From: Adhemerval Zanella <adhemerval.zanella@linaro.org>
To: Lukasz Majewski <lukma@denx.de>
Cc: GNU C Library <libc-alpha@sourceware.org>
Subject: Re: [PATCH] tst: Provide test for getrusage
Date: Mon, 22 Mar 2021 16:47:25 -0300 [thread overview]
Message-ID: <85d5e1b7-46c7-5904-0856-fc5f90493295@linaro.org> (raw)
In-Reply-To: <20210315104223.32542-1-lukma@denx.de>
On 15/03/2021 07:42, Lukasz Majewski wrote:
> This commit adds test to check if getrusage provides correct time
> accounting for some busy wait loop.
>
> To be more specific - the time is measured in a loop by wrappers
> from glibc's libsupport and checked if timeout has passed.
>
> Afterwards, the data from struct rusage, namely ru_stime and ru_utime
> is summed up and compared with time from busy waiting.
> ---
> sysdeps/unix/sysv/linux/Makefile | 2 +-
> sysdeps/unix/sysv/linux/tst-getrusage.c | 55 +++++++++++++++++++++++++
> 2 files changed, 56 insertions(+), 1 deletion(-)
> create mode 100644 sysdeps/unix/sysv/linux/tst-getrusage.c
>
> diff --git a/sysdeps/unix/sysv/linux/Makefile b/sysdeps/unix/sysv/linux/Makefile
> index 6a45ac820a..a93fa4ca50 100644
> --- a/sysdeps/unix/sysv/linux/Makefile
> +++ b/sysdeps/unix/sysv/linux/Makefile
> @@ -113,7 +113,7 @@ tests += tst-clone tst-clone2 tst-clone3 tst-fanotify tst-personality \
> tst-tgkill tst-sysvsem-linux tst-sysvmsg-linux tst-sysvshm-linux \
> tst-timerfd tst-ppoll tst-futimens tst-utime tst-utimes \
> tst-clock_adjtime tst-adjtimex tst-ntp_adjtime tst-futimes \
> - tst-lutimes tst-ntp_gettime tst-ntp_gettimex
> + tst-lutimes tst-ntp_gettime tst-ntp_gettimex tst-getrusage
> tests-internal += tst-ofdlocks-compat tst-sigcontext-get_pc
> tests-time64 += \
> tst-futimes-time64\
> diff --git a/sysdeps/unix/sysv/linux/tst-getrusage.c b/sysdeps/unix/sysv/linux/tst-getrusage.c
> new file mode 100644
> index 0000000000..068becdadb
> --- /dev/null
> +++ b/sysdeps/unix/sysv/linux/tst-getrusage.c
> @@ -0,0 +1,55 @@
> +/* Test for getrusage
> + Copyright (C) 2021 Free Software Foundation, Inc.
> + This file is part of the GNU C Library.
> +
> + The GNU C Library is free software; you can redistribute it and/or
> + modify it under the terms of the GNU Lesser General Public
> + License as published by the Free Software Foundation; either
> + version 2.1 of the License, or (at your option) any later version.
> +
> + The GNU C Library is distributed in the hope that it will be useful,
> + but WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + Lesser General Public License for more details.
> +
> + You should have received a copy of the GNU Lesser General Public
> + License along with the GNU C Library; if not, see
> + <https://www.gnu.org/licenses/>. */
> +
> +#include <time.h>
> +#include <sys/resource.h>
> +#include <support/check.h>
> +#include <support/timespec.h>
> +
> +#define TST_GETRUSAGE_TIMEOUT 1
> +
> +static int
> +do_test (void)
> +{
> + struct timespec tv_then, tv_now, gru_u, gru_s, s, r;
> + struct rusage ru;
> + int ret;
> +
> + tv_then = xclock_now (CLOCK_REALTIME);
> + /* Busy wait for 1 second. */
> + do
> + {
> + tv_now = xclock_now (CLOCK_REALTIME);
> + r = timespec_sub (tv_now, tv_then);
> + } while ( r.tv_sec != TST_GETRUSAGE_TIMEOUT );
> +
> + ret = getrusage (RUSAGE_SELF, &ru);
> + if (ret == -1)
> + FAIL_EXIT1 ("getrusage failed: %m\n");
> +
> + gru_u = make_timespec (ru.ru_utime.tv_sec, ru.ru_utime.tv_usec * 1000);
> + gru_s = make_timespec (ru.ru_stime.tv_sec, ru.ru_stime.tv_usec * 1000);
> + s = timespec_add (gru_u, gru_s);
> +
> + TEST_COMPARE (support_timespec_check_in_range
> + ((struct timespec) { 1, 0 }, s, 0.8, 1.2), 1);
> +
> + return 0;
> +}
> +
> +#include <support/test-driver.c>
>
Timer accounting tests are tricky to get right and prone to show a
lot of false positive failures and has bitten us in the past (check
04deeaa9ea74 and 9a29f1a2ae3d4b).
So I think we should be as simple as possible and take in consideration
that tests will run a timer shared machine so even with a busy wait the
result timer might not take in consideration the desirable timeout.
So for geturage (which afaik is also deprecated on future POSIX) I
more simpler tests would be to:
1. Check for invalid 'who' (the EFAULT result from man-pages only
trigger where syscall is issued directly, which is not always
the case for glibc).
2. Issue getusage and check if the ru_utime.tv_nsec and
ru_stime.tv_nsec are within the microseconds range.
3. Do some very short busy wait (10ms) and check if the
ru_utime.tv_sec and ru_stime.tv_sec are equal or larger
than the previous one.
Also, I think this tests should generic, so move to resource
folder.
prev parent reply other threads:[~2021-03-22 19:47 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-15 10:42 Lukasz Majewski
2021-03-15 10:55 ` Andreas Schwab
2021-03-15 13:56 ` Lukasz Majewski
2021-03-17 16:04 ` Lukasz Majewski
2021-03-17 16:24 ` Andreas Schwab
2021-03-22 11:29 ` Lukasz Majewski
2021-03-22 19:47 ` Adhemerval Zanella [this message]
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=85d5e1b7-46c7-5904-0856-fc5f90493295@linaro.org \
--to=adhemerval.zanella@linaro.org \
--cc=libc-alpha@sourceware.org \
--cc=lukma@denx.de \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).