From: Richard Palethorpe <rpalethorpe@suse.de>
To: Li Wang <liwang@redhat.com>
Cc: ltp@lists.linux.it
Subject: Re: [LTP] [PATCH 2/2] getitimer01: add checking for nonzero timer
Date: Mon, 14 Nov 2022 16:02:51 +0000 [thread overview]
Message-ID: <87leodseg2.fsf@suse.de> (raw)
In-Reply-To: <20221112040107.3953862-2-liwang@redhat.com>
Hello,
Li Wang <liwang@redhat.com> writes:
> By default a new process disabled the timer and getitimer()
> returned zero value. But we also need to check if the timer
> is correct when reset to nonzero.
>
> Signed-off-by: Li Wang <liwang@redhat.com>
> ---
>
> Notes:
> The reason for using jiffy instead of time_step is that it only
> checks the timer is set expectedly but not really expired. So we
> use a rough value in the macro definition is enough.
>
> .../kernel/syscalls/getitimer/getitimer01.c | 84 ++++++++++++++++---
> 1 file changed, 71 insertions(+), 13 deletions(-)
>
> diff --git a/testcases/kernel/syscalls/getitimer/getitimer01.c b/testcases/kernel/syscalls/getitimer/getitimer01.c
> index 5ecfac55c..a49f63a85 100644
> --- a/testcases/kernel/syscalls/getitimer/getitimer01.c
> +++ b/testcases/kernel/syscalls/getitimer/getitimer01.c
> @@ -12,25 +12,83 @@
> */
>
> #include "tst_test.h"
> +#include "tst_safe_clocks.h"
>
> -static int itimer_name[] = {
> - ITIMER_REAL,
> - ITIMER_VIRTUAL,
> - ITIMER_PROF,
> +#define SEC 100
> +#define USEC 10000
> +
> +static struct itimerval *value;
> +static long jiffy;
> +
> +static struct tcase {
> + int which;
> + char *des;
> +} tcases[] = {
> + {ITIMER_REAL, "ITIMER_REAL"},
> + {ITIMER_VIRTUAL, "ITIMER_VIRTUAL"},
> + {ITIMER_PROF, "ITIMER_PROF"},
> };
>
> -static void run(void)
> +static void set_setitimer_value(int sec, int usec)
> {
> - long unsigned int i;
> - struct itimerval value;
> + value->it_value.tv_sec = sec;
> + value->it_value.tv_usec = usec;
> + value->it_interval.tv_sec = sec;
> + value->it_interval.tv_usec = usec;
> +}
>
> - for (i = 0; i < ARRAY_SIZE(itimer_name); i++) {
> - TST_EXP_PASS(getitimer(itimer_name[i], &value));
> - TST_EXP_EQ_LI(value.it_value.tv_sec, 0);
> - TST_EXP_EQ_LI(value.it_value.tv_usec, 0);
> - }
> +static void verify_getitimer(unsigned int i)
> +{
> + struct tcase *tc = &tcases[i];
> +
> + tst_res(TINFO, "tc->which = %s", tc->des);
> +
> + TST_EXP_PASS(getitimer(tc->which, value));
> + TST_EXP_EQ_LI(value->it_value.tv_sec, 0);
> + TST_EXP_EQ_LI(value->it_value.tv_usec, 0);
> + TST_EXP_EQ_LI(value->it_interval.tv_sec, 0);
> + TST_EXP_EQ_LI(value->it_interval.tv_usec, 0);
> +
> + set_setitimer_value(SEC, USEC);
> + TST_EXP_PASS(setitimer(tc->which, value, NULL));
> +
> + set_setitimer_value(0, 0);
> + TST_EXP_PASS(getitimer(tc->which, value));
> +
> + TST_EXP_EQ_LI(value->it_interval.tv_sec, SEC);
> + TST_EXP_EQ_LI(value->it_interval.tv_usec, USEC);
> +
> + tst_res(TINFO, "value->it_value.tv_sec=%ld, value->it_value.tv_usec=%ld",
> + value->it_value.tv_sec, value->it_value.tv_usec);
> +
> + /*
> + * ITIMER_VIRTUAL and ITIMER_PROF timers always expire a
> + * TICK_NSEC (jiffy) afterward the elapsed time to make
> + * sure that at least time counters take effect.
> + */
> + long margin = (tc->which == ITIMER_REAL) ? 0 : jiffy;
> +
> + if (value->it_value.tv_sec > SEC ||
> + value->it_value.tv_usec > USEC + margin)
If a second is able to elapse then it_value.tv_usec can be greater than
USEC + margin. Or is there something I am missing?
I guess again there is the issue with wall clock time jumping around.
> + tst_res(TFAIL, "timer value is not within the expected range");
> + else
> + tst_res(TPASS, "timer value is within the expected
> range");
Also we want to print all the relevant values in case of failure.
> +}
> +
> +static void setup(void)
> +{
> + struct timespec time_res;
> +
> + SAFE_CLOCK_GETRES(CLOCK_MONOTONIC_COARSE, &time_res);
> + jiffy = (time_res.tv_nsec + 999) / 1000;
> }
>
> static struct tst_test test = {
> - .test_all = run
> + .tcnt = ARRAY_SIZE(tcases),
> + .setup = setup,
> + .test = verify_getitimer,
> + .bufs = (struct tst_buffers[]) {
> + {&value, .size = sizeof(struct itimerval)},
> + {}
> + }
> };
> --
> 2.35.3
--
Thank you,
Richard.
--
Mailing list info: https://lists.linux.it/listinfo/ltp
next prev parent reply other threads:[~2022-11-14 16:12 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-11-12 4:01 [LTP] [PATCH 1/2] setitimer01: add interval timer test Li Wang
2022-11-12 4:01 ` [LTP] [PATCH 2/2] getitimer01: add checking for nonzero timer Li Wang
2022-11-14 16:02 ` Richard Palethorpe [this message]
2022-11-14 14:32 ` [LTP] [PATCH 1/2] setitimer01: add interval timer test Richard Palethorpe
2022-11-15 4:00 ` Li Wang
2022-11-15 8:44 ` Richard Palethorpe
2022-11-15 10:00 ` Li Wang
2022-11-15 10:08 ` Li Wang
2022-11-15 11:04 ` Richard Palethorpe
2022-11-16 10:25 ` Li Wang
2022-11-15 9:27 ` Andrea Cervesato via ltp
-- strict thread matches above, loose matches on Subject: below --
2022-10-25 12:18 Li Wang
2022-10-25 12:18 ` [LTP] [PATCH 2/2] getitimer01: add checking for nonzero timer Li Wang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=87leodseg2.fsf@suse.de \
--to=rpalethorpe@suse.de \
--cc=liwang@redhat.com \
--cc=ltp@lists.linux.it \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.