deltas if test fails.
> Remove DEBUG ifdefs and main parameters to avoid unused variable
> warning.
Also thanks for cleanup.
Reviewed-by: Petr Vorel <pvorel@suse.cz>
Kind regards,
Petr
^ permalink raw reply [flat|nested] 6+ messages in thread
* [LTP] [PATCH] clocks/invaliddates.c: relax acceptable delta
2020-06-18 10:24 [LTP] [PATCH] clocks/invaliddates.c: relax acceptable delta Jan Stancek
2020-06-18 20:37 ` Petr Vorel
@ 2020-06-19 1:31 ` Li Wang
2020-06-19 1:54 ` Li Wang
2020-06-19 2:25 ` [LTP] 答复: " Xu, Yang
2 siblings, 1 reply; 6+ messages in thread
From: Li Wang @ 2020-06-19 1:31 UTC (permalink / raw)
To: ltp
On Thu, Jun 18, 2020 at 6:24 PM Jan Stancek <jstancek@redhat.com> wrote:
> Test allows just 5ms delta for PASS, and test randomly fails in
> environments with shared resources and increased steal time.
>
> Relax the condition and also print deltas if test fails.
> Remove DEBUG ifdefs and main parameters to avoid unused variable
> warning.
>
> Signed-off-by: Jan Stancek <jstancek@redhat.com>
> ---
> .../functional/timers/clocks/invaliddates.c | 25 +++++++------------
> 1 file changed, 9 insertions(+), 16 deletions(-)
>
> diff --git
> a/testcases/open_posix_testsuite/functional/timers/clocks/invaliddates.c
> b/testcases/open_posix_testsuite/functional/timers/clocks/invaliddates.c
> index face334fd250..d4116b1e9bc0 100644
> ---
> a/testcases/open_posix_testsuite/functional/timers/clocks/invaliddates.c
> +++
> b/testcases/open_posix_testsuite/functional/timers/clocks/invaliddates.c
> @@ -18,8 +18,7 @@
>
> #define NUMTESTS 5
>
> -#define ACCEPTABLESECDELTA 0
> -#define ACCEPTABLENSECDELTA 5000000
> +#define ACCEPTABLESECDELTA 2
>
It's hard to say what size of the second-delta is a proper value, but to
increase it can obviously decrease the failure probability. If there is no
better way I'd go with this patch.
Reviewed-by: Li Wang <liwang@redhat.com>
>
> static int testtimes[NUMTESTS][2] = {
> {INT32_MAX, 999999999}, // large number
> @@ -29,7 +28,7 @@ static int testtimes[NUMTESTS][2] = {
> {1049623200, 999999999}, // daylight savings 2003
> };
>
> -int main(int argc, char *argv[])
> +int main(void)
> {
> struct timespec tpset, tpget, tsreset;
> int secdelta, nsecdelta;
> @@ -44,9 +43,6 @@ int main(int argc, char *argv[])
> for (i = 0; i < NUMTESTS; i++) {
> tpset.tv_sec = testtimes[i][0];
> tpset.tv_nsec = testtimes[i][1];
> -#ifdef DEBUG
> - printf("Test: %ds %dns\n", testtimes[i][0],
> testtimes[i][1]);
> -#endif
> if (clock_settime(CLOCK_REALTIME, &tpset) == 0) {
> if (clock_gettime(CLOCK_REALTIME, &tpget) == -1) {
> printf("Error in clock_gettime()\n");
> @@ -58,16 +54,13 @@ int main(int argc, char *argv[])
> nsecdelta = nsecdelta + 1000000000;
> secdelta = secdelta - 1;
> }
> -#ifdef DEBUG
> - printf("Delta: %ds %dns\n", secdelta, nsecdelta);
> -#endif
> - if ((secdelta > ACCEPTABLESECDELTA) || (secdelta <
> 0)) {
> - printf("clock does not appear to be
> set\n");
> - failure = 1;
> - }
> - if ((nsecdelta > ACCEPTABLENSECDELTA) ||
> - (nsecdelta < 0)) {
> - printf("clock does not appear to be
> set\n");
> +
> + if ((secdelta > ACCEPTABLESECDELTA)
> + || (secdelta < 0)) {
> + printf("FAIL: test(%d,%d), secdelta: %d,"
> + " nsecdelta: %d\n",
> + testtimes[i][0], testtimes[i][1],
> + secdelta, nsecdelta);
> failure = 1;
> }
> } else {
> --
> 2.18.1
>
>
--
Regards,
Li Wang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linux.it/pipermail/ltp/attachments/20200619/1ffe0c9f/attachment-0001.htm>
^ permalink raw reply [flat|nested] 6+ messages in thread* [LTP] [PATCH] clocks/invaliddates.c: relax acceptable delta
2020-06-19 1:31 ` Li Wang
@ 2020-06-19 1:54 ` Li Wang
0 siblings, 0 replies; 6+ messages in thread
From: Li Wang @ 2020-06-19 1:54 UTC (permalink / raw)
To: ltp
Li Wang <liwang@redhat.com> wrote:
> On Thu, Jun 18, 2020 at 6:24 PM Jan Stancek <jstancek@redhat.com> wrote:
>
>> Test allows just 5ms delta for PASS, and test randomly fails in
>> environments with shared resources and increased steal time.
>>
>> Relax the condition and also print deltas if test fails.
>> Remove DEBUG ifdefs and main parameters to avoid unused variable
>> warning.
>>
>> Signed-off-by: Jan Stancek <jstancek@redhat.com>
>> ---
>> .../functional/timers/clocks/invaliddates.c | 25 +++++++------------
>> 1 file changed, 9 insertions(+), 16 deletions(-)
>>
>> diff --git
>> a/testcases/open_posix_testsuite/functional/timers/clocks/invaliddates.c
>> b/testcases/open_posix_testsuite/functional/timers/clocks/invaliddates.c
>> index face334fd250..d4116b1e9bc0 100644
>> ---
>> a/testcases/open_posix_testsuite/functional/timers/clocks/invaliddates.c
>> +++
>> b/testcases/open_posix_testsuite/functional/timers/clocks/invaliddates.c
>> @@ -18,8 +18,7 @@
>>
>> #define NUMTESTS 5...
>>
>> -#define ACCEPTABLESECDELTA 0
>> -#define ACCEPTABLENSECDELTA 5000000
>> +#define ACCEPTABLESECDELTA 2
>>
>
> It's hard to say what size of the second-delta is a proper value, but to
> increase it can obviously decrease the failure probability. If there is no
> better way I'd go with this patch.
>
But in another word, this test is just focused on invalid dates[1], so
the deviation(caused by sharing resource) in seconds(even 10 seconds) bound
is acceptable I think.
If someone insists to make thing more precise, we can make use of
function tst_is_virt() to apply this method only for virtual machines. But
anyway that's not very important:).
[1].
static int testtimes[NUMTESTS][2] = {
{INT32_MAX, 999999999}, // large number
{946713600, 999999999}, // Y2K - Jan 1, 2000
{951811200, 999999999}, // Y2K - Feb 29, 2000
{1078041600, 999999999}, // Y2K - Feb 29, 2004
{1049623200, 999999999}, // daylight savings 2003
};
--
Regards,
Li Wang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linux.it/pipermail/ltp/attachments/20200619/32028224/attachment.htm>
^ permalink raw reply [flat|nested] 6+ messages in thread
* [LTP] 答复: [PATCH] clocks/invaliddates.c: relax acceptable delta
2020-06-18 10:24 [LTP] [PATCH] clocks/invaliddates.c: relax acceptable delta Jan Stancek
2020-06-18 20:37 ` Petr Vorel
2020-06-19 1:31 ` Li Wang
@ 2020-06-19 2:25 ` Xu, Yang
2020-06-19 6:37 ` Jan Stancek
2 siblings, 1 reply; 6+ messages in thread
From: Xu, Yang @ 2020-06-19 2:25 UTC (permalink / raw)
To: ltp
Hi Jan
Acked-by.
ps: I like leaving a comment in delta, just my personal habit.
Best Regards
Yang Xu
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2020-06-19 6:37 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2020-06-18 10:24 [LTP] [PATCH] clocks/invaliddates.c: relax acceptable delta Jan Stancek
2020-06-18 20:37 ` Petr Vorel
2020-06-19 1:31 ` Li Wang
2020-06-19 1:54 ` Li Wang
2020-06-19 2:25 ` [LTP] 答复: " Xu, Yang
2020-06-19 6:37 ` Jan Stancek
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox