* [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants
@ 2020-05-08 4:23 Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 01/17] syscalls/timer_gettime: Add support for time64 tests Viresh Kumar
` (16 more replies)
0 siblings, 17 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
Hi Cyril/Petr,
This replaces all my pending patches that I have sent earlier and not
applied yet. I am sending them all together (along with few new patches)
to avoid any rebase conflicts as reported by Petr.
I have also fixed the GPL licensing thing that Petr asked for at several
files in this V2. Otherwise no difference in the patches.
With this I have completed the time64 support for all the syscalls
tests.
Viresh Kumar (17):
syscalls/timer_gettime: Add support for time64 tests
syscalls/timer_settime: Add support for time64 tests
syscalls/timerfd: Add support for time64 tests
syscalls/sched_rr_get_interval: Add support for time64 tests
syscalls/futex: Merge futex_wait_bitset tests
syscalls/futex: Add support for time64 tests
syscalls/io_pgetevents: Add support for time64 tests
syscalls/sigwaitinfo: Migrate to new test framework
syscalls/rt_sigtimedwait: Add support for time64 tests
syscalls/mq_timed{send|receive}: Add support for time64 tests
syscalls/recvmmsg: Add support for time64 tests
syscalls/ppoll: Add support for time64 tests
syscalls/select6: Add support for time64 tests
syscalls/semop: Migrate to new test framework
syscalls/semtimedop: Add support for semtimedop and its time64 version
syscalls/utimensat: Migrate to new test framework
syscalls/utimensat: Add support for time64 tests
include/lapi/io_pgetevents.h | 15 +-
include/tst_timer.h | 191 +++++++
runtest/syscalls | 1 -
testcases/kernel/syscalls/futex/.gitignore | 1 -
testcases/kernel/syscalls/futex/Makefile | 1 -
.../syscalls/futex/futex_cmp_requeue01.c | 35 +-
.../syscalls/futex/futex_cmp_requeue02.c | 25 +-
testcases/kernel/syscalls/futex/futex_utils.h | 52 +-
.../kernel/syscalls/futex/futex_wait01.c | 83 +--
.../kernel/syscalls/futex/futex_wait02.c | 104 ++--
.../kernel/syscalls/futex/futex_wait03.c | 91 +--
.../kernel/syscalls/futex/futex_wait04.c | 85 +--
.../kernel/syscalls/futex/futex_wait05.c | 2 +-
.../kernel/syscalls/futex/futex_wait_bitset.h | 75 ---
.../syscalls/futex/futex_wait_bitset01.c | 103 +++-
.../syscalls/futex/futex_wait_bitset02.c | 18 -
.../kernel/syscalls/futex/futex_wake01.c | 71 ++-
.../kernel/syscalls/futex/futex_wake02.c | 99 ++--
.../kernel/syscalls/futex/futex_wake03.c | 111 ++--
.../kernel/syscalls/futex/futex_wake04.c | 150 ++---
testcases/kernel/syscalls/futex/futextest.h | 111 ++--
.../syscalls/io_pgetevents/io_pgetevents01.c | 41 +-
.../syscalls/io_pgetevents/io_pgetevents02.c | 59 +-
testcases/kernel/syscalls/ipc/semop/Makefile | 2 +-
testcases/kernel/syscalls/ipc/semop/semop.h | 52 ++
testcases/kernel/syscalls/ipc/semop/semop01.c | 145 +++--
testcases/kernel/syscalls/ipc/semop/semop02.c | 153 +++---
testcases/kernel/syscalls/ipc/semop/semop03.c | 160 ++----
testcases/kernel/syscalls/ipc/semop/semop04.c | 173 +++---
testcases/kernel/syscalls/ipc/semop/semop05.c | 309 +++++------
.../mq_timedreceive/mq_timedreceive01.c | 97 +++-
.../syscalls/mq_timedsend/mq_timedsend01.c | 101 ++--
testcases/kernel/syscalls/ppoll/ppoll01.c | 76 ++-
.../sched_rr_get_interval01.c | 131 ++---
.../sched_rr_get_interval02.c | 137 ++---
.../sched_rr_get_interval03.c | 151 +++--
testcases/kernel/syscalls/select/select_var.h | 27 +-
.../kernel/syscalls/sendmmsg/sendmmsg01.c | 45 +-
.../kernel/syscalls/sendmmsg/sendmmsg_var.h | 55 +-
.../syscalls/sigwaitinfo/sigwaitinfo01.c | 317 ++++++-----
.../syscalls/timer_gettime/timer_gettime01.c | 135 ++---
.../syscalls/timer_settime/timer_settime01.c | 51 +-
.../syscalls/timer_settime/timer_settime02.c | 65 ++-
testcases/kernel/syscalls/timerfd/timerfd01.c | 58 +-
testcases/kernel/syscalls/timerfd/timerfd04.c | 56 +-
.../syscalls/timerfd/timerfd_gettime01.c | 138 +++--
.../syscalls/timerfd/timerfd_settime01.c | 139 +++--
.../syscalls/timerfd/timerfd_settime02.c | 33 +-
testcases/kernel/syscalls/utils/mq_timed.h | 42 +-
.../kernel/syscalls/utimensat/utimensat01.c | 469 ++++++++--------
.../syscalls/utimensat/utimensat_tests.sh | 517 ------------------
51 files changed, 2657 insertions(+), 2701 deletions(-)
delete mode 100644 testcases/kernel/syscalls/futex/futex_wait_bitset.h
delete mode 100644 testcases/kernel/syscalls/futex/futex_wait_bitset02.c
create mode 100644 testcases/kernel/syscalls/ipc/semop/semop.h
delete mode 100755 testcases/kernel/syscalls/utimensat/utimensat_tests.sh
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 01/17] syscalls/timer_gettime: Add support for time64 tests
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-08 8:30 ` Arnd Bergmann
2020-05-08 4:23 ` [LTP] [PATCH V2 02/17] syscalls/timer_settime: " Viresh Kumar
` (15 subsequent siblings)
16 siblings, 1 reply; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
This adds support for time64 tests to the existing timer_gettime()
syscall tests.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
include/tst_timer.h | 40 ++++++
.../syscalls/timer_gettime/timer_gettime01.c | 135 ++++++++----------
2 files changed, 100 insertions(+), 75 deletions(-)
diff --git a/include/tst_timer.h b/include/tst_timer.h
index 256e1d71e1bc..601f934da670 100644
--- a/include/tst_timer.h
+++ b/include/tst_timer.h
@@ -15,6 +15,7 @@
#include <sys/time.h>
#include <time.h>
#include "tst_test.h"
+#include "lapi/common_timers.h"
#include "lapi/syscalls.h"
/*
@@ -112,6 +113,11 @@ struct __kernel_timespec {
__kernel_time64_t tv_sec; /* seconds */
long long tv_nsec; /* nanoseconds */
};
+
+struct __kernel_itimerspec {
+ struct __kernel_timespec it_interval; /* timer period */
+ struct __kernel_timespec it_value; /* timer expiration */
+};
#endif
enum tst_ts_type {
@@ -129,6 +135,14 @@ struct tst_ts {
} ts;
};
+struct tst_its {
+ enum tst_ts_type type;
+ union {
+ struct itimerspec libc_its;
+ struct __kernel_itimerspec kern_its;
+ } ts;
+};
+
static inline void *tst_ts_get(struct tst_ts *t)
{
if (!t)
@@ -147,6 +161,22 @@ static inline void *tst_ts_get(struct tst_ts *t)
}
}
+static inline void *tst_its_get(struct tst_its *t)
+{
+ if (!t)
+ return NULL;
+
+ switch (t->type) {
+ case TST_LIBC_TIMESPEC:
+ return &t->ts.libc_its;
+ case TST_KERN_TIMESPEC:
+ return &t->ts.kern_its;
+ default:
+ tst_brk(TBROK, "Invalid type: %d", t->type);
+ return NULL;
+ }
+}
+
static inline int libc_clock_getres(clockid_t clk_id, void *ts)
{
return clock_getres(clk_id, ts);
@@ -212,6 +242,16 @@ static inline int sys_clock_nanosleep64(clockid_t clk_id, int flags,
request, remain);
}
+static inline int sys_timer_gettime(kernel_timer_t timerid, void *its)
+{
+ return tst_syscall(__NR_timer_gettime, timerid, its);
+}
+
+static inline int sys_timer_gettime64(kernel_timer_t timerid, void *its)
+{
+ return tst_syscall(__NR_timer_gettime64, timerid, its);
+}
+
/*
* Returns tst_ts seconds.
*/
diff --git a/testcases/kernel/syscalls/timer_gettime/timer_gettime01.c b/testcases/kernel/syscalls/timer_gettime/timer_gettime01.c
index 1c75f1cf0e45..4e9230fb6e9a 100644
--- a/testcases/kernel/syscalls/timer_gettime/timer_gettime01.c
+++ b/testcases/kernel/syscalls/timer_gettime/timer_gettime01.c
@@ -1,24 +1,5 @@
-/******************************************************************************
- * Copyright (c) Crackerjack Project., 2007 *
- * Porting from Crackerjack to LTP is done by: *
- * Manas Kumar Nayak <maknayak@in.ibm.com> *
- * Copyright (c) 2013 Cyril Hrubis <chrubis@suse.cz> *
- * *
- * This program is free software; you can redistribute it and/or modify *
- * it under the terms of the GNU General Public License as published by *
- * the Free Software Foundation; either version 2 of the License, or *
- * (at your option) any later version. *
- * *
- * This program 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 General Public License for more details. *
- * *
- * You should have received a copy of the GNU General Public License *
- * along with this program; if not, write to the Free Software Foundation, *
- * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA *
- * *
- ******************************************************************************/
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Copyright (c) Crackerjack Project., 2007 */
#include <time.h>
#include <signal.h>
@@ -26,71 +7,75 @@
#include <stdio.h>
#include <errno.h>
-#include "test.h"
-#include "lapi/syscalls.h"
+#include "tst_timer.h"
+#include "lapi/abisize.h"
-char *TCID = "timer_gettime01";
-int TST_TOTAL = 3;
+static struct test_variants {
+ int (*func)(kernel_timer_t timer, void *its);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .func = sys_timer_gettime, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+#endif
-static void cleanup(void)
-{
- tst_rmdir();
-}
+#if defined(TST_ABI64)
+ { .func = sys_timer_gettime, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
-static void setup(void)
-{
- TEST_PAUSE;
- tst_tmpdir();
-}
+#if (__NR_timer_gettime64 != __LTP__NR_INVALID_SYSCALL)
+ { .func = sys_timer_gettime64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
-int main(int ac, char **av)
-{
- int lc;
+static kernel_timer_t timer;
+static void setup(void)
+{
struct sigevent ev;
- struct itimerspec spec;
- int timer;
- tst_parse_opts(ac, av, NULL, NULL);
-
- setup();
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
ev.sigev_value = (union sigval) 0;
ev.sigev_signo = SIGALRM;
ev.sigev_notify = SIGEV_SIGNAL;
- TEST(ltp_syscall(__NR_timer_create, CLOCK_REALTIME, &ev, &timer));
-
- if (TEST_RETURN != 0)
- tst_brkm(TBROK | TERRNO, cleanup, "Failed to create timer");
-
- for (lc = 0; TEST_LOOPING(lc); ++lc) {
- tst_count = 0;
-
- TEST(ltp_syscall(__NR_timer_gettime, timer, &spec));
- if (TEST_RETURN == 0) {
- tst_resm(TPASS, "timer_gettime(CLOCK_REALTIME) Passed");
- } else {
- tst_resm(TFAIL | TERRNO,
- "timer_gettime(CLOCK_REALTIME) Failed");
- }
-
- TEST(ltp_syscall(__NR_timer_gettime, -1, &spec));
- if (TEST_RETURN == -1 && TEST_ERRNO == EINVAL) {
- tst_resm(TPASS, "timer_gettime(-1) Failed: EINVAL");
- } else {
- tst_resm(TFAIL | TERRNO,
- "timer_gettime(-1) = %li", TEST_RETURN);
- }
-
- TEST(ltp_syscall(__NR_timer_gettime, timer, NULL));
- if (TEST_RETURN == -1 && TEST_ERRNO == EFAULT) {
- tst_resm(TPASS, "timer_gettime(NULL) Failed: EFAULT");
- } else {
- tst_resm(TFAIL | TERRNO,
- "timer_gettime(-1) = %li", TEST_RETURN);
- }
+
+ TEST(tst_syscall(__NR_timer_create, CLOCK_REALTIME, &ev, &timer));
+
+ if (TST_RET) {
+ tst_res(TFAIL | TTERRNO, "timer_create() failed");
+ return;
+ }
+}
+
+static void verify(void)
+{
+ struct test_variants *tv = &variants[tst_variant];
+ struct tst_its spec = {.type = tv->type, };
+
+ TEST(tv->func(timer, tst_its_get(&spec)));
+ if (TST_RET == 0) {
+ tst_res(TPASS, "timer_gettime() Passed");
+ } else {
+ tst_res(TFAIL | TTERRNO, "timer_gettime() Failed");
}
- cleanup();
- tst_exit();
+ TEST(tv->func((kernel_timer_t)-1, tst_its_get(&spec)));
+ if (TST_RET == -1 && TST_ERR == EINVAL)
+ tst_res(TPASS, "timer_gettime(-1) Failed: EINVAL");
+ else
+ tst_res(TFAIL | TTERRNO, "timer_gettime(-1) = %li", TST_RET);
+
+ TEST(tv->func(timer, NULL));
+ if (TST_RET == -1 && TST_ERR == EFAULT)
+ tst_res(TPASS, "timer_gettime(NULL) Failed: EFAULT");
+ else
+ tst_res(TFAIL | TTERRNO, "timer_gettime(-1) = %li", TST_RET);
}
+
+static struct tst_test test = {
+ .test_all = verify,
+ .test_variants = ARRAY_SIZE(variants),
+ .setup = setup,
+ .needs_tmpdir = 1,
+};
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 02/17] syscalls/timer_settime: Add support for time64 tests
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 01/17] syscalls/timer_gettime: Add support for time64 tests Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 03/17] syscalls/timerfd: " Viresh Kumar
` (14 subsequent siblings)
16 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
This adds support for time64 tests to the existing timer_settime()
syscall tests.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
include/tst_timer.h | 34 ++++++++++
.../syscalls/timer_settime/timer_settime01.c | 51 +++++++++++----
.../syscalls/timer_settime/timer_settime02.c | 65 ++++++++++++++-----
3 files changed, 122 insertions(+), 28 deletions(-)
diff --git a/include/tst_timer.h b/include/tst_timer.h
index 601f934da670..94d03cecdabe 100644
--- a/include/tst_timer.h
+++ b/include/tst_timer.h
@@ -252,6 +252,40 @@ static inline int sys_timer_gettime64(kernel_timer_t timerid, void *its)
return tst_syscall(__NR_timer_gettime64, timerid, its);
}
+static inline int sys_timer_settime(kernel_timer_t timerid, int flags,
+ void *its, void *old_its)
+{
+ return tst_syscall(__NR_timer_settime, timerid, flags, its, old_its);
+}
+
+static inline int sys_timer_settime64(kernel_timer_t timerid, int flags,
+ void *its, void *old_its)
+{
+ return tst_syscall(__NR_timer_settime64, timerid, flags, its, old_its);
+}
+
+static inline void tst_its_set_time(struct tst_its *its, long long value_sec,
+ long long value_nsec, long long interval_sec,
+ long long interval_nsec)
+{
+ switch (its->type) {
+ case TST_LIBC_TIMESPEC:
+ its->ts.libc_its.it_value.tv_sec = value_sec;
+ its->ts.libc_its.it_value.tv_nsec = value_nsec;
+ its->ts.libc_its.it_interval.tv_sec = interval_sec;
+ its->ts.libc_its.it_interval.tv_nsec = interval_nsec;
+ break;
+ case TST_KERN_TIMESPEC:
+ its->ts.kern_its.it_value.tv_sec = value_sec;
+ its->ts.kern_its.it_value.tv_nsec = value_nsec;
+ its->ts.kern_its.it_interval.tv_sec = interval_sec;
+ its->ts.kern_its.it_interval.tv_nsec = interval_nsec;
+ break;
+ default:
+ tst_brk(TBROK, "Invalid type: %d", its->type);
+ }
+}
+
/*
* Returns tst_ts seconds.
*/
diff --git a/testcases/kernel/syscalls/timer_settime/timer_settime01.c b/testcases/kernel/syscalls/timer_settime/timer_settime01.c
index da365d221086..8c00718eedaf 100644
--- a/testcases/kernel/syscalls/timer_settime/timer_settime01.c
+++ b/testcases/kernel/syscalls/timer_settime/timer_settime01.c
@@ -25,15 +25,15 @@
#include <errno.h>
#include <time.h>
#include <signal.h>
-#include "tst_test.h"
-#include "lapi/common_timers.h"
+#include "tst_timer.h"
+#include "lapi/abisize.h"
-static struct timespec timenow;
-static struct itimerspec new_set, old_set;
+static struct tst_ts timenow;
+static struct tst_its new_set, old_set;
static kernel_timer_t timer;
static struct testcase {
- struct itimerspec *old_ptr;
+ struct tst_its *old_ptr;
int it_value_tv_sec;
int it_interval_tv_sec;
int flag;
@@ -45,10 +45,32 @@ static struct testcase {
{&old_set, 5, 0, TIMER_ABSTIME, "using absolute time"},
};
+static struct test_variants {
+ int (*gettime)(clockid_t clk_id, void *ts);
+ int (*func)(kernel_timer_t timerid, int flags, void *its,
+ void *old_its);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .gettime = sys_clock_gettime, .func = sys_timer_settime, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+#endif
+
+#if defined(TST_ABI64)
+ { .gettime = sys_clock_gettime, .func = sys_timer_settime, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_timer_settime64 != __LTP__NR_INVALID_SYSCALL)
+ { .gettime = sys_clock_gettime64, .func = sys_timer_settime64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
static void run(unsigned int n)
{
- unsigned int i;
+ struct test_variants *tv = &variants[tst_variant];
struct testcase *tc = &tcases[n];
+ long long val;
+ unsigned int i;
tst_res(TINFO, "Testing for %s:", tc->description);
@@ -78,21 +100,24 @@ static void run(unsigned int n)
memset(&new_set, 0, sizeof(new_set));
memset(&old_set, 0, sizeof(old_set));
- new_set.it_value.tv_sec = tc->it_value_tv_sec;
- new_set.it_interval.tv_sec = tc->it_interval_tv_sec;
+ new_set.type = old_set.type = tv->type;
+
+ val = tc->it_value_tv_sec;
if (tc->flag & TIMER_ABSTIME) {
- if (clock_gettime(clock, &timenow) < 0) {
+ timenow.type = tv->type;
+ if (tv->gettime(clock, tst_ts_get(&timenow)) < 0) {
tst_res(TFAIL,
"clock_gettime(%s) failed - skipping the test",
get_clock_str(clock));
continue;
}
- new_set.it_value.tv_sec += timenow.tv_sec;
+ val += tst_ts_get_sec(timenow);
}
- TEST(tst_syscall(__NR_timer_settime, timer,
- tc->flag, &new_set, tc->old_ptr));
+ tst_its_set_time(&new_set, val, 0, tc->it_interval_tv_sec, 0);
+
+ TEST(tv->func(timer, tc->flag, tst_its_get(&new_set), tst_its_get(tc->old_ptr)));
if (TST_RET != 0) {
tst_res(TFAIL | TTERRNO, "%s failed",
@@ -116,6 +141,7 @@ static void sighandler(int sig)
static void setup(void)
{
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
SAFE_SIGNAL(SIGALRM, sighandler);
}
@@ -123,6 +149,7 @@ static struct tst_test test = {
.test = run,
.needs_root = 1,
.tcnt = ARRAY_SIZE(tcases),
+ .test_variants = ARRAY_SIZE(variants),
.setup = setup,
.tags = (const struct tst_tag[]) {
{"linux-git", "f18ddc13af98"},
diff --git a/testcases/kernel/syscalls/timer_settime/timer_settime02.c b/testcases/kernel/syscalls/timer_settime/timer_settime02.c
index bcabb76956f8..9ca62e15b690 100644
--- a/testcases/kernel/syscalls/timer_settime/timer_settime02.c
+++ b/testcases/kernel/syscalls/timer_settime/timer_settime02.c
@@ -25,10 +25,12 @@
#include <errno.h>
#include <time.h>
-#include "tst_test.h"
-#include "lapi/common_timers.h"
+#include "tst_timer.h"
+#include "lapi/abisize.h"
-static struct itimerspec new_set, old_set;
+static struct tst_its new_set, old_set;
+static struct tst_its *pnew_set = &new_set, *pold_set = &old_set, *null_set = NULL;
+static void *faulty_set;
static kernel_timer_t timer;
static kernel_timer_t timer_inval = -1;
@@ -44,23 +46,50 @@ static const char * const descriptions[] = {
static struct testcase {
kernel_timer_t *timer_id;
- struct itimerspec *new_ptr;
- struct itimerspec *old_ptr;
+ struct tst_its **new_ptr;
+ struct tst_its **old_ptr;
int it_value_tv_nsec;
int error;
} tcases[] = {
- {&timer, NULL, &old_set, 0, EINVAL},
- {&timer, &new_set, &old_set, -1, EINVAL},
- {&timer, &new_set, &old_set, NSEC_PER_SEC + 1, EINVAL},
- {&timer_inval, &new_set, &old_set, 0, EINVAL},
- {&timer, (struct itimerspec *) -1, &old_set, 0, EFAULT},
- {&timer, &new_set, (struct itimerspec *) -1, 0, EFAULT},
+ {&timer, &null_set, &pold_set, 0, EINVAL},
+ {&timer, &pnew_set, &pold_set, -1, EINVAL},
+ {&timer, &pnew_set, &pold_set, NSEC_PER_SEC + 1, EINVAL},
+ {&timer_inval, &pnew_set, &pold_set, 0, EINVAL},
+ {&timer, (struct tst_its **)&faulty_set, &pold_set, 0, EFAULT},
+ {&timer, &pnew_set, (struct tst_its **)&faulty_set, 0, EFAULT},
};
+static struct test_variants {
+ int (*func)(kernel_timer_t timerid, int flags, void *its,
+ void *old_its);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .func = sys_timer_settime, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+#endif
+
+#if defined(TST_ABI64)
+ { .func = sys_timer_settime, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_timer_settime64 != __LTP__NR_INVALID_SYSCALL)
+ { .func = sys_timer_settime64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
+static void setup(void)
+{
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ faulty_set = tst_get_bad_addr(NULL);
+}
+
static void run(unsigned int n)
{
- unsigned int i;
+ struct test_variants *tv = &variants[tst_variant];
struct testcase *tc = &tcases[n];
+ void *new, *old;
+ unsigned int i;
tst_res(TINFO, "Testing for %s:", descriptions[n]);
@@ -91,11 +120,13 @@ static void run(unsigned int n)
memset(&new_set, 0, sizeof(new_set));
memset(&old_set, 0, sizeof(old_set));
- new_set.it_value.tv_sec = 5;
- new_set.it_value.tv_nsec = tc->it_value_tv_nsec;
+ new_set.type = old_set.type = tv->type;
+ tst_its_set_time(&new_set, 5, tc->it_value_tv_nsec, 0, 0);
+
+ new = (tc->new_ptr == (struct tst_its **)&faulty_set) ? faulty_set : tst_its_get(*tc->new_ptr);
+ old = (tc->old_ptr == (struct tst_its **)&faulty_set) ? faulty_set : tst_its_get(*tc->old_ptr);
- TEST(tst_syscall(__NR_timer_settime, *tc->timer_id,
- 0, tc->new_ptr, tc->old_ptr));
+ TEST(tv->func(*tc->timer_id, 0, new, old));
if (tc->error != TST_ERR) {
tst_res(TFAIL | TTERRNO,
@@ -119,6 +150,8 @@ static struct tst_test test = {
.test = run,
.needs_root = 1,
.tcnt = ARRAY_SIZE(tcases),
+ .test_variants = ARRAY_SIZE(variants),
+ .setup = setup,
.tags = (const struct tst_tag[]) {
{"linux-git", "f18ddc13af98"},
{}
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 03/17] syscalls/timerfd: Add support for time64 tests
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 01/17] syscalls/timer_gettime: Add support for time64 tests Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 02/17] syscalls/timer_settime: " Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 04/17] syscalls/sched_rr_get_interval: " Viresh Kumar
` (13 subsequent siblings)
16 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
This adds support for time64 tests to the existing timerfd_gettime() and
timerfd_settime() syscall tests.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
include/tst_timer.h | 48 ++++++
testcases/kernel/syscalls/timerfd/timerfd01.c | 58 ++++++--
testcases/kernel/syscalls/timerfd/timerfd04.c | 56 ++++++-
.../syscalls/timerfd/timerfd_gettime01.c | 138 ++++++++---------
.../syscalls/timerfd/timerfd_settime01.c | 139 +++++++++---------
.../syscalls/timerfd/timerfd_settime02.c | 33 ++++-
6 files changed, 303 insertions(+), 169 deletions(-)
diff --git a/include/tst_timer.h b/include/tst_timer.h
index 94d03cecdabe..a728d78d7def 100644
--- a/include/tst_timer.h
+++ b/include/tst_timer.h
@@ -264,6 +264,54 @@ static inline int sys_timer_settime64(kernel_timer_t timerid, int flags,
return tst_syscall(__NR_timer_settime64, timerid, flags, its, old_its);
}
+static inline int sys_timerfd_gettime(int fd, void *its)
+{
+ return tst_syscall(__NR_timerfd_gettime, fd, its);
+}
+
+static inline int sys_timerfd_gettime64(int fd, void *its)
+{
+ return tst_syscall(__NR_timerfd_gettime64, fd, its);
+}
+
+static inline int sys_timerfd_settime(int fd, int flags, void *its,
+ void *old_its)
+{
+ return tst_syscall(__NR_timerfd_settime, fd, flags, its, old_its);
+}
+
+static inline int sys_timerfd_settime64(int fd, int flags, void *its,
+ void *old_its)
+{
+ return tst_syscall(__NR_timerfd_settime64, fd, flags, its, old_its);
+}
+
+static inline long long tst_its_get_val_sec(struct tst_its its)
+{
+ switch (its.type) {
+ case TST_LIBC_TIMESPEC:
+ return its.ts.libc_its.it_value.tv_sec;
+ case TST_KERN_TIMESPEC:
+ return its.ts.kern_its.it_value.tv_sec;
+ default:
+ tst_brk(TBROK, "Invalid type: %d", its.type);
+ return -1;
+ }
+}
+
+static inline long long tst_its_get_val_nsec(struct tst_its its)
+{
+ switch (its.type) {
+ case TST_LIBC_TIMESPEC:
+ return its.ts.libc_its.it_value.tv_nsec;
+ case TST_KERN_TIMESPEC:
+ return its.ts.kern_its.it_value.tv_nsec;
+ default:
+ tst_brk(TBROK, "Invalid type: %d", its.type);
+ return -1;
+ }
+}
+
static inline void tst_its_set_time(struct tst_its *its, long long value_sec,
long long value_nsec, long long interval_sec,
long long interval_nsec)
diff --git a/testcases/kernel/syscalls/timerfd/timerfd01.c b/testcases/kernel/syscalls/timerfd/timerfd01.c
index 45d4c5d5ea0d..99882bfc9782 100644
--- a/testcases/kernel/syscalls/timerfd/timerfd01.c
+++ b/testcases/kernel/syscalls/timerfd/timerfd01.c
@@ -17,8 +17,8 @@
#define _GNU_SOURCE
#include <poll.h>
-#include "tst_test.h"
#include "tst_timer.h"
+#include "lapi/abisize.h"
#include "tst_safe_timerfd.h"
static struct tcase {
@@ -29,25 +29,50 @@ static struct tcase {
{CLOCK_REALTIME, "CLOCK REALTIME"},
};
+static struct test_variants {
+ int (*cgettime)(clockid_t clk_id, void *ts);
+ int (*tfd_gettime)(int fd, void *its);
+ int (*tfd_settime)(int fd, int flags, void *new_value, void *old_value);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .cgettime = sys_clock_gettime, .tfd_gettime = sys_timerfd_gettime, .tfd_settime = sys_timerfd_settime, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+#endif
+
+#if defined(TST_ABI64)
+ { .cgettime = sys_clock_gettime, .tfd_gettime = sys_timerfd_gettime, .tfd_settime = sys_timerfd_settime, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_timerfd_gettime64 != __LTP__NR_INVALID_SYSCALL)
+ { .cgettime = sys_clock_gettime64, .tfd_gettime = sys_timerfd_gettime64, .tfd_settime = sys_timerfd_settime64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
static unsigned long long getustime(int clockid)
{
- struct timespec tp;
+ struct test_variants *tv = &variants[tst_variant];
+ struct tst_ts tp = {.type = tv->type, };
- if (clock_gettime((clockid_t) clockid, &tp)) {
+ if (tv->cgettime((clockid_t) clockid, tst_ts_get(&tp))) {
tst_res(TFAIL | TERRNO, "clock_gettime() failed");
return 0;
}
- return 1000000ULL * tp.tv_sec + tp.tv_nsec / 1000;
+ return 1000000ULL * tst_ts_get_sec(tp) + tst_ts_get_nsec(tp) / 1000;
}
-static void settime(int tfd, struct itimerspec *tmr, int tflags,
+static void settime(int tfd, struct tst_its *tmr, int tflags,
unsigned long long tvalue, int tinterval)
{
- tmr->it_value = tst_timespec_from_us(tvalue);
- tmr->it_interval = tst_timespec_from_us(tinterval);
+ struct test_variants *tv = &variants[tst_variant];
+ struct timespec value = tst_timespec_from_us(tvalue);
+ struct timespec interval = tst_timespec_from_us(tinterval);
- SAFE_TIMERFD_SETTIME(tfd, tflags, tmr, NULL);
+ tst_its_set_time(tmr, value.tv_sec, value.tv_nsec, interval.tv_sec, interval.tv_nsec);
+
+ if (tv->tfd_settime(tfd, tflags, tst_its_get(tmr), NULL))
+ tst_res(TFAIL, "timerfd_settime() failed");
}
static void waittmr(int tfd, unsigned int exp_ticks)
@@ -78,10 +103,11 @@ static void waittmr(int tfd, unsigned int exp_ticks)
static void run(unsigned int n)
{
+ struct test_variants *tv = &variants[tst_variant];
int tfd;
unsigned long long tnow;
uint64_t uticks;
- struct itimerspec tmr;
+ struct tst_its tmr = {.type = tv->type, };
struct tcase *clks = &tcases[n];
tst_res(TINFO, "testing %s", clks->name);
@@ -102,11 +128,12 @@ static void run(unsigned int n)
settime(tfd, &tmr, TFD_TIMER_ABSTIME, tnow + 50 * 1000, 50 * 1000);
memset(&tmr, 0, sizeof(tmr));
- if (timerfd_gettime(tfd, &tmr))
- tst_res(TFAIL | TERRNO, "timerfd_gettime() failed");
+ tmr.type = tv->type;
+ if (tv->tfd_gettime(tfd, tst_its_get(&tmr)))
+ tst_res(TFAIL | TERRNO, "timerfd_gettime() failed");
- if (tmr.it_value.tv_sec != 0 || tmr.it_value.tv_nsec > 50 * 1000000)
+ if (tst_its_get_val_sec(tmr) != 0 || tst_its_get_val_nsec(tmr) > 50 * 1000000)
tst_res(TFAIL, "Timer read back value not relative");
else
tst_res(TPASS, "Timer read back value is relative");
@@ -132,8 +159,15 @@ static void run(unsigned int n)
SAFE_CLOSE(tfd);
}
+static void setup(void)
+{
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+}
+
static struct tst_test test = {
.test = run,
.tcnt = ARRAY_SIZE(tcases),
+ .test_variants = ARRAY_SIZE(variants),
+ .setup = setup,
.min_kver = "2.6.25",
};
diff --git a/testcases/kernel/syscalls/timerfd/timerfd04.c b/testcases/kernel/syscalls/timerfd/timerfd04.c
index 7197fc67ed47..75fa7882539a 100644
--- a/testcases/kernel/syscalls/timerfd/timerfd04.c
+++ b/testcases/kernel/syscalls/timerfd/timerfd04.c
@@ -18,8 +18,8 @@
#include "tst_safe_clocks.h"
#include "tst_safe_timerfd.h"
#include "tst_timer.h"
+#include "lapi/abisize.h"
#include "lapi/namespaces_constants.h"
-#include "tst_test.h"
#define SLEEP_US 40000
@@ -35,26 +35,60 @@ static struct tcase {
{CLOCK_BOOTTIME, CLOCK_BOOTTIME, -10},
};
+static struct test_variants {
+ int (*cgettime)(clockid_t clk_id, void *ts);
+ int (*tfd_settime)(int fd, int flags, void *new_value, void *old_value);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .cgettime = sys_clock_gettime, .tfd_settime = sys_timerfd_settime, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+#endif
+
+#if defined(TST_ABI64)
+ { .cgettime = sys_clock_gettime, .tfd_settime = sys_timerfd_settime, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_timerfd_settime64 != __LTP__NR_INVALID_SYSCALL)
+ { .cgettime = sys_clock_gettime64, .tfd_settime = sys_timerfd_settime64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
+static void setup(void)
+{
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+}
+
static void verify_timerfd(unsigned int n)
{
- struct timespec start, end;
- struct itimerspec it = {};
+ struct test_variants *tv = &variants[tst_variant];
+ struct tst_ts start, end;
+ struct tst_its it;
struct tcase *tc = &tcases[n];
+ start.type = end.type = it.type = tv->type;
SAFE_UNSHARE(CLONE_NEWTIME);
SAFE_FILE_PRINTF("/proc/self/timens_offsets", "%d %d 0",
tc->clk_off, tc->off);
- SAFE_CLOCK_GETTIME(tc->clk_id, &start);
+ if (tv->cgettime(tc->clk_id, tst_ts_get(&start))) {
+ tst_res(TFAIL | TTERRNO, "clock_gettime(2) failed for clock %s",
+ tst_clock_name(tc->clk_id));
+ return;
+ }
- it.it_value = tst_timespec_add_us(start, 1000000 * tc->off + SLEEP_US);
+ end = tst_ts_add_us(start, 1000000 * tc->off + SLEEP_US);
+ tst_its_set_time(&it, tst_ts_get_sec(end), tst_ts_get_nsec(end), 0, 0);
if (!SAFE_FORK()) {
uint64_t exp;
int fd = SAFE_TIMERFD_CREATE(tc->clk_id, 0);
- SAFE_TIMERFD_SETTIME(fd, TFD_TIMER_ABSTIME, &it, NULL);
+ if (tv->tfd_settime(fd, TFD_TIMER_ABSTIME, tst_its_get(&it), NULL)) {
+ tst_res(TFAIL, "timerfd_settime() failed");
+ return;
+ }
SAFE_READ(1, fd, &exp, sizeof(exp));
@@ -67,9 +101,13 @@ static void verify_timerfd(unsigned int n)
SAFE_WAIT(NULL);
- SAFE_CLOCK_GETTIME(CLOCK_MONOTONIC, &end);
+ if (tv->cgettime(CLOCK_MONOTONIC, tst_ts_get(&end))) {
+ tst_res(TFAIL | TTERRNO, "clock_gettime(2) failed for clock %s",
+ tst_clock_name(CLOCK_MONOTONIC));
+ return;
+ }
- long long diff = tst_timespec_diff_us(end, start);
+ long long diff = tst_ts_diff_us(end, start);
if (diff > 5 * SLEEP_US) {
tst_res(TFAIL, "timerfd %s slept too long %lli",
@@ -90,6 +128,8 @@ static void verify_timerfd(unsigned int n)
static struct tst_test test = {
.tcnt = ARRAY_SIZE(tcases),
.test = verify_timerfd,
+ .test_variants = ARRAY_SIZE(variants),
+ .setup = setup,
.needs_root = 1,
.forks_child = 1,
.needs_kconfigs = (const char *[]) {
diff --git a/testcases/kernel/syscalls/timerfd/timerfd_gettime01.c b/testcases/kernel/syscalls/timerfd/timerfd_gettime01.c
index 5f3240bdc7d7..7cfad51d3218 100644
--- a/testcases/kernel/syscalls/timerfd/timerfd_gettime01.c
+++ b/testcases/kernel/syscalls/timerfd/timerfd_gettime01.c
@@ -1,20 +1,8 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (c) 2014 Fujitsu Ltd.
* Author: Zeng Linggang <zenglg.jy@cn.fujitsu.com>
*
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of version 2 of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
- *
- * You should have received a copy of the GNU General Public License along
- * with this program; if not, write the Free Software Foundation, Inc.,
- * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
- */
-/*
* DESCRIPTION
* Verify that,
* 1. fd is not a valid file descriptor, EBADF would return.
@@ -24,99 +12,101 @@
#define _GNU_SOURCE
-#include <errno.h>
-#include <sys/types.h>
-#include <sys/stat.h>
-#include <fcntl.h>
-
-#include "test.h"
-#include "safe_macros.h"
-#include "lapi/timerfd.h"
+#include "tst_timer.h"
+#include "lapi/abisize.h"
+#include "tst_safe_timerfd.h"
char *TCID = "timerfd_gettime01";
static int bad_clockfd = -1;
static int clockfd;
static int fd;
+static void *bad_addr;
static struct test_case_t {
int *fd;
- struct itimerspec *curr_value;
+ struct tst_its *curr_value;
int exp_errno;
} test_cases[] = {
{&bad_clockfd, NULL, EBADF},
- {&clockfd, (struct itimerspec *)-1, EFAULT},
+ {&clockfd, NULL, EFAULT},
{&fd, NULL, EINVAL},
};
-int TST_TOTAL = ARRAY_SIZE(test_cases);
-static void setup(void);
-static void timerfd_gettime_verify(const struct test_case_t *);
-static void cleanup(void);
+static struct test_variants {
+ int (*tfd_gettime)(int fd, void *its);
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .tfd_gettime = sys_timerfd_gettime, .desc = "syscall with libc spec"},
+#endif
+
+#if defined(TST_ABI64)
+ { .tfd_gettime = sys_timerfd_gettime, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_timerfd_gettime64 != __LTP__NR_INVALID_SYSCALL)
+ { .tfd_gettime = sys_timerfd_gettime64, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
-int main(int argc, char *argv[])
+static void setup(void)
{
- int lc;
- int i;
-
- tst_parse_opts(argc, argv, NULL, NULL);
-
- setup();
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ bad_addr = tst_get_bad_addr(NULL);
- for (lc = 0; TEST_LOOPING(lc); lc++) {
- tst_count = 0;
- for (i = 0; i < TST_TOTAL; i++)
- timerfd_gettime_verify(&test_cases[i]);
+ clockfd = timerfd_create(CLOCK_REALTIME, 0);
+ if (clockfd == -1) {
+ tst_brk(TFAIL | TERRNO, "timerfd_create() fail");
+ return;
}
- cleanup();
- tst_exit();
+ fd = SAFE_OPEN("test_file", O_RDWR | O_CREAT, 0644);
}
-static void setup(void)
+static void cleanup(void)
{
- if ((tst_kvercmp(2, 6, 25)) < 0)
- tst_brkm(TCONF, NULL, "This test needs kernel 2.6.25 or newer");
-
- tst_sig(NOFORK, DEF_HANDLER, cleanup);
-
- TEST_PAUSE;
-
- tst_tmpdir();
-
- clockfd = timerfd_create(CLOCK_REALTIME, 0);
- if (clockfd == -1)
- tst_brkm(TBROK | TERRNO, cleanup, "timerfd_create() fail");
+ if (clockfd > 0)
+ close(clockfd);
- fd = SAFE_OPEN(cleanup, "test_file", O_RDWR | O_CREAT, 0644);
+ if (fd > 0)
+ close(fd);
}
-static void timerfd_gettime_verify(const struct test_case_t *test)
+static void run(unsigned int n)
{
- TEST(timerfd_gettime(*test->fd, test->curr_value));
+ struct test_variants *tv = &variants[tst_variant];
+ struct test_case_t *test = &test_cases[n];
+ struct itimerspec *its;
- if (TEST_RETURN != -1) {
- tst_resm(TFAIL, "timerfd_gettime() succeeded unexpectedly");
+ if (test->exp_errno == EFAULT)
+ its = bad_addr;
+ else
+ its = tst_its_get(test->curr_value);
+
+ TEST(tv->tfd_gettime(*test->fd, its));
+
+ if (TST_RET != -1) {
+ tst_res(TFAIL, "timerfd_gettime() succeeded unexpectedly");
return;
}
- if (TEST_ERRNO == test->exp_errno) {
- tst_resm(TPASS | TTERRNO,
- "timerfd_gettime() failed as expected");
+ if (TST_ERR == test->exp_errno) {
+ tst_res(TPASS | TTERRNO,
+ "timerfd_gettime() failed as expected");
} else {
- tst_resm(TFAIL | TTERRNO,
- "timerfd_gettime() failed unexpectedly; expected: "
- "%d - %s", test->exp_errno, strerror(test->exp_errno));
+ tst_res(TFAIL | TTERRNO,
+ "timerfd_gettime() failed unexpectedly; expected: "
+ "%d - %s", test->exp_errno, strerror(test->exp_errno));
}
}
-static void cleanup(void)
-{
- if (clockfd > 0)
- close(clockfd);
-
- if (fd > 0)
- close(fd);
-
- tst_rmdir();
-}
+static struct tst_test test = {
+ .test = run,
+ .tcnt = ARRAY_SIZE(test_cases),
+ .test_variants = ARRAY_SIZE(variants),
+ .setup = setup,
+ .cleanup = cleanup,
+ .needs_tmpdir = 1,
+ .min_kver = "2.6.25",
+};
diff --git a/testcases/kernel/syscalls/timerfd/timerfd_settime01.c b/testcases/kernel/syscalls/timerfd/timerfd_settime01.c
index 2e65d23ae825..747447aff754 100644
--- a/testcases/kernel/syscalls/timerfd/timerfd_settime01.c
+++ b/testcases/kernel/syscalls/timerfd/timerfd_settime01.c
@@ -1,20 +1,8 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (c) 2014 Fujitsu Ltd.
* Author: Zeng Linggang <zenglg.jy@cn.fujitsu.com>
*
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of version 2 of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
- *
- * You should have received a copy of the GNU General Public License along
- * with this program; if not, write the Free Software Foundation, Inc.,
- * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
- */
-/*
* DESCRIPTION
* Verify that,
* 1. fd is not a valid file descriptor, EBADF would return.
@@ -25,13 +13,8 @@
#define _GNU_SOURCE
-#include <errno.h>
-#include <sys/types.h>
-#include <sys/stat.h>
-#include <fcntl.h>
-
-#include "test.h"
-#include "safe_macros.h"
+#include "tst_timer.h"
+#include "lapi/abisize.h"
#include "lapi/timerfd.h"
char *TCID = "timerfd_settime01";
@@ -39,80 +22,55 @@ char *TCID = "timerfd_settime01";
static int bad_clockfd = -1;
static int clockfd;
static int fd;
+static void *bad_addr;
static struct test_case_t {
int *fd;
int flags;
- struct itimerspec *old_value;
+ struct tst_its *old_value;
int exp_errno;
} test_cases[] = {
{&bad_clockfd, 0, NULL, EBADF},
- {&clockfd, 0, (struct itimerspec *)-1, EFAULT},
+ {&clockfd, 0, NULL, EFAULT},
{&fd, 0, NULL, EINVAL},
{&clockfd, -1, NULL, EINVAL},
};
-int TST_TOTAL = ARRAY_SIZE(test_cases);
-static void setup(void);
-static void timerfd_settime_verify(const struct test_case_t *);
-static void cleanup(void);
-static struct itimerspec new_value;
-
-int main(int argc, char *argv[])
-{
- int lc;
- int i;
-
- tst_parse_opts(argc, argv, NULL, NULL);
+static struct tst_its new_value;
- setup();
+static struct test_variants {
+ int (*tfd_settime)(int fd, int flags, void *new_value, void *old_value);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .tfd_settime = sys_timerfd_settime, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+#endif
- for (lc = 0; TEST_LOOPING(lc); lc++) {
- tst_count = 0;
- for (i = 0; i < TST_TOTAL; i++)
- timerfd_settime_verify(&test_cases[i]);
- }
+#if defined(TST_ABI64)
+ { .tfd_settime = sys_timerfd_settime, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
- cleanup();
- tst_exit();
-}
+#if (__NR_timerfd_settime64 != __LTP__NR_INVALID_SYSCALL)
+ { .tfd_settime = sys_timerfd_settime64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
static void setup(void)
{
- if ((tst_kvercmp(2, 6, 25)) < 0)
- tst_brkm(TCONF, NULL, "This test needs kernel 2.6.25 or newer");
-
- tst_sig(NOFORK, DEF_HANDLER, cleanup);
+ struct test_variants *tv = &variants[tst_variant];
- TEST_PAUSE;
-
- tst_tmpdir();
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ bad_addr = tst_get_bad_addr(NULL);
+ new_value.type = tv->type;
clockfd = timerfd_create(CLOCK_REALTIME, 0);
- if (clockfd == -1)
- tst_brkm(TBROK | TERRNO, cleanup, "timerfd_create() fail");
-
- fd = SAFE_OPEN(cleanup, "test_file", O_RDWR | O_CREAT, 0644);
-}
-
-static void timerfd_settime_verify(const struct test_case_t *test)
-{
- TEST(timerfd_settime(*test->fd, test->flags, &new_value,
- test->old_value));
-
- if (TEST_RETURN != -1) {
- tst_resm(TFAIL, "timerfd_settime() succeeded unexpectedly");
+ if (clockfd == -1) {
+ tst_brk(TFAIL | TERRNO, "timerfd_create() fail");
return;
}
- if (TEST_ERRNO == test->exp_errno) {
- tst_resm(TPASS | TTERRNO,
- "timerfd_settime() failed as expected");
- } else {
- tst_resm(TFAIL | TTERRNO,
- "timerfd_settime() failed unexpectedly; expected: "
- "%d - %s", test->exp_errno, strerror(test->exp_errno));
- }
+ fd = SAFE_OPEN("test_file", O_RDWR | O_CREAT, 0644);
}
static void cleanup(void)
@@ -122,6 +80,43 @@ static void cleanup(void)
if (fd > 0)
close(fd);
+}
+
+static void run(unsigned int n)
+{
+ struct test_variants *tv = &variants[tst_variant];
+ struct test_case_t *test = &test_cases[n];
+ struct itimerspec *its;
+
+ if (test->exp_errno == EFAULT)
+ its = bad_addr;
+ else
+ its = tst_its_get(test->old_value);
- tst_rmdir();
+ TEST(tv->tfd_settime(*test->fd, test->flags, tst_its_get(&new_value),
+ its));
+
+ if (TST_RET != -1) {
+ tst_res(TFAIL, "timerfd_settime() succeeded unexpectedly");
+ return;
+ }
+
+ if (TST_ERR == test->exp_errno) {
+ tst_res(TPASS | TTERRNO,
+ "timerfd_settime() failed as expected");
+ } else {
+ tst_res(TFAIL | TTERRNO,
+ "timerfd_settime() failed unexpectedly; expected: "
+ "%d - %s", test->exp_errno, strerror(test->exp_errno));
+ }
}
+
+static struct tst_test test = {
+ .test = run,
+ .tcnt = ARRAY_SIZE(test_cases),
+ .test_variants = ARRAY_SIZE(variants),
+ .setup = setup,
+ .cleanup = cleanup,
+ .needs_tmpdir = 1,
+ .min_kver = "2.6.25",
+};
diff --git a/testcases/kernel/syscalls/timerfd/timerfd_settime02.c b/testcases/kernel/syscalls/timerfd/timerfd_settime02.c
index 0565802f4288..b6444a6db7ba 100644
--- a/testcases/kernel/syscalls/timerfd/timerfd_settime02.c
+++ b/testcases/kernel/syscalls/timerfd/timerfd_settime02.c
@@ -15,8 +15,9 @@
* timerfd: Protect the might cancel mechanism proper
*/
#include <unistd.h>
+#include "tst_timer.h"
+#include "lapi/abisize.h"
#include "tst_safe_timerfd.h"
-#include "tst_test.h"
#include "tst_fuzzy_sync.h"
#include "tst_taint.h"
@@ -27,11 +28,34 @@
#endif
static int fd = -1;
-static struct itimerspec its;
+static struct tst_its its;
static struct tst_fzsync_pair fzsync_pair;
+static struct test_variants {
+ int (*tfd_settime)(int fd, int flags, void *new_value, void *old_value);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .tfd_settime = sys_timerfd_settime, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+#endif
+
+#if defined(TST_ABI64)
+ { .tfd_settime = sys_timerfd_settime, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_timerfd_settime64 != __LTP__NR_INVALID_SYSCALL)
+ { .tfd_settime = sys_timerfd_settime64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
static void setup(void)
{
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ its.type = tv->type;
+
tst_taint_init(TST_TAINT_W | TST_TAINT_D);
fd = SAFE_TIMERFD_CREATE(CLOCK_REALTIME, 0);
@@ -48,7 +72,9 @@ static void cleanup(void)
static int punch_clock(int flags)
{
- return timerfd_settime(fd, flags, &its, NULL);
+ return variants[tst_variant].tfd_settime(fd, flags, tst_its_get(&its),
+ NULL);
+
}
static void *thread_run(void *arg)
@@ -91,6 +117,7 @@ static void run(void)
static struct tst_test test = {
.test_all = run,
+ .test_variants = ARRAY_SIZE(variants),
.setup = setup,
.cleanup = cleanup,
.min_kver = "2.6.25",
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 04/17] syscalls/sched_rr_get_interval: Add support for time64 tests
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (2 preceding siblings ...)
2020-05-08 4:23 ` [LTP] [PATCH V2 03/17] syscalls/timerfd: " Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 05/17] syscalls/futex: Merge futex_wait_bitset tests Viresh Kumar
` (12 subsequent siblings)
16 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
This adds support for time64 tests to the existing
sched_rr_get_interval() syscall tests.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
include/tst_timer.h | 16 ++
.../sched_rr_get_interval01.c | 131 +++++++--------
.../sched_rr_get_interval02.c | 137 +++++++---------
.../sched_rr_get_interval03.c | 151 ++++++++----------
4 files changed, 198 insertions(+), 237 deletions(-)
diff --git a/include/tst_timer.h b/include/tst_timer.h
index a728d78d7def..a726298b407d 100644
--- a/include/tst_timer.h
+++ b/include/tst_timer.h
@@ -12,6 +12,7 @@
#ifndef TST_TIMER
#define TST_TIMER
+#include <sched.h>
#include <sys/time.h>
#include <time.h>
#include "tst_test.h"
@@ -242,6 +243,21 @@ static inline int sys_clock_nanosleep64(clockid_t clk_id, int flags,
request, remain);
}
+static inline int libc_sched_rr_get_interval(pid_t pid, void *ts)
+{
+ return sched_rr_get_interval(pid, ts);
+}
+
+static inline int sys_sched_rr_get_interval(pid_t pid, void *ts)
+{
+ return tst_syscall(__NR_sched_rr_get_interval, pid, ts);
+}
+
+static inline int sys_sched_rr_get_interval64(pid_t pid, void *ts)
+{
+ return tst_syscall(__NR_sched_rr_get_interval_time64, pid, ts);
+}
+
static inline int sys_timer_gettime(kernel_timer_t timerid, void *its)
{
return tst_syscall(__NR_timer_gettime, timerid, its);
diff --git a/testcases/kernel/syscalls/sched_rr_get_interval/sched_rr_get_interval01.c b/testcases/kernel/syscalls/sched_rr_get_interval/sched_rr_get_interval01.c
index b6084f673ca7..deae3314221f 100644
--- a/testcases/kernel/syscalls/sched_rr_get_interval/sched_rr_get_interval01.c
+++ b/testcases/kernel/syscalls/sched_rr_get_interval/sched_rr_get_interval01.c
@@ -1,20 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (c) Wipro Technologies Ltd, 2002. All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of version 2 of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
- *
- * You should have received a copy of the GNU General Public License along
- * with this program; if not, write the Free Software Foundation, Inc.,
- * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
- *
- */
-/**********************************************************
*
* TEST IDENTIFIER : sched_rr_get_interval01
*
@@ -62,81 +48,74 @@
* -P x : Pause for x seconds between iterations.
* -t : Turn on syscall timing.
*
- ****************************************************************/
+ */
-#include <errno.h>
#include <sched.h>
-#include "test.h"
-
-static void setup();
-static void cleanup();
-
-char *TCID = "sched_rr_get_interval01";
-int TST_TOTAL = 1;
-
-struct timespec tp;
-
-int main(int ac, char **av)
+#include "tst_timer.h"
+#include "lapi/abisize.h"
+
+struct tst_ts tp;
+
+static struct test_variants {
+ int (*func)(pid_t pid, void *ts);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .func = libc_sched_rr_get_interval, .type = TST_LIBC_TIMESPEC, .desc = "vDSO or syscall with libc spec"},
+ { .func = sys_sched_rr_get_interval, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+ { .func = sys_sched_rr_get_interval, .type = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .func = sys_sched_rr_get_interval, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_sched_rr_get_interval_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .func = sys_sched_rr_get_interval64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
+static void setup(void)
{
-
- int lc;
-
- tst_parse_opts(ac, av, NULL, NULL);
-
- setup();
-
- for (lc = 0; TEST_LOOPING(lc); lc++) {
-
- tst_count = 0;
-
- /*
- * Call sched_rr_get_interval(2) with pid=0 so that it will
- * write into the timespec structure pointed to by tp, the
- * round robin time quantum for the current process.
- */
- TEST(sched_rr_get_interval(0, &tp));
-
- if (TEST_RETURN == 0) {
- tst_resm(TPASS, "sched_rr_get_interval() returned %ld",
- TEST_RETURN);
- } else {
- tst_resm(TFAIL | TTERRNO,
- "Test Failed, sched_rr_get_interval()"
- "returned %ld", TEST_RETURN);
- }
- }
-
- /* cleanup and exit */
- cleanup();
- tst_exit();
-
-}
-
-/* setup() - performs all ONE TIME setup for this test */
-void setup(void)
-{
- tst_require_root();
+ struct test_variants *tv = &variants[tst_variant];
/*
* Initialize scheduling parameter structure to use with
* sched_setscheduler()
*/
struct sched_param p = { 1 };
- tst_sig(NOFORK, DEF_HANDLER, cleanup);
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
- TEST_PAUSE;
+ tp.type = tv->type;
/* Change scheduling policy to SCHED_RR */
- if ((sched_setscheduler(0, SCHED_RR, &p)) == -1) {
- tst_brkm(TBROK, cleanup, "sched_setscheduler() failed");
- }
+ if ((sched_setscheduler(0, SCHED_RR, &p)) == -1)
+ tst_res(TFAIL | TTERRNO, "sched_setscheduler() failed");
}
-/*
- *cleanup() - performs all ONE TIME cleanup for this test at
- * completion or premature exit.
- */
-void cleanup(void)
+static void run(void)
{
+ struct test_variants *tv = &variants[tst_variant];
+ /*
+ * Call sched_rr_get_interval(2) with pid=0 so that it will
+ * write into the timespec structure pointed to by tp, the
+ * round robin time quantum for the current process.
+ */
+ TEST(tv->func(0, tst_ts_get(&tp)));
+
+ if (!TST_RET) {
+ tst_res(TPASS, "sched_rr_get_interval() passed");
+ } else {
+ tst_res(TFAIL | TTERRNO, "Test Failed, sched_rr_get_interval() returned %ld",
+ TST_RET);
+ }
}
+
+static struct tst_test test = {
+ .test_all = run,
+ .test_variants = ARRAY_SIZE(variants),
+ .setup = setup,
+ .needs_root = 1,
+};
diff --git a/testcases/kernel/syscalls/sched_rr_get_interval/sched_rr_get_interval02.c b/testcases/kernel/syscalls/sched_rr_get_interval/sched_rr_get_interval02.c
index 6bf743f78c5e..019f2c1bdfb3 100644
--- a/testcases/kernel/syscalls/sched_rr_get_interval/sched_rr_get_interval02.c
+++ b/testcases/kernel/syscalls/sched_rr_get_interval/sched_rr_get_interval02.c
@@ -1,20 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (c) Wipro Technologies Ltd, 2002. All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of version 2 of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
- *
- * You should have received a copy of the GNU General Public License along
- * with this program; if not, write the Free Software Foundation, Inc.,
- * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
- *
- */
-/**********************************************************
*
* TEST IDENTIFIER : sched_rr_get_interval02
*
@@ -63,84 +49,77 @@
* -P x : Pause for x seconds between iterations.
* -t : Turn on syscall timing.
*
- ****************************************************************/
+ */
-#include <errno.h>
#include <sched.h>
-#include "test.h"
-
-static void setup();
-static void cleanup();
-
-char *TCID = "sched_rr_get_interval02";
-int TST_TOTAL = 1;
-
-struct timespec tp;
-
-int main(int ac, char **av)
-{
-
- int lc;
-
- tst_parse_opts(ac, av, NULL, NULL);
-
- setup();
-
- for (lc = 0; TEST_LOOPING(lc); lc++) {
-
- tst_count = 0;
-
- tp.tv_sec = 99;
- tp.tv_nsec = 99;
- /*
- * Call sched_rr_get_interval(2) with pid=0 sothat it will
- * write into the timespec structure pointed to by tp the
- * round robin time quantum for the current process.
- */
- TEST(sched_rr_get_interval(0, &tp));
-
- if ((TEST_RETURN == 0) && (tp.tv_sec == 0) && (tp.tv_nsec == 0)) {
- tst_resm(TPASS, "Test passed");
- } else {
- tst_resm(TFAIL, "Test Failed, sched_rr_get_interval()"
- "returned %ld, errno = %d : %s, tp.tv_sec = %d,"
- " tp.tv_nsec = %ld", TEST_RETURN, TEST_ERRNO,
- strerror(TEST_ERRNO), (int)tp.tv_sec,
- tp.tv_nsec);
- }
- }
-
- /* cleanup and exit */
- cleanup();
- tst_exit();
-
-}
-
-/* setup() - performs all ONE TIME setup for this test */
-void setup(void)
+#include "tst_timer.h"
+#include "lapi/abisize.h"
+
+struct tst_ts tp;
+
+static struct test_variants {
+ int (*func)(pid_t pid, void *ts);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .func = libc_sched_rr_get_interval, .type = TST_LIBC_TIMESPEC, .desc = "vDSO or syscall with libc spec"},
+ { .func = sys_sched_rr_get_interval, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+ { .func = sys_sched_rr_get_interval, .type = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .func = sys_sched_rr_get_interval, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_sched_rr_get_interval_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .func = sys_sched_rr_get_interval64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
+static void setup(void)
{
- tst_require_root();
+ struct test_variants *tv = &variants[tst_variant];
/*
* Initialize scheduling parameter structure to use with
* sched_setscheduler()
*/
struct sched_param p = { 1 };
- tst_sig(NOFORK, DEF_HANDLER, cleanup);
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
- TEST_PAUSE;
+ tp.type = tv->type;
/* Change scheduling policy to SCHED_FIFO */
- if ((sched_setscheduler(0, SCHED_FIFO, &p)) == -1) {
- tst_brkm(TBROK, cleanup, "sched_setscheduler() failed");
- }
+ if ((sched_setscheduler(0, SCHED_FIFO, &p)) == -1)
+ tst_res(TFAIL | TTERRNO, "sched_setscheduler() failed");
}
-/*
- *cleanup() - performs all ONE TIME cleanup for this test at
- * completion or premature exit.
- */
-void cleanup(void)
+static void run(void)
{
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_ts_set_sec(&tp, 99);
+ tst_ts_set_nsec(&tp, 99);
+
+ /*
+ * Call sched_rr_get_interval(2) with pid=0 so that it will
+ * write into the timespec structure pointed to by tp the
+ * round robin time quantum for the current process.
+ */
+ TEST(tv->func(0, tst_ts_get(&tp)));
+ if (!TST_RET && tst_ts_valid(&tp) == -1) {
+ tst_res(TPASS, "sched_rr_get_interval() passed");
+ } else {
+ tst_res(TFAIL | TTERRNO, "Test Failed, sched_rr_get_interval() returned %ld, tp.tv_sec = %lld, tp.tv_nsec = %lld",
+ TST_RET, tst_ts_get_sec(tp), tst_ts_get_nsec(tp));
+ }
}
+
+static struct tst_test test = {
+ .test_all = run,
+ .test_variants = ARRAY_SIZE(variants),
+ .setup = setup,
+ .needs_root = 1,
+};
diff --git a/testcases/kernel/syscalls/sched_rr_get_interval/sched_rr_get_interval03.c b/testcases/kernel/syscalls/sched_rr_get_interval/sched_rr_get_interval03.c
index 56f2fcf28269..1f66ff6b11e9 100644
--- a/testcases/kernel/syscalls/sched_rr_get_interval/sched_rr_get_interval03.c
+++ b/testcases/kernel/syscalls/sched_rr_get_interval/sched_rr_get_interval03.c
@@ -1,20 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (c) Wipro Technologies Ltd, 2002. All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of version 2 of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
- *
- * You should have received a copy of the GNU General Public License along
- * with this program; if not, write the Free Software Foundation, Inc.,
- * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
- *
- */
-/**********************************************************
*
* TEST IDENTIFIER : sched_rr_get_interval03
*
@@ -68,104 +54,105 @@
* -P x : Pause for x seconds between iterations.
* -t : Turn on syscall timing.
*
- ****************************************************************/
+ */
-#include <errno.h>
#include <sched.h>
-#include "test.h"
-
-static void setup();
-static void cleanup();
-
-char *TCID = "sched_rr_get_interval03";
-struct timespec tp;
+#include "tst_timer.h"
+#include "lapi/abisize.h"
static pid_t unused_pid;
static pid_t inval_pid = -1;
static pid_t zero_pid;
+struct tst_ts tp;
+static void *bad_addr;
+
struct test_cases_t {
pid_t *pid;
- struct timespec *tp;
+ struct tst_ts *tp;
int exp_errno;
} test_cases[] = {
- {
- &inval_pid, &tp, EINVAL}, {
- &unused_pid, &tp, ESRCH},
+ { &inval_pid, &tp, EINVAL},
+ { &unused_pid, &tp, ESRCH},
#ifndef UCLINUX
- /* Skip since uClinux does not implement memory protection */
- {
- &zero_pid, (struct timespec *)-1, EFAULT}
+ /* Skip since uClinux does not implement memory protection */
+ { &zero_pid, NULL, EFAULT}
#endif
};
-int TST_TOTAL = sizeof(test_cases) / sizeof(test_cases[0]);
-
-int main(int ac, char **av)
-{
-
- int lc, i;
-
- tst_parse_opts(ac, av, NULL, NULL);
-
- setup();
-
- for (lc = 0; TEST_LOOPING(lc); lc++) {
-
- tst_count = 0;
-
- for (i = 0; i < TST_TOTAL; ++i) {
- /*
- * Call sched_rr_get_interval(2)
- */
- TEST(sched_rr_get_interval(*(test_cases[i].pid),
- test_cases[i].tp));
-
- if ((TEST_RETURN == -1) &&
- (TEST_ERRNO == test_cases[i].exp_errno)) {
- tst_resm(TPASS, "Test Passed");
- } else {
- tst_resm(TFAIL | TTERRNO, "Test Failed,"
- " sched_rr_get_interval() returned %ld",
- TEST_RETURN);
- }
- }
- }
-
- /* cleanup and exit */
- cleanup();
+static struct test_variants {
+ int (*func)(pid_t pid, void *ts);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .func = libc_sched_rr_get_interval, .type = TST_LIBC_TIMESPEC, .desc = "vDSO or syscall with libc spec"},
+ { .func = sys_sched_rr_get_interval, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+ { .func = sys_sched_rr_get_interval, .type = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
- tst_exit();
+#if defined(TST_ABI64)
+ { .func = sys_sched_rr_get_interval, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
-}
+#if (__NR_sched_rr_get_interval_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .func = sys_sched_rr_get_interval64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
-/* setup() - performs all ONE TIME setup for this test */
-void setup(void)
+static void setup(void)
{
- tst_require_root();
+ struct test_variants *tv = &variants[tst_variant];
/*
* Initialize scheduling parameter structure to use with
* sched_setscheduler()
*/
struct sched_param p = { 1 };
- tst_sig(NOFORK, DEF_HANDLER, cleanup);
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
- TEST_PAUSE;
+ bad_addr = tst_get_bad_addr(NULL);
+ tp.type = tv->type;
/* Change scheduling policy to SCHED_RR */
- if ((sched_setscheduler(0, SCHED_RR, &p)) == -1) {
- tst_brkm(TBROK, cleanup, "sched_setscheduler() failed");
- }
+ if ((sched_setscheduler(0, SCHED_RR, &p)) == -1)
+ tst_res(TFAIL | TTERRNO, "sched_setscheduler() failed");
- unused_pid = tst_get_unused_pid(cleanup);
+ unused_pid = tst_get_unused_pid();
}
-/*
- *cleanup() - performs all ONE TIME cleanup for this test at
- * completion or premature exit.
- */
-void cleanup(void)
+static void run(unsigned int i)
{
+ struct test_variants *tv = &variants[tst_variant];
+ struct test_cases_t *tc = &test_cases[i];
+ struct timerspec *ts;
+
+ if (tc->exp_errno == EFAULT)
+ ts = bad_addr;
+ else
+ ts = tst_ts_get(tc->tp);
+
+ /*
+ * Call sched_rr_get_interval(2)
+ */
+ TEST(tv->func(*tc->pid, ts));
+ if (TST_RET != -1) {
+ tst_res(TFAIL, "sched_rr_get_interval() passed unexcpectedly");
+ return;
+ }
+
+ if (tc->exp_errno == TST_ERR)
+ tst_res(TPASS | TTERRNO, "sched_rr_get_interval() failed as excpected");
+ else
+ tst_res(TFAIL | TTERRNO, "sched_rr_get_interval() failed unexcpectedly: %s",
+ tst_strerrno(tc->exp_errno));
}
+
+static struct tst_test test = {
+ .test = run,
+ .tcnt = ARRAY_SIZE(test_cases),
+ .test_variants = ARRAY_SIZE(variants),
+ .setup = setup,
+ .needs_root = 1,
+};
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 05/17] syscalls/futex: Merge futex_wait_bitset tests
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (3 preceding siblings ...)
2020-05-08 4:23 ` [LTP] [PATCH V2 04/17] syscalls/sched_rr_get_interval: " Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 06/17] syscalls/futex: Add support for time64 tests Viresh Kumar
` (11 subsequent siblings)
16 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
We can handle them easily in a single test, lets do it.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Petr Vorel <pvorel@suse.cz>
---
runtest/syscalls | 1 -
testcases/kernel/syscalls/futex/.gitignore | 1 -
testcases/kernel/syscalls/futex/Makefile | 1 -
.../kernel/syscalls/futex/futex_wait_bitset.h | 75 -------------------
.../syscalls/futex/futex_wait_bitset01.c | 73 +++++++++++++++++-
.../syscalls/futex/futex_wait_bitset02.c | 18 -----
6 files changed, 71 insertions(+), 98 deletions(-)
delete mode 100644 testcases/kernel/syscalls/futex/futex_wait_bitset.h
delete mode 100644 testcases/kernel/syscalls/futex/futex_wait_bitset02.c
diff --git a/runtest/syscalls b/runtest/syscalls
index edd3e8de7861..4a19075cd131 100644
--- a/runtest/syscalls
+++ b/runtest/syscalls
@@ -1664,7 +1664,6 @@ futex_wake02 futex_wake02
futex_wake03 futex_wake03
futex_wake04 futex_wake04
futex_wait_bitset01 futex_wait_bitset01
-futex_wait_bitset02 futex_wait_bitset02
memfd_create01 memfd_create01
memfd_create02 memfd_create02
diff --git a/testcases/kernel/syscalls/futex/.gitignore b/testcases/kernel/syscalls/futex/.gitignore
index 68bc2028e1b2..54cd02b0223f 100644
--- a/testcases/kernel/syscalls/futex/.gitignore
+++ b/testcases/kernel/syscalls/futex/.gitignore
@@ -6,7 +6,6 @@
/futex_wait04
/futex_wait05
/futex_wait_bitset01
-/futex_wait_bitset02
/futex_wake01
/futex_wake02
/futex_wake03
diff --git a/testcases/kernel/syscalls/futex/Makefile b/testcases/kernel/syscalls/futex/Makefile
index 8dd239301e63..c88af7c9643b 100644
--- a/testcases/kernel/syscalls/futex/Makefile
+++ b/testcases/kernel/syscalls/futex/Makefile
@@ -12,7 +12,6 @@ futex_wake02: CFLAGS+=-pthread
futex_wake04: CFLAGS+=-pthread
futex_wait05: LDLIBS+=-lrt
futex_wait_bitset01: LDLIBS+=-lrt
-futex_wait_bitset02: LDLIBS+=-lrt
include $(top_srcdir)/include/mk/testcases.mk
include $(top_srcdir)/include/mk/generic_leaf_target.mk
diff --git a/testcases/kernel/syscalls/futex/futex_wait_bitset.h b/testcases/kernel/syscalls/futex/futex_wait_bitset.h
deleted file mode 100644
index 84712816bfe8..000000000000
--- a/testcases/kernel/syscalls/futex/futex_wait_bitset.h
+++ /dev/null
@@ -1,75 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * Copyright (C) 2015 Cyril Hrubis <chrubis@suse.cz>
- */
-
-#define THRESHOLD_US 100000
-#define DEFAULT_TIMEOUT_US 100010
-
-static void verify_futex_wait_bitset(long long wait_us, clock_t clk_id)
-{
- struct timespec start, to, end;
- futex_t futex = FUTEX_INITIALIZER;
- u_int32_t bitset = 0xffffffff;
- int flags = clk_id == CLOCK_REALTIME ? FUTEX_CLOCK_REALTIME : 0;
-
- tst_res(TINFO, "testing futex_wait_bitset() timeout with %s",
- clk_id == CLOCK_REALTIME ? "CLOCK_REALTIME" : "CLOCK_MONOTONIC");
-
- clock_gettime(clk_id, &start);
- to = tst_timespec_add_us(start, wait_us);
-
- TEST(futex_wait_bitset(&futex, futex, &to, bitset, flags));
-
- clock_gettime(clk_id, &end);
-
- if (TST_RET != -1) {
- tst_res(TFAIL, "futex_wait_bitset() returned %li, expected -1",
- TST_RET);
- return;
- }
-
- if (TST_ERR == ENOSYS) {
- tst_res(TCONF,
- "In this kernel, futex() does not support FUTEX_WAIT_BITSET operation");
- return;
- }
-
- if (TST_ERR != ETIMEDOUT) {
- tst_res(TFAIL | TTERRNO, "expected %s",
- tst_strerrno(ETIMEDOUT));
- return;
- }
-
- if (tst_timespec_lt(end, to)) {
- tst_res(TFAIL,
- "futex_wait_bitset() woken up prematurely %llius, expected %llius",
- tst_timespec_diff_us(end, start), wait_us);
- return;
- }
-
- if (tst_timespec_diff_us(end, to) > THRESHOLD_US) {
- tst_res(TFAIL,
- "futex_wait_bitset() waited too long %llius, expected %llius",
- tst_timespec_diff_us(end, start), wait_us);
- return;
- }
-
- tst_res(TPASS, "futex_wait_bitset() waited %llius, expected %llius",
- tst_timespec_diff_us(end, start), wait_us);
-}
-
-static void setup(void)
-{
- tst_timer_check(USE_CLOCK);
-}
-
-static void run(void)
-{
- verify_futex_wait_bitset(DEFAULT_TIMEOUT_US, USE_CLOCK);
-}
-
-static struct tst_test test = {
- .setup = setup,
- .test_all = run,
-};
diff --git a/testcases/kernel/syscalls/futex/futex_wait_bitset01.c b/testcases/kernel/syscalls/futex/futex_wait_bitset01.c
index 1ae3253cfe13..b8546c3e640c 100644
--- a/testcases/kernel/syscalls/futex/futex_wait_bitset01.c
+++ b/testcases/kernel/syscalls/futex/futex_wait_bitset01.c
@@ -13,7 +13,76 @@
#include "tst_timer.h"
#include "futextest.h"
-#define USE_CLOCK CLOCK_MONOTONIC
+#define THRESHOLD_US 100000
+#define DEFAULT_TIMEOUT_US 100010
-#include "futex_wait_bitset.h"
+static struct test_case_t {
+ clockid_t clk_id;
+} tcases[] = {
+ { CLOCK_MONOTONIC },
+ { CLOCK_REALTIME }
+};
+static void verify_futex_wait_bitset(long long wait_us, clock_t clk_id)
+{
+ struct timespec start, to, end;
+ futex_t futex = FUTEX_INITIALIZER;
+ u_int32_t bitset = 0xffffffff;
+ int flags = clk_id == CLOCK_REALTIME ? FUTEX_CLOCK_REALTIME : 0;
+
+ tst_res(TINFO, "testing futex_wait_bitset() timeout with %s",
+ clk_id == CLOCK_REALTIME ? "CLOCK_REALTIME" : "CLOCK_MONOTONIC");
+
+ clock_gettime(clk_id, &start);
+ to = tst_timespec_add_us(start, wait_us);
+
+ TEST(futex_wait_bitset(&futex, futex, &to, bitset, flags));
+
+ clock_gettime(clk_id, &end);
+
+ if (TST_RET != -1) {
+ tst_res(TFAIL, "futex_wait_bitset() returned %li, expected -1",
+ TST_RET);
+ return;
+ }
+
+ if (TST_ERR == ENOSYS) {
+ tst_res(TCONF,
+ "In this kernel, futex() does not support FUTEX_WAIT_BITSET operation");
+ return;
+ }
+
+ if (TST_ERR != ETIMEDOUT) {
+ tst_res(TFAIL | TTERRNO, "expected %s",
+ tst_strerrno(ETIMEDOUT));
+ return;
+ }
+
+ if (tst_timespec_lt(end, to)) {
+ tst_res(TFAIL,
+ "futex_wait_bitset() woken up prematurely %llius, expected %llius",
+ tst_timespec_diff_us(end, start), wait_us);
+ return;
+ }
+
+ if (tst_timespec_diff_us(end, to) > THRESHOLD_US) {
+ tst_res(TFAIL,
+ "futex_wait_bitset() waited too long %llius, expected %llius",
+ tst_timespec_diff_us(end, start), wait_us);
+ return;
+ }
+
+ tst_res(TPASS, "futex_wait_bitset() waited %llius, expected %llius",
+ tst_timespec_diff_us(end, start), wait_us);
+}
+
+static void run(unsigned int n)
+{
+ tst_timer_check(tcases[n].clk_id);
+ verify_futex_wait_bitset(DEFAULT_TIMEOUT_US, tcases[n].clk_id);
+}
+
+static struct tst_test test = {
+ .test = run,
+ .tcnt = ARRAY_SIZE(tcases),
+};
diff --git a/testcases/kernel/syscalls/futex/futex_wait_bitset02.c b/testcases/kernel/syscalls/futex/futex_wait_bitset02.c
deleted file mode 100644
index 4d532fc07775..000000000000
--- a/testcases/kernel/syscalls/futex/futex_wait_bitset02.c
+++ /dev/null
@@ -1,18 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-or-later
-/*
- * Copyright (C) 2015 Cyril Hrubis <chrubis@suse.cz>
- *
- * 1. Block on a bitset futex and wait for timeout, the difference between
- * normal futex and bitset futex is that that the later have absolute timeout.
- * 2. Check that the futex waited for expected time.
- */
-
-#include <errno.h>
-
-#include "tst_test.h"
-#include "tst_timer.h"
-#include "futextest.h"
-
-#define USE_CLOCK CLOCK_REALTIME
-
-#include "futex_wait_bitset.h"
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 06/17] syscalls/futex: Add support for time64 tests
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (4 preceding siblings ...)
2020-05-08 4:23 ` [LTP] [PATCH V2 05/17] syscalls/futex: Merge futex_wait_bitset tests Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-09 9:02 ` Li Wang
2020-05-08 4:23 ` [LTP] [PATCH V2 07/17] syscalls/io_pgetevents: " Viresh Kumar
` (10 subsequent siblings)
16 siblings, 1 reply; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
This adds support for time64 tests to the existing futex() syscall
tests.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Petr Vorel <pvorel@suse.cz>
---
include/tst_timer.h | 12 ++
.../syscalls/futex/futex_cmp_requeue01.c | 35 +++-
.../syscalls/futex/futex_cmp_requeue02.c | 25 ++-
testcases/kernel/syscalls/futex/futex_utils.h | 52 ++++--
.../kernel/syscalls/futex/futex_wait01.c | 83 +++++-----
.../kernel/syscalls/futex/futex_wait02.c | 104 ++++++------
.../kernel/syscalls/futex/futex_wait03.c | 91 +++++------
.../kernel/syscalls/futex/futex_wait04.c | 85 +++++-----
.../kernel/syscalls/futex/futex_wait05.c | 2 +-
.../syscalls/futex/futex_wait_bitset01.c | 50 ++++--
.../kernel/syscalls/futex/futex_wake01.c | 71 ++++-----
.../kernel/syscalls/futex/futex_wake02.c | 99 ++++++------
.../kernel/syscalls/futex/futex_wake03.c | 111 +++++++------
.../kernel/syscalls/futex/futex_wake04.c | 150 +++++++++---------
testcases/kernel/syscalls/futex/futextest.h | 111 ++++++++-----
15 files changed, 609 insertions(+), 472 deletions(-)
diff --git a/include/tst_timer.h b/include/tst_timer.h
index a726298b407d..b4280042c77f 100644
--- a/include/tst_timer.h
+++ b/include/tst_timer.h
@@ -243,6 +243,18 @@ static inline int sys_clock_nanosleep64(clockid_t clk_id, int flags,
request, remain);
}
+static inline int sys_futex(int *uaddr, int futex_op, int val, void *to,
+ int *uaddr2, int val3)
+{
+ return tst_syscall(__NR_futex, uaddr, futex_op, val, to, uaddr2, val3);
+}
+
+static inline int sys_futex_time64(int *uaddr, int futex_op, int val, void *to,
+ int *uaddr2, int val3)
+{
+ return tst_syscall(__NR_futex_time64, uaddr, futex_op, val, to, uaddr2, val3);
+}
+
static inline int libc_sched_rr_get_interval(pid_t pid, void *ts)
{
return sched_rr_get_interval(pid, ts);
diff --git a/testcases/kernel/syscalls/futex/futex_cmp_requeue01.c b/testcases/kernel/syscalls/futex/futex_cmp_requeue01.c
index c50fade9f969..bc9d2a6bee37 100644
--- a/testcases/kernel/syscalls/futex/futex_cmp_requeue01.c
+++ b/testcases/kernel/syscalls/futex/futex_cmp_requeue01.c
@@ -18,6 +18,7 @@
#include "tst_timer_test.h"
#include "tst_test.h"
#include "futextest.h"
+#include "lapi/abisize.h"
struct shared_data {
futex_t futexes[2];
@@ -42,14 +43,33 @@ static struct tcase {
{1000, 300, 500},
};
+static struct test_variants {
+ enum futex_fn_type fntype;
+ enum tst_ts_type tstype;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .fntype = FUTEX_FN_FUTEX, .tstype = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .fntype = FUTEX_FN_FUTEX, .tstype = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_futex_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .fntype = FUTEX_FN_FUTEX64, .tstype = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
static void do_child(void)
{
+ struct test_variants *tv = &variants[tst_variant];
+ struct tst_ts usec = tst_ts_from_ms(tv->tstype, max_sleep_ms);
int slept_for_ms = 0;
- struct timespec usec = tst_timespec_from_ms(max_sleep_ms);
int pid = getpid();
int ret = 0;
- if (futex_wait(&sd->futexes[0], sd->futexes[0], &usec, 0) == -1) {
+ if (futex_wait(tv->fntype, &sd->futexes[0], sd->futexes[0], &usec, 0) == -1) {
if (errno == EAGAIN) {
/* spurious wakeup or signal */
tst_atomic_inc(&sd->spurious);
@@ -72,6 +92,7 @@ static void do_child(void)
static void verify_futex_cmp_requeue(unsigned int n)
{
+ struct test_variants *tv = &variants[tst_variant];
int num_requeues = 0, num_waits = 0, num_total = 0;
int i, status, spurious, woken_up;
struct tcase *tc = &tcases[n];
@@ -104,8 +125,8 @@ static void verify_futex_cmp_requeue(unsigned int n)
* specifies an upper limit on the number of waiters that are requeued.
* Returns the total number of waiters that were woken up or requeued.
*/
- TEST(futex_cmp_requeue(&sd->futexes[0], sd->futexes[0], &sd->futexes[1],
- tc->set_wakes, tc->set_requeues, 0));
+ TEST(futex_cmp_requeue(tv->fntype, &sd->futexes[0], sd->futexes[0],
+ &sd->futexes[1], tc->set_wakes, tc->set_requeues, 0));
/* Fail if more than requested wakes + requeues were returned */
if (TST_RET > exp_ret) {
@@ -115,8 +136,8 @@ static void verify_futex_cmp_requeue(unsigned int n)
tst_res(TINFO, "futex_cmp_requeue() returned %ld", TST_RET);
}
- num_requeues = futex_wake(&sd->futexes[1], tc->num_waiters, 0);
- num_waits = futex_wake(&sd->futexes[0], tc->num_waiters, 0);
+ num_requeues = futex_wake(tv->fntype, &sd->futexes[1], tc->num_waiters, 0);
+ num_waits = futex_wake(tv->fntype, &sd->futexes[0], tc->num_waiters, 0);
tst_atomic_store(1, &sd->test_done);
for (i = 0; i < tc->num_waiters; i++) {
@@ -178,6 +199,7 @@ static void verify_futex_cmp_requeue(unsigned int n)
static void setup(void)
{
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
max_sleep_ms = tst_multiply_timeout(5000);
sd = SAFE_MMAP(NULL, sizeof(*sd), PROT_READ | PROT_WRITE,
@@ -198,5 +220,6 @@ static struct tst_test test = {
.cleanup = cleanup,
.tcnt = ARRAY_SIZE(tcases),
.test = verify_futex_cmp_requeue,
+ .test_variants = ARRAY_SIZE(variants),
.forks_child = 1,
};
diff --git a/testcases/kernel/syscalls/futex/futex_cmp_requeue02.c b/testcases/kernel/syscalls/futex/futex_cmp_requeue02.c
index 228d8cff6ea0..5559b72a3a45 100644
--- a/testcases/kernel/syscalls/futex/futex_cmp_requeue02.c
+++ b/testcases/kernel/syscalls/futex/futex_cmp_requeue02.c
@@ -18,6 +18,7 @@
#include "tst_test.h"
#include "futextest.h"
+#include "lapi/abisize.h"
static futex_t *futexes;
@@ -32,12 +33,30 @@ static struct tcase {
{1, 1, FUTEX_INITIALIZER + 1, EAGAIN},
};
+static struct test_variants {
+ enum futex_fn_type fntype;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .fntype = FUTEX_FN_FUTEX, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .fntype = FUTEX_FN_FUTEX, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_futex_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .fntype = FUTEX_FN_FUTEX64, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
static void verify_futex_cmp_requeue(unsigned int n)
{
+ struct test_variants *tv = &variants[tst_variant];
struct tcase *tc = &tcases[n];
- TEST(futex_cmp_requeue(&futexes[0], tc->exp_val, &futexes[1],
- tc->set_wakes, tc->set_requeues, 0));
+ TEST(futex_cmp_requeue(tv->fntype, &futexes[0], tc->exp_val,
+ &futexes[1], tc->set_wakes, tc->set_requeues, 0));
if (TST_RET != -1) {
tst_res(TFAIL, "futex_cmp_requeue() succeeded unexpectedly");
return;
@@ -55,6 +74,7 @@ static void verify_futex_cmp_requeue(unsigned int n)
static void setup(void)
{
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
futexes = SAFE_MMAP(NULL, sizeof(futex_t) * 2, PROT_READ | PROT_WRITE,
MAP_ANONYMOUS | MAP_SHARED, -1, 0);
@@ -73,6 +93,7 @@ static struct tst_test test = {
.cleanup = cleanup,
.test = verify_futex_cmp_requeue,
.tcnt = ARRAY_SIZE(tcases),
+ .test_variants = ARRAY_SIZE(variants),
.tags = (const struct tst_tag[]) {
{"CVE", "2018-6927"},
{"linux-git", "fbe0e839d1e2"},
diff --git a/testcases/kernel/syscalls/futex/futex_utils.h b/testcases/kernel/syscalls/futex/futex_utils.h
index 63b97c00b631..f77b9d6b3b8a 100644
--- a/testcases/kernel/syscalls/futex/futex_utils.h
+++ b/testcases/kernel/syscalls/futex/futex_utils.h
@@ -20,10 +20,13 @@
#ifndef FUTEX_UTILS_H__
#define FUTEX_UTILS_H__
+#include <stdio.h>
+#include <stdlib.h>
+
/*
* Wait for nr_threads to be sleeping
*/
-static int wait_for_threads(unsigned int nr_threads)
+static inline int wait_for_threads(unsigned int nr_threads)
{
char thread_state, name[1024];
DIR *dir;
@@ -32,34 +35,63 @@ static int wait_for_threads(unsigned int nr_threads)
snprintf(name, sizeof(name), "/proc/%i/task/", getpid());
- dir = SAFE_OPENDIR(NULL, name);
+ dir = SAFE_OPENDIR(name);
- while ((dent = SAFE_READDIR(NULL, dir))) {
+ while ((dent = SAFE_READDIR(dir))) {
/* skip ".", ".." and the main thread */
if (atoi(dent->d_name) == getpid() || atoi(dent->d_name) == 0)
continue;
snprintf(name, sizeof(name), "/proc/%i/task/%s/stat",
- getpid(), dent->d_name);
+ getpid(), dent->d_name);
- SAFE_FILE_SCANF(NULL, name, "%*i %*s %c", &thread_state);
+ SAFE_FILE_SCANF(name, "%*i %*s %c", &thread_state);
if (thread_state != 'S') {
- tst_resm(TINFO, "Thread %s not sleeping yet", dent->d_name);
- SAFE_CLOSEDIR(NULL, dir);
+ tst_res(TINFO, "Thread %s not sleeping yet", dent->d_name);
+ SAFE_CLOSEDIR(dir);
return 1;
}
cnt++;
}
- SAFE_CLOSEDIR(NULL, dir);
+ SAFE_CLOSEDIR(dir);
if (cnt != nr_threads) {
- tst_resm(TINFO, "%u threads sleeping, expected %u",
- cnt, nr_threads);
+ tst_res(TINFO, "%u threads sleeping, expected %u", cnt,
+ nr_threads);
}
return 0;
}
+static inline int process_state_wait2(pid_t pid, const char state)
+{
+ char proc_path[128], cur_state;
+
+ snprintf(proc_path, sizeof(proc_path), "/proc/%i/stat", pid);
+
+ for (;;) {
+ FILE *f = fopen(proc_path, "r");
+ if (!f) {
+ tst_res(TFAIL, "Failed to open '%s': %s\n", proc_path,
+ strerror(errno));
+ return 1;
+ }
+
+ if (fscanf(f, "%*i %*s %c", &cur_state) != 1) {
+ fclose(f);
+ tst_res(TFAIL, "Failed to read '%s': %s\n", proc_path,
+ strerror(errno));
+ return 1;
+ }
+ fclose(f);
+
+ if (state == cur_state)
+ return 0;
+
+ usleep(10000);
+ }
+}
+
#endif /* FUTEX_UTILS_H__ */
diff --git a/testcases/kernel/syscalls/futex/futex_wait01.c b/testcases/kernel/syscalls/futex/futex_wait01.c
index 51a540d9e315..66fc901e4b2c 100644
--- a/testcases/kernel/syscalls/futex/futex_wait01.c
+++ b/testcases/kernel/syscalls/futex/futex_wait01.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (C) 2015 Cyril Hrubis <chrubis@suse.cz>
*
@@ -5,33 +6,15 @@
* written by Darren Hart <dvhltc@us.ibm.com>
* Gowrishankar <gowrishankar.m@in.ibm.com>
*
- * Licensed under the GNU GPLv2 or later.
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ * 1. Block on a futex and wait for timeout.
+ * 2. Test if FUTEX_WAIT op returns -EWOULDBLOCK if the futex value differs
+ * from the expected one.
*/
- /*
- * 1. Block on a futex and wait for timeout.
- * 2. Test if FUTEX_WAIT op returns -EWOULDBLOCK if the futex value differs
- * from the expected one.
- */
#include <errno.h>
-#include "test.h"
#include "futextest.h"
-
-const char *TCID="futex_wait01";
+#include "lapi/abisize.h"
struct testcase {
futex_t *f_addr;
@@ -41,7 +24,6 @@ struct testcase {
};
static futex_t futex = FUTEX_INITIALIZER;
-static struct timespec to = {.tv_sec = 0, .tv_nsec = 10000};
static struct testcase testcases[] = {
{&futex, FUTEX_INITIALIZER, 0, ETIMEDOUT},
@@ -50,38 +32,59 @@ static struct testcase testcases[] = {
{&futex, FUTEX_INITIALIZER+1, FUTEX_PRIVATE_FLAG, EWOULDBLOCK},
};
-const int TST_TOTAL=ARRAY_SIZE(testcases);
+static struct test_variants {
+ enum futex_fn_type fntype;
+ enum tst_ts_type tstype;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .fntype = FUTEX_FN_FUTEX, .tstype = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .fntype = FUTEX_FN_FUTEX, .tstype = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_futex_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .fntype = FUTEX_FN_FUTEX64, .tstype = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
-static void verify_futex_wait(struct testcase *tc)
+static void run(unsigned int n)
{
+ struct test_variants *tv = &variants[tst_variant];
+ struct testcase *tc = &testcases[n];
+ static struct tst_ts to;
int res;
- res = futex_wait(tc->f_addr, tc->f_val, &to, tc->opflags);
+ to.type = tv->tstype;
+ tst_ts_set_sec(&to, 0);
+ tst_ts_set_nsec(&to, 10000);
+
+ res = futex_wait(tv->fntype, tc->f_addr, tc->f_val, &to, tc->opflags);
if (res != -1) {
- tst_resm(TFAIL, "futex_wait() returned %i, expected -1", res);
+ tst_res(TFAIL, "futex_wait() succeeded unexpectedly");
return;
}
if (errno != tc->exp_errno) {
- tst_resm(TFAIL | TERRNO, "expected errno=%s",
+ tst_res(TFAIL | TTERRNO, "futex_wait() failed with incorrect error, expected errno=%s",
tst_strerrno(tc->exp_errno));
return;
}
- tst_resm(TPASS | TERRNO, "futex_wait()");
+ tst_res(TPASS | TERRNO, "futex_wait() passed");
}
-int main(int argc, char *argv[])
+static void setup(void)
{
- int lc, i;
-
- tst_parse_opts(argc, argv, NULL, NULL);
-
- for (lc = 0; TEST_LOOPING(lc); lc++) {
- for (i = 0; i < TST_TOTAL; i++)
- verify_futex_wait(testcases + i);
- }
-
- tst_exit();
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
}
+
+static struct tst_test test = {
+ .setup = setup,
+ .test = run,
+ .tcnt = ARRAY_SIZE(testcases),
+ .test_variants = ARRAY_SIZE(variants),
+};
diff --git a/testcases/kernel/syscalls/futex/futex_wait02.c b/testcases/kernel/syscalls/futex/futex_wait02.c
index 1ca1df457b4a..b0f2ca2c2474 100644
--- a/testcases/kernel/syscalls/futex/futex_wait02.c
+++ b/testcases/kernel/syscalls/futex/futex_wait02.c
@@ -1,105 +1,95 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (C) 2015 Cyril Hrubis <chrubis@suse.cz>
*
- * Licensed under the GNU GPLv2 or later.
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
+ * Block on a futex and wait for wakeup.
*
- * This program 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ * This tests uses shared memory page to store the mutex variable.
*/
- /*
- * Block on a futex and wait for wakeup.
- *
- * This tests uses shared memory page to store the mutex variable.
- */
#include <sys/mman.h>
#include <sys/wait.h>
#include <errno.h>
-#include "test.h"
-#include "safe_macros.h"
#include "futextest.h"
-
-const char *TCID="futex_wait02";
-const int TST_TOTAL=1;
+#include "futex_utils.h"
+#include "lapi/abisize.h"
static futex_t *futex;
+static struct test_variants {
+ enum futex_fn_type fntype;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .fntype = FUTEX_FN_FUTEX, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .fntype = FUTEX_FN_FUTEX, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_futex_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .fntype = FUTEX_FN_FUTEX64, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
static void do_child(void)
{
+ struct test_variants *tv = &variants[tst_variant];
int ret;
- tst_process_state_wait2(getppid(), 'S');
+ process_state_wait2(getppid(), 'S');
- ret = futex_wake(futex, 1, 0);
+ ret = futex_wake(tv->fntype, futex, 1, 0);
if (ret != 1)
- tst_brkm(TFAIL, NULL, "futex_wake() returned %i", ret);
+ tst_res(TFAIL | TTERRNO, "futex_wake() failed");
- exit(TPASS);
+ exit(0);
}
-static void verify_futex_wait(void)
+static void run(void)
{
- int res;
- int pid;
+ struct test_variants *tv = &variants[tst_variant];
+ int res, pid;
- pid = tst_fork();
+ pid = SAFE_FORK();
switch (pid) {
case 0:
do_child();
- break;
- case -1:
- tst_brkm(TBROK | TERRNO, NULL, "fork() failed");
- break;
default:
- break;
+ break;
}
- res = futex_wait(futex, *futex, NULL, 0);
-
+ res = futex_wait(tv->fntype, futex, *futex, NULL, 0);
if (res) {
- tst_resm(TFAIL, "futex_wait() returned %i, errno %s",
- res, tst_strerrno(errno));
+ tst_res(TFAIL | TTERRNO, "futex_wait() failed");
+ return;
}
- SAFE_WAIT(NULL, &res);
+ SAFE_WAIT(&res);
if (WIFEXITED(res) && WEXITSTATUS(res) == TPASS)
- tst_resm(TPASS, "futex_wait() woken up");
+ tst_res(TPASS, "futex_wait() woken up");
else
- tst_resm(TFAIL, "child failed");
+ tst_res(TFAIL, "child failed");
}
static void setup(void)
{
- futex = SAFE_MMAP(NULL, NULL, sizeof(*futex), PROT_READ | PROT_WRITE,
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+
+ futex = SAFE_MMAP(NULL, sizeof(*futex), PROT_READ | PROT_WRITE,
MAP_ANONYMOUS | MAP_SHARED, -1, 0);
*futex = FUTEX_INITIALIZER;
}
-int main(int argc, char *argv[])
-{
- int lc;
-
- tst_parse_opts(argc, argv, NULL, NULL);
-
- setup();
-
- for (lc = 0; TEST_LOOPING(lc); lc++)
- verify_futex_wait();
-
- tst_exit();
-}
+static struct tst_test test = {
+ .setup = setup,
+ .test_all = run,
+ .test_variants = ARRAY_SIZE(variants),
+ .forks_child = 1,
+};
diff --git a/testcases/kernel/syscalls/futex/futex_wait03.c b/testcases/kernel/syscalls/futex/futex_wait03.c
index 9683e76503c0..a5ddb95c0000 100644
--- a/testcases/kernel/syscalls/futex/futex_wait03.c
+++ b/testcases/kernel/syscalls/futex/futex_wait03.c
@@ -1,87 +1,88 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (C) 2015 Cyril Hrubis <chrubis@suse.cz>
*
- * Licensed under the GNU GPLv2 or later.
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
+ * Block on a futex and wait for wakeup.
*
- * This program 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ * This tests uses private mutexes with threads.
*/
- /*
- * Block on a futex and wait for wakeup.
- *
- * This tests uses private mutexes with threads.
- */
#include <errno.h>
#include <pthread.h>
-#include "test.h"
#include "futextest.h"
-
-const char *TCID="futex_wait03";
-const int TST_TOTAL=1;
+#include "futex_utils.h"
+#include "lapi/abisize.h"
static futex_t futex = FUTEX_INITIALIZER;
+static struct test_variants {
+ enum futex_fn_type fntype;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .fntype = FUTEX_FN_FUTEX, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .fntype = FUTEX_FN_FUTEX, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_futex_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .fntype = FUTEX_FN_FUTEX64, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
static void *threaded(void *arg LTP_ATTRIBUTE_UNUSED)
{
+ struct test_variants *tv = &variants[tst_variant];
long ret;
- tst_process_state_wait2(getpid(), 'S');
+ process_state_wait2(getpid(), 'S');
- ret = futex_wake(&futex, 1, FUTEX_PRIVATE_FLAG);
+ ret = futex_wake(tv->fntype, &futex, 1, FUTEX_PRIVATE_FLAG);
return (void*)ret;
}
-static void verify_futex_wait(void)
+static void run(void)
{
- long ret;
- int res;
+ struct test_variants *tv = &variants[tst_variant];
+ long ret, res;
pthread_t t;
res = pthread_create(&t, NULL, threaded, NULL);
if (res) {
- tst_brkm(TBROK, NULL, "pthread_create(): %s",
- tst_strerrno(res));
+ tst_res(TFAIL | TTERRNO, "pthread_create() failed");
+ return;
}
- res = futex_wait(&futex, futex, NULL, FUTEX_PRIVATE_FLAG);
+ res = futex_wait(tv->fntype, &futex, futex, NULL, FUTEX_PRIVATE_FLAG);
if (res) {
- tst_resm(TFAIL, "futex_wait() returned %i, errno %s",
- res, tst_strerrno(errno));
+ tst_res(TFAIL | TTERRNO, "futex_wait() failed");
pthread_join(t, NULL);
return;
}
res = pthread_join(t, (void*)&ret);
- if (res)
- tst_brkm(TBROK, NULL, "pthread_join(): %s", tst_strerrno(res));
+ if (res) {
+ tst_res(TFAIL | TTERRNO, "pthread_join() failed");
+ return;
+ }
if (ret != 1)
- tst_resm(TFAIL, "futex_wake() returned %li", ret);
+ tst_res(TFAIL, "futex_wake() returned %li", ret);
else
- tst_resm(TPASS, "futex_wait() woken up");
+ tst_res(TPASS, "futex_wait() woken up");
}
-int main(int argc, char *argv[])
+static void setup(void)
{
- int lc;
-
- tst_parse_opts(argc, argv, NULL, NULL);
-
- for (lc = 0; TEST_LOOPING(lc); lc++)
- verify_futex_wait();
-
- tst_exit();
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
}
+
+static struct tst_test test = {
+ .setup = setup,
+ .test_all = run,
+ .test_variants = ARRAY_SIZE(variants),
+};
diff --git a/testcases/kernel/syscalls/futex/futex_wait04.c b/testcases/kernel/syscalls/futex/futex_wait04.c
index b8360ff80914..3a611494b694 100644
--- a/testcases/kernel/syscalls/futex/futex_wait04.c
+++ b/testcases/kernel/syscalls/futex/futex_wait04.c
@@ -1,65 +1,68 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (C) 2015 Cyril Hrubis <chrubis@suse.cz>
*
* Based on futextest (futext_wait_uninitialized_heap.c)
* written by KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
*
- * Licensed under the GNU GPLv2 or later.
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ * Wait on uninitialized heap. It shold be zero and FUTEX_WAIT should return
+ * immediately. This test tests zero page handling in futex code.
*/
- /*
- * Wait on uninitialized heap. It shold be zero and FUTEX_WAIT should return
- * immediately. This test tests zero page handling in futex code.
- */
#include <errno.h>
-#include "test.h"
-#include "safe_macros.h"
#include "futextest.h"
+#include "lapi/abisize.h"
+
+static struct test_variants {
+ enum futex_fn_type fntype;
+ enum tst_ts_type tstype;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .fntype = FUTEX_FN_FUTEX, .tstype = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .fntype = FUTEX_FN_FUTEX, .tstype = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
-const char *TCID="futex_wait04";
-const int TST_TOTAL=1;
-static struct timespec to = {.tv_sec = 0, .tv_nsec = 10000};
+#if (__NR_futex_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .fntype = FUTEX_FN_FUTEX64, .tstype = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
-static void verify_futex_wait(void)
+static void run(void)
{
- int res;
- void *buf;
+ struct test_variants *tv = &variants[tst_variant];
+ static struct tst_ts to;
size_t pagesize = getpagesize();
- buf = SAFE_MMAP(NULL, NULL, pagesize, PROT_READ|PROT_WRITE,
- MAP_PRIVATE|MAP_ANONYMOUS, 0, 0);
+ void *buf;
+ int res;
+
+ buf = SAFE_MMAP(NULL, pagesize, PROT_READ|PROT_WRITE,
+ MAP_PRIVATE|MAP_ANONYMOUS, 0, 0);
- res = futex_wait(buf, 1, &to, 0);
+ to.type = tv->tstype;
+ tst_ts_set_sec(&to, 0);
+ tst_ts_set_nsec(&to, 10000);
+ res = futex_wait(tv->fntype, buf, 1, &to, 0);
if (res == -1 && errno == EWOULDBLOCK)
- tst_resm(TPASS | TERRNO, "futex_wait() returned %i", res);
+ tst_res(TPASS | TERRNO, "futex_wait() returned %i", res);
else
- tst_resm(TFAIL | TERRNO, "futex_wait() returned %i", res);
+ tst_res(TFAIL | TERRNO, "futex_wait() returned %i", res);
- SAFE_MUNMAP(NULL, buf, pagesize);
+ SAFE_MUNMAP(buf, pagesize);
}
-int main(int argc, char *argv[])
+static void setup(void)
{
- int lc;
-
- tst_parse_opts(argc, argv, NULL, NULL);
-
- for (lc = 0; TEST_LOOPING(lc); lc++)
- verify_futex_wait();
-
- tst_exit();
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
}
+
+static struct tst_test test = {
+ .setup = setup,
+ .test_all = run,
+ .test_variants = ARRAY_SIZE(variants),
+};
diff --git a/testcases/kernel/syscalls/futex/futex_wait05.c b/testcases/kernel/syscalls/futex/futex_wait05.c
index 2573ae177d5b..8fad5d858716 100644
--- a/testcases/kernel/syscalls/futex/futex_wait05.c
+++ b/testcases/kernel/syscalls/futex/futex_wait05.c
@@ -19,7 +19,7 @@ int sample_fn(int clk_id, long long usec)
futex_t futex = FUTEX_INITIALIZER;
tst_timer_start(clk_id);
- TEST(futex_wait(&futex, futex, &to, 0));
+ TEST(syscall(SYS_futex, &futex, FUTEX_WAIT, futex, &to, NULL, 0));
tst_timer_stop();
tst_timer_sample();
diff --git a/testcases/kernel/syscalls/futex/futex_wait_bitset01.c b/testcases/kernel/syscalls/futex/futex_wait_bitset01.c
index b8546c3e640c..f3856697a964 100644
--- a/testcases/kernel/syscalls/futex/futex_wait_bitset01.c
+++ b/testcases/kernel/syscalls/futex/futex_wait_bitset01.c
@@ -12,6 +12,7 @@
#include "tst_test.h"
#include "tst_timer.h"
#include "futextest.h"
+#include "lapi/abisize.h"
#define THRESHOLD_US 100000
#define DEFAULT_TIMEOUT_US 100010
@@ -23,22 +24,44 @@ static struct test_case_t {
{ CLOCK_REALTIME }
};
+static struct test_variants {
+ enum futex_fn_type fntype;
+ enum tst_ts_type tstype;
+ int (*gettime)(clockid_t clk_id, void *ts);
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .fntype = FUTEX_FN_FUTEX, .tstype = TST_KERN_OLD_TIMESPEC, .gettime = sys_clock_gettime, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .fntype = FUTEX_FN_FUTEX, .tstype = TST_KERN_TIMESPEC, .gettime = sys_clock_gettime, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_futex_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .fntype = FUTEX_FN_FUTEX64, .tstype = TST_KERN_TIMESPEC, .gettime = sys_clock_gettime64, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
static void verify_futex_wait_bitset(long long wait_us, clock_t clk_id)
{
- struct timespec start, to, end;
+ struct test_variants *tv = &variants[tst_variant];
+ struct tst_ts start, to, end;
futex_t futex = FUTEX_INITIALIZER;
u_int32_t bitset = 0xffffffff;
int flags = clk_id == CLOCK_REALTIME ? FUTEX_CLOCK_REALTIME : 0;
+ start.type = end.type = to.type = tv->tstype;
+
tst_res(TINFO, "testing futex_wait_bitset() timeout with %s",
clk_id == CLOCK_REALTIME ? "CLOCK_REALTIME" : "CLOCK_MONOTONIC");
- clock_gettime(clk_id, &start);
- to = tst_timespec_add_us(start, wait_us);
+ tv->gettime(clk_id, tst_ts_get(&start));
+ to = tst_ts_add_us(start, wait_us);
- TEST(futex_wait_bitset(&futex, futex, &to, bitset, flags));
+ TEST(futex_wait_bitset(tv->fntype, &futex, futex, &to, bitset, flags));
- clock_gettime(clk_id, &end);
+ tv->gettime(clk_id, tst_ts_get(&end));
if (TST_RET != -1) {
tst_res(TFAIL, "futex_wait_bitset() returned %li, expected -1",
@@ -58,22 +81,22 @@ static void verify_futex_wait_bitset(long long wait_us, clock_t clk_id)
return;
}
- if (tst_timespec_lt(end, to)) {
+ if (tst_ts_lt(end, to)) {
tst_res(TFAIL,
"futex_wait_bitset() woken up prematurely %llius, expected %llius",
- tst_timespec_diff_us(end, start), wait_us);
+ tst_ts_diff_us(end, start), wait_us);
return;
}
- if (tst_timespec_diff_us(end, to) > THRESHOLD_US) {
+ if (tst_ts_diff_us(end, to) > THRESHOLD_US) {
tst_res(TFAIL,
"futex_wait_bitset() waited too long %llius, expected %llius",
- tst_timespec_diff_us(end, start), wait_us);
+ tst_ts_diff_us(end, start), wait_us);
return;
}
tst_res(TPASS, "futex_wait_bitset() waited %llius, expected %llius",
- tst_timespec_diff_us(end, start), wait_us);
+ tst_ts_diff_us(end, start), wait_us);
}
static void run(unsigned int n)
@@ -82,7 +105,14 @@ static void run(unsigned int n)
verify_futex_wait_bitset(DEFAULT_TIMEOUT_US, tcases[n].clk_id);
}
+static void setup(void)
+{
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+}
+
static struct tst_test test = {
+ .setup = setup,
.test = run,
.tcnt = ARRAY_SIZE(tcases),
+ .test_variants = ARRAY_SIZE(variants),
};
diff --git a/testcases/kernel/syscalls/futex/futex_wake01.c b/testcases/kernel/syscalls/futex/futex_wake01.c
index 42ea6f6e6a75..e6176adfbb22 100644
--- a/testcases/kernel/syscalls/futex/futex_wake01.c
+++ b/testcases/kernel/syscalls/futex/futex_wake01.c
@@ -1,32 +1,15 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (C) 2015 Cyril Hrubis <chrubis@suse.cz>
*
- * Licensed under the GNU GPLv2 or later.
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ * futex_wake() returns 0 (0 woken up processes) when no processes wait on the mutex.
*/
- /*
- * futex_wake() returns 0 (0 woken up processes) when no processes wait on the mutex.
- */
#include <errno.h>
#include <limits.h>
-#include "test.h"
#include "futextest.h"
-
-const char *TCID="futex_wake01";
+#include "lapi/abisize.h"
struct testcase {
futex_t *f_addr;
@@ -46,32 +29,46 @@ static struct testcase testcases[] = {
{&futex, INT_MAX, FUTEX_PRIVATE_FLAG},
};
-const int TST_TOTAL=ARRAY_SIZE(testcases);
+static struct test_variants {
+ enum futex_fn_type fntype;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .fntype = FUTEX_FN_FUTEX, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .fntype = FUTEX_FN_FUTEX, .desc = "syscall with kernel spec64"},
+#endif
-static void verify_futex_wake(struct testcase *tc)
+#if (__NR_futex_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .fntype = FUTEX_FN_FUTEX64, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
+static void run(unsigned int n)
{
+ struct test_variants *tv = &variants[tst_variant];
+ struct testcase *tc = &testcases[n];
int res;
- res = futex_wake(tc->f_addr, tc->nr_wake, tc->opflags);
-
+ res = futex_wake(tv->fntype, tc->f_addr, tc->nr_wake, tc->opflags);
if (res != 0) {
- tst_resm(TFAIL, "futex_wake() returned %i, expected 0", res);
+ tst_res(TFAIL | TTERRNO, "futex_wake() failed");
return;
}
- tst_resm(TPASS, "futex_wake() returned 0");
+ tst_res(TPASS, "futex_wake() passed");
}
-int main(int argc, char *argv[])
+static void setup(void)
{
- int lc, i;
-
- tst_parse_opts(argc, argv, NULL, NULL);
-
- for (lc = 0; TEST_LOOPING(lc); lc++) {
- for (i = 0; i < TST_TOTAL; i++)
- verify_futex_wake(testcases + i);
- }
-
- tst_exit();
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
}
+
+static struct tst_test test = {
+ .setup = setup,
+ .test = run,
+ .tcnt = ARRAY_SIZE(testcases),
+ .test_variants = ARRAY_SIZE(variants),
+};
diff --git a/testcases/kernel/syscalls/futex/futex_wake02.c b/testcases/kernel/syscalls/futex/futex_wake02.c
index 8a4c2d58e4b9..e37df08e5fac 100644
--- a/testcases/kernel/syscalls/futex/futex_wake02.c
+++ b/testcases/kernel/syscalls/futex/futex_wake02.c
@@ -1,40 +1,39 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (C) 2015 Cyril Hrubis <chrubis@suse.cz>
*
- * Licensed under the GNU GPLv2 or later.
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ * Block several threads on a private mutex, then wake them up.
*/
- /*
- * Block several threads on a private mutex, then wake them up.
- */
#include <errno.h>
#include <pthread.h>
+#include <sys/types.h>
-#include "test.h"
-#include "safe_macros.h"
#include "futextest.h"
#include "futex_utils.h"
-
-const char *TCID="futex_wake02";
-const int TST_TOTAL=11;
+#include "lapi/abisize.h"
static futex_t futex = FUTEX_INITIALIZER;
static volatile int threads_flags[55];
+static struct test_variants {
+ enum futex_fn_type fntype;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .fntype = FUTEX_FN_FUTEX, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .fntype = FUTEX_FN_FUTEX, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_futex_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .fntype = FUTEX_FN_FUTEX64, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
static int threads_awake(void)
{
int ret = 0;
@@ -58,9 +57,10 @@ static void clear_threads_awake(void)
static void *threaded(void *arg)
{
+ struct test_variants *tv = &variants[tst_variant];
long i = (long)arg;
- futex_wait(&futex, futex, NULL, FUTEX_PRIVATE_FLAG);
+ futex_wait(tv->fntype, &futex, futex, NULL, FUTEX_PRIVATE_FLAG);
threads_flags[i] = 1;
@@ -69,14 +69,15 @@ static void *threaded(void *arg)
static void do_child(void)
{
+ struct test_variants *tv = &variants[tst_variant];
int res, i, j, awake;
pthread_t t[55];
for (i = 0; i < (int)ARRAY_SIZE(t); i++) {
res = pthread_create(&t[i], NULL, threaded, (void*)((long)i));
if (res) {
- tst_brkm(TBROK, NULL, "pthread_create(): %s",
- tst_strerrno(res));
+ tst_res(TFAIL | TTERRNO, "pthread_create() failed");
+ return;
}
}
@@ -85,9 +86,9 @@ static void do_child(void)
for (i = 1; i <= 10; i++) {
clear_threads_awake();
- res = futex_wake(&futex, i, FUTEX_PRIVATE_FLAG);
+ res = futex_wake(tv->fntype, &futex, i, FUTEX_PRIVATE_FLAG);
if (i != res) {
- tst_resm(TFAIL,
+ tst_res(TFAIL | TTERRNO,
"futex_wake() woken up %i threads, expected %i",
res, i);
}
@@ -101,26 +102,26 @@ static void do_child(void)
}
if (awake == i) {
- tst_resm(TPASS, "futex_wake() woken up %i threads", i);
+ tst_res(TPASS, "futex_wake() woken up %i threads", i);
} else {
- tst_resm(TFAIL, "Woken up %i threads, expected %i",
- awake, i);
+ tst_res(TFAIL | TTERRNO, "Woken up %i threads, expected %i",
+ awake, i);
}
}
- res = futex_wake(&futex, 1, FUTEX_PRIVATE_FLAG);
+ res = futex_wake(tv->fntype, &futex, 1, FUTEX_PRIVATE_FLAG);
if (res) {
- tst_resm(TFAIL, "futex_wake() woken up %i, none were waiting",
- res);
+ tst_res(TFAIL | TTERRNO, "futex_wake() woken up %i, none were waiting",
+ res);
} else {
- tst_resm(TPASS, "futex_wake() woken up 0 threads");
+ tst_res(TPASS, "futex_wake() woken up 0 threads");
}
for (i = 0; i < (int)ARRAY_SIZE(t); i++)
pthread_join(t[i], NULL);
- tst_exit();
+ exit(0);
}
/*
@@ -135,30 +136,28 @@ static void do_child(void)
* under /proc/$PID/tasks/, but the subsequent open() fails with ENOENT because
* the thread was removed meanwhile.
*/
-static void verify_futex_wake(void)
+static void run(void)
{
- int pid;
+ int status, pid;
- pid = tst_fork();
+ pid = SAFE_FORK();
switch (pid) {
case 0:
do_child();
- case -1:
- tst_brkm(TBROK | TERRNO, NULL, "fork() failed");
default:
- tst_record_childstatus(NULL, pid);
+ SAFE_WAITPID(pid, &status, 0);
}
}
-int main(int argc, char *argv[])
+static void setup(void)
{
- int lc;
-
- tst_parse_opts(argc, argv, NULL, NULL);
-
- for (lc = 0; TEST_LOOPING(lc); lc++)
- verify_futex_wake();
-
- tst_exit();
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
}
+
+static struct tst_test test = {
+ .setup = setup,
+ .test_all = run,
+ .test_variants = ARRAY_SIZE(variants),
+ .forks_child = 1,
+};
diff --git a/testcases/kernel/syscalls/futex/futex_wake03.c b/testcases/kernel/syscalls/futex/futex_wake03.c
index d6e5e5422db9..b6c716d7afec 100644
--- a/testcases/kernel/syscalls/futex/futex_wake03.c
+++ b/testcases/kernel/syscalls/futex/futex_wake03.c
@@ -1,54 +1,56 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (C) 2015 Cyril Hrubis <chrubis@suse.cz>
*
- * Licensed under the GNU GPLv2 or later.
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ * Block several processes on a mutex, then wake them up.
*/
- /*
- * Block several processes on a mutex, then wake them up.
- */
#include <errno.h>
#include <sys/types.h>
#include <sys/wait.h>
-#include "test.h"
-#include "safe_macros.h"
#include "futextest.h"
-
-const char *TCID="futex_wake03";
-const int TST_TOTAL=11;
+#include "futex_utils.h"
+#include "lapi/abisize.h"
static futex_t *futex;
+static struct test_variants {
+ enum futex_fn_type fntype;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .fntype = FUTEX_FN_FUTEX, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .fntype = FUTEX_FN_FUTEX, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_futex_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .fntype = FUTEX_FN_FUTEX64, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
static void do_child(void)
{
- futex_wait(futex, *futex, NULL, 0);
+ struct test_variants *tv = &variants[tst_variant];
+
+ futex_wait(tv->fntype, futex, *futex, NULL, 0);
exit(0);
}
static void do_wake(int nr_children)
{
+ struct test_variants *tv = &variants[tst_variant];
int res, i, cnt;
- res = futex_wake(futex, nr_children, 0);
+ res = futex_wake(tv->fntype, futex, nr_children, 0);
if (res != nr_children) {
- tst_resm(TFAIL,
- "futex_wake() woken up %i children, expected %i",
- res, nr_children);
+ tst_res(TFAIL | TTERRNO,
+ "futex_wake() woken up %i children, expected %i",
+ res, nr_children);
return;
}
@@ -63,65 +65,60 @@ static void do_wake(int nr_children)
}
if (cnt != nr_children) {
- tst_resm(TFAIL, "reaped only %i childs, expected %i",
- cnt, nr_children);
+ tst_res(TFAIL | TTERRNO, "reaped only %i childs, expected %i",
+ cnt, nr_children);
} else {
- tst_resm(TPASS, "futex_wake() woken up %i childs", cnt);
+ tst_res(TPASS, "futex_wake() woken up %i childs", cnt);
}
}
-static void verify_futex_wake(void)
+static void run(void)
{
- int i, res;
+ struct test_variants *tv = &variants[tst_variant];
pid_t pids[55];
+ unsigned int i;
+ int res;
- for (i = 0; i < (int)ARRAY_SIZE(pids); i++) {
- pids[i] = tst_fork();
+ for (i = 0; i < ARRAY_SIZE(pids); i++) {
+ pids[i] = SAFE_FORK();
switch (pids[i]) {
- case -1:
- tst_brkm(TBROK | TERRNO, NULL, "fork()");
case 0:
do_child();
default:
- break;
+ break;
}
}
- for (i = 0; i < (int)ARRAY_SIZE(pids); i++)
- tst_process_state_wait2(pids[i], 'S');
+ for (i = 0; i < ARRAY_SIZE(pids); i++)
+ process_state_wait2(pids[i], 'S');
for (i = 1; i <= 10; i++)
do_wake(i);
- res = futex_wake(futex, 1, 0);
+ res = futex_wake(tv->fntype, futex, 1, 0);
if (res) {
- tst_resm(TFAIL, "futex_wake() woken up %u, none were waiting",
- res);
+ tst_res(TFAIL | TTERRNO, "futex_wake() woken up %u, none were waiting",
+ res);
} else {
- tst_resm(TPASS, "futex_wake() woken up 0 children");
+ tst_res(TPASS, "futex_wake() woken up 0 children");
}
}
static void setup(void)
{
- futex = SAFE_MMAP(NULL, NULL, sizeof(*futex), PROT_READ | PROT_WRITE,
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+
+ futex = SAFE_MMAP(NULL, sizeof(*futex), PROT_READ | PROT_WRITE,
MAP_ANONYMOUS | MAP_SHARED, -1, 0);
*futex = FUTEX_INITIALIZER;
}
-int main(int argc, char *argv[])
-{
- int lc;
-
- tst_parse_opts(argc, argv, NULL, NULL);
-
- setup();
-
- for (lc = 0; TEST_LOOPING(lc); lc++)
- verify_futex_wake();
-
- tst_exit();
-}
+static struct tst_test test = {
+ .setup = setup,
+ .test_all = run,
+ .test_variants = ARRAY_SIZE(variants),
+ .forks_child = 1,
+};
diff --git a/testcases/kernel/syscalls/futex/futex_wake04.c b/testcases/kernel/syscalls/futex/futex_wake04.c
index f92bda53f6c7..683a6263fba7 100644
--- a/testcases/kernel/syscalls/futex/futex_wake04.c
+++ b/testcases/kernel/syscalls/futex/futex_wake04.c
@@ -1,23 +1,9 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (C) 2015 Yi Zhang <wetpzy@gmail.com>
* Li Wang <liwang@redhat.com>
*
- * Licensed under the GNU GPLv2 or later.
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
- /* DESCRIPTION:
+ * DESCRIPTION:
*
* It is a regression test for commit:
* http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/
@@ -42,53 +28,63 @@
#include <sys/time.h>
#include <string.h>
-#include "test.h"
-#include "safe_macros.h"
#include "futextest.h"
#include "futex_utils.h"
#include "lapi/mmap.h"
+#include "lapi/abisize.h"
+#include "tst_safe_stdio.h"
#define PATH_MEMINFO "/proc/meminfo"
#define PATH_NR_HUGEPAGES "/proc/sys/vm/nr_hugepages"
#define PATH_HUGEPAGES "/sys/kernel/mm/hugepages/"
-const char *TCID = "futex_wake04";
-const int TST_TOTAL = 1;
-
static futex_t *futex1, *futex2;
-static struct timespec to = {.tv_sec = 30, .tv_nsec = 0};
+static struct tst_ts to;
static long orig_hugepages;
+static struct test_variants {
+ enum futex_fn_type fntype;
+ enum tst_ts_type tstype;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .fntype = FUTEX_FN_FUTEX, .tstype = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .fntype = FUTEX_FN_FUTEX, .tstype = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_futex_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .fntype = FUTEX_FN_FUTEX64, .tstype = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
static void setup(void)
{
- tst_require_root();
+ struct test_variants *tv = &variants[tst_variant];
- if ((tst_kvercmp(2, 6, 32)) < 0) {
- tst_brkm(TCONF, NULL, "This test can only run on kernels "
- "that are 2.6.32 or higher");
- }
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
- if (access(PATH_HUGEPAGES, F_OK))
- tst_brkm(TCONF, NULL, "Huge page is not supported.");
+ to.type = tv->tstype;
+ tst_ts_set_sec(&to, 30);
+ tst_ts_set_nsec(&to, 0);
- tst_tmpdir();
+ if (access(PATH_HUGEPAGES, F_OK))
+ tst_brk(TCONF, "Huge page is not supported.");
- SAFE_FILE_SCANF(NULL, PATH_NR_HUGEPAGES, "%ld", &orig_hugepages);
+ SAFE_FILE_SCANF(PATH_NR_HUGEPAGES, "%ld", &orig_hugepages);
if (orig_hugepages <= 0)
- SAFE_FILE_PRINTF(NULL, PATH_NR_HUGEPAGES, "%d", 1);
-
- TEST_PAUSE;
+ SAFE_FILE_PRINTF(PATH_NR_HUGEPAGES, "%d", 1);
}
static void cleanup(void)
{
if (orig_hugepages <= 0)
- SAFE_FILE_PRINTF(NULL, PATH_NR_HUGEPAGES, "%ld", orig_hugepages);
-
- tst_rmdir();
+ SAFE_FILE_PRINTF(PATH_NR_HUGEPAGES, "%ld", orig_hugepages);
}
static int read_hugepagesize(void)
@@ -97,60 +93,64 @@ static int read_hugepagesize(void)
char line[BUFSIZ], buf[BUFSIZ];
int val;
- fp = SAFE_FOPEN(cleanup, PATH_MEMINFO, "r");
+ fp = SAFE_FOPEN(PATH_MEMINFO, "r");
while (fgets(line, BUFSIZ, fp) != NULL) {
if (sscanf(line, "%64s %d", buf, &val) == 2)
if (strcmp(buf, "Hugepagesize:") == 0) {
- SAFE_FCLOSE(cleanup, fp);
+ SAFE_FCLOSE(fp);
return 1024 * val;
}
}
- SAFE_FCLOSE(cleanup, fp);
- tst_brkm(TBROK, NULL, "can't find \"%s\" in %s",
- "Hugepagesize:", PATH_MEMINFO);
+ SAFE_FCLOSE(fp);
+ tst_res(TFAIL, "can't find \"%s\" in %s", "Hugepagesize:",
+ PATH_MEMINFO);
+ return 0;
}
static void *wait_thread1(void *arg LTP_ATTRIBUTE_UNUSED)
{
- futex_wait(futex1, *futex1, &to, 0);
+ struct test_variants *tv = &variants[tst_variant];
+
+ futex_wait(tv->fntype, futex1, *futex1, &to, 0);
return NULL;
}
static void *wait_thread2(void *arg LTP_ATTRIBUTE_UNUSED)
{
+ struct test_variants *tv = &variants[tst_variant];
int res;
- res = futex_wait(futex2, *futex2, &to, 0);
+ res = futex_wait(tv->fntype, futex2, *futex2, &to, 0);
if (!res)
- tst_resm(TPASS, "Hi hydra, thread2 awake!");
+ tst_res(TPASS, "Hi hydra, thread2 awake!");
else
- tst_resm(TFAIL, "Bug: wait_thread2 did not wake after 30 secs.");
+ tst_res(TFAIL | TTERRNO, "Bug: wait_thread2 did not wake after 30 secs.");
return NULL;
}
static void wakeup_thread2(void)
{
+ struct test_variants *tv = &variants[tst_variant];
void *addr;
int hpsz, pgsz, res;
pthread_t th1, th2;
hpsz = read_hugepagesize();
- tst_resm(TINFO, "Hugepagesize %i", hpsz);
+ tst_res(TINFO, "Hugepagesize %i", hpsz);
/*allocate some shared memory*/
addr = mmap(NULL, hpsz, PROT_WRITE | PROT_READ,
MAP_SHARED | MAP_ANONYMOUS | MAP_HUGETLB, -1, 0);
if (addr == MAP_FAILED) {
- if (errno == ENOMEM) {
- tst_brkm(TCONF, NULL,
- "Cannot allocate hugepage, memory too fragmented?");
- }
+ if (errno == ENOMEM)
+ tst_res(TCONF, "Cannot allocate hugepage, memory too fragmented?");
- tst_brkm(TBROK | TERRNO, NULL, "Cannot allocate hugepage");
+ tst_res(TFAIL | TERRNO, "Cannot allocate hugepage");
+ return;
}
pgsz = getpagesize();
@@ -165,45 +165,47 @@ static void wakeup_thread2(void)
/*thread1 block on futex1 first,then thread2 block on futex2*/
res = pthread_create(&th1, NULL, wait_thread1, NULL);
if (res) {
- tst_brkm(TBROK, NULL, "pthread_create(): %s",
- tst_strerrno(res));
+ tst_res(TFAIL | TTERRNO, "pthread_create() failed");
+ return;
}
res = pthread_create(&th2, NULL, wait_thread2, NULL);
if (res) {
- tst_brkm(TBROK, NULL, "pthread_create(): %s",
- tst_strerrno(res));
+ tst_res(TFAIL | TTERRNO, "pthread_create() failed");
+ return;
}
while (wait_for_threads(2))
usleep(1000);
- futex_wake(futex2, 1, 0);
+ futex_wake(tv->fntype, futex2, 1, 0);
res = pthread_join(th2, NULL);
- if (res)
- tst_brkm(TBROK, NULL, "pthread_join(): %s", tst_strerrno(res));
+ if (res) {
+ tst_res(TFAIL | TTERRNO, "pthread_join() failed");
+ return;
+ }
- futex_wake(futex1, 1, 0);
+ futex_wake(tv->fntype, futex1, 1, 0);
res = pthread_join(th1, NULL);
if (res)
- tst_brkm(TBROK, NULL, "pthread_join(): %s", tst_strerrno(res));
+ tst_res(TFAIL | TTERRNO, "pthread_join() failed");
- SAFE_MUNMAP(NULL, addr, hpsz);
+ SAFE_MUNMAP(addr, hpsz);
}
-int main(int argc, char *argv[])
+static void run(void)
{
- int lc;
-
- tst_parse_opts(argc, argv, NULL, NULL);
-
- setup();
-
- for (lc = 0; TEST_LOOPING(lc); lc++)
- wakeup_thread2();
-
- cleanup();
- tst_exit();
+ wakeup_thread2();
}
+
+static struct tst_test test = {
+ .setup = setup,
+ .cleanup = cleanup,
+ .test_all = run,
+ .test_variants = ARRAY_SIZE(variants),
+ .needs_root = 1,
+ .min_kver = "2.6.32",
+ .needs_tmpdir = 1,
+};
diff --git a/testcases/kernel/syscalls/futex/futextest.h b/testcases/kernel/syscalls/futex/futextest.h
index 5754d36dae51..a9dcbbedd249 100644
--- a/testcases/kernel/syscalls/futex/futextest.h
+++ b/testcases/kernel/syscalls/futex/futextest.h
@@ -39,6 +39,7 @@
#include <sys/types.h>
#include <linux/futex.h>
#include "lapi/futex.h"
+#include "tst_timer.h"
#define FUTEX_INITIALIZER 0
@@ -82,18 +83,24 @@
# define FUTEX_CLOCK_REALTIME 256
#endif
+enum futex_fn_type {
+ FUTEX_FN_FUTEX,
+ FUTEX_FN_FUTEX64,
+};
+
/**
- * futex() - SYS_futex syscall wrapper
+ * futex_syscall() - futex syscall wrapper
+ * @fntype: Futex function type
* @uaddr: address of first futex
* @op: futex op code
* @val: typically expected value of uaddr, but varies by op
- * @timeout: typically an absolute struct timespec (except where noted
+ * @timeout: typically an absolute struct tst_ts (except where noted
* otherwise). Overloaded by some ops
* @uaddr2: address of second futex for some ops\
* @val3: varies by op
* @opflags: flags to be bitwise OR'd with op, such as FUTEX_PRIVATE_FLAG
*
- * futex() is used by all the following futex op wrappers. It can also be
+ * futex_syscall() is used by all the following futex op wrappers. It can also be
* used for misuse and abuse testing. Generally, the specific op wrappers
* should be used instead. It is a macro instead of an static inline function as
* some of the types over overloaded (timeout is used for nr_requeue for
@@ -102,17 +109,30 @@
* These argument descriptions are the defaults for all
* like-named arguments in the following wrappers except where noted below.
*/
-#define futex(uaddr, op, val, timeout, uaddr2, val3, opflags) \
- syscall(SYS_futex, uaddr, op | opflags, val, timeout, uaddr2, val3)
+static inline int futex_syscall(enum futex_fn_type fntype, futex_t *uaddr,
+ int futex_op, futex_t val, void *timeout,
+ futex_t *uaddr2, int val3, int opflags)
+{
+ int (*func)(int *uaddr, int futex_op, int val, void *to, int *uaddr2, int val3);
+
+ if (fntype == FUTEX_FN_FUTEX)
+ func = sys_futex;
+ else
+ func = sys_futex_time64;
+
+ return func((int *)uaddr, futex_op | opflags, val, timeout, (int *)uaddr2, val3);
+}
/**
* futex_wait() - block on uaddr with optional timeout
* @timeout: relative timeout
*/
static inline int
-futex_wait(futex_t *uaddr, futex_t val, struct timespec *timeout, int opflags)
+futex_wait(enum futex_fn_type fntype, futex_t *uaddr, futex_t val,
+ struct tst_ts *timeout, int opflags)
{
- return futex(uaddr, FUTEX_WAIT, val, timeout, NULL, 0, opflags);
+ return futex_syscall(fntype, uaddr, FUTEX_WAIT, val,
+ tst_ts_get(timeout), NULL, 0, opflags);
}
/**
@@ -120,9 +140,10 @@ futex_wait(futex_t *uaddr, futex_t val, struct timespec *timeout, int opflags)
* @nr_wake: wake up to this many tasks
*/
static inline int
-futex_wake(futex_t *uaddr, int nr_wake, int opflags)
+futex_wake(enum futex_fn_type fntype, futex_t *uaddr, int nr_wake, int opflags)
{
- return futex(uaddr, FUTEX_WAKE, nr_wake, NULL, NULL, 0, opflags);
+ return futex_syscall(fntype, uaddr, FUTEX_WAKE, nr_wake, NULL, NULL, 0,
+ opflags);
}
/**
@@ -130,11 +151,11 @@ futex_wake(futex_t *uaddr, int nr_wake, int opflags)
* @bitset: bitset to be used with futex_wake_bitset
*/
static inline int
-futex_wait_bitset(futex_t *uaddr, futex_t val, struct timespec *timeout,
- u_int32_t bitset, int opflags)
+futex_wait_bitset(enum futex_fn_type fntype, futex_t *uaddr, futex_t val,
+ struct tst_ts *timeout, u_int32_t bitset, int opflags)
{
- return futex(uaddr, FUTEX_WAIT_BITSET, val, timeout, NULL, bitset,
- opflags);
+ return futex_syscall(fntype, uaddr, FUTEX_WAIT_BITSET, val,
+ tst_ts_get(timeout), NULL, bitset, opflags);
}
/**
@@ -142,10 +163,11 @@ futex_wait_bitset(futex_t *uaddr, futex_t val, struct timespec *timeout,
* @bitset: bitset to compare with that used in futex_wait_bitset
*/
static inline int
-futex_wake_bitset(futex_t *uaddr, int nr_wake, u_int32_t bitset, int opflags)
+futex_wake_bitset(enum futex_fn_type fntype, futex_t *uaddr, int nr_wake,
+ u_int32_t bitset, int opflags)
{
- return futex(uaddr, FUTEX_WAKE_BITSET, nr_wake, NULL, NULL, bitset,
- opflags);
+ return futex_syscall(fntype, uaddr, FUTEX_WAKE_BITSET, nr_wake, NULL,
+ NULL, bitset, opflags);
}
/**
@@ -153,30 +175,32 @@ futex_wake_bitset(futex_t *uaddr, int nr_wake, u_int32_t bitset, int opflags)
* @detect: whether (1) or not (0) to perform deadlock detection
*/
static inline int
-futex_lock_pi(futex_t *uaddr, struct timespec *timeout, int detect,
- int opflags)
+futex_lock_pi(enum futex_fn_type fntype, futex_t *uaddr, struct tst_ts *timeout,
+ int detect, int opflags)
{
- return futex(uaddr, FUTEX_LOCK_PI, detect, timeout, NULL, 0, opflags);
+ return futex_syscall(fntype, uaddr, FUTEX_LOCK_PI, detect,
+ tst_ts_get(timeout), NULL, 0, opflags);
}
/**
* futex_unlock_pi() - release uaddr as a PI mutex, waking the top waiter
*/
static inline int
-futex_unlock_pi(futex_t *uaddr, int opflags)
+futex_unlock_pi(enum futex_fn_type fntype, futex_t *uaddr, int opflags)
{
- return futex(uaddr, FUTEX_UNLOCK_PI, 0, NULL, NULL, 0, opflags);
-}
+ return futex_syscall(fntype, uaddr, FUTEX_UNLOCK_PI, 0, NULL, NULL, 0,
+ opflags); }
/**
* futex_wake_op() - FIXME: COME UP WITH A GOOD ONE LINE DESCRIPTION
*/
static inline int
-futex_wake_op(futex_t *uaddr, futex_t *uaddr2, int nr_wake, int nr_wake2,
- int wake_op, int opflags)
+futex_wake_op(enum futex_fn_type fntype, futex_t *uaddr, futex_t *uaddr2,
+ int nr_wake, int nr_wake2, int wake_op, int opflags)
{
- return futex(uaddr, FUTEX_WAKE_OP, nr_wake, nr_wake2, uaddr2, wake_op,
- opflags);
+ return futex_syscall(fntype, uaddr, FUTEX_WAKE_OP, nr_wake,
+ (void *)((unsigned long)nr_wake2), uaddr2, wake_op,
+ opflags);
}
/**
@@ -188,11 +212,12 @@ futex_wake_op(futex_t *uaddr, futex_t *uaddr2, int nr_wake, int nr_wake2,
* favor of futex_cmp_requeue().
*/
static inline int
-futex_requeue(futex_t *uaddr, futex_t *uaddr2, int nr_wake, int nr_requeue,
- int opflags)
+futex_requeue(enum futex_fn_type fntype, futex_t *uaddr, futex_t *uaddr2,
+ int nr_wake, int nr_requeue, int opflags)
{
- return futex(uaddr, FUTEX_REQUEUE, nr_wake, nr_requeue, uaddr2, 0,
- opflags);
+ return futex_syscall(fntype, uaddr, FUTEX_REQUEUE, nr_wake,
+ (void *)((unsigned long)nr_requeue), uaddr2, 0,
+ opflags);
}
/**
@@ -201,11 +226,12 @@ futex_requeue(futex_t *uaddr, futex_t *uaddr2, int nr_wake, int nr_requeue,
* @nr_requeue: requeue up to this many tasks
*/
static inline int
-futex_cmp_requeue(futex_t *uaddr, futex_t val, futex_t *uaddr2, int nr_wake,
- int nr_requeue, int opflags)
+futex_cmp_requeue(enum futex_fn_type fntype, futex_t *uaddr, futex_t val,
+ futex_t *uaddr2, int nr_wake, int nr_requeue, int opflags)
{
- return futex(uaddr, FUTEX_CMP_REQUEUE, nr_wake, nr_requeue, uaddr2,
- val, opflags);
+ return futex_syscall(fntype, uaddr, FUTEX_CMP_REQUEUE, nr_wake,
+ (void *)((unsigned long)nr_requeue), uaddr2, val,
+ opflags);
}
/**
@@ -217,11 +243,11 @@ futex_cmp_requeue(futex_t *uaddr, futex_t val, futex_t *uaddr2, int nr_wake,
* paired with futex_cmp_requeue_pi().
*/
static inline int
-futex_wait_requeue_pi(futex_t *uaddr, futex_t val, futex_t *uaddr2,
- struct timespec *timeout, int opflags)
+futex_wait_requeue_pi(enum futex_fn_type fntype, futex_t *uaddr, futex_t val,
+ futex_t *uaddr2, struct tst_ts *timeout, int opflags)
{
- return futex(uaddr, FUTEX_WAIT_REQUEUE_PI, val, timeout, uaddr2, 0,
- opflags);
+ return futex_syscall(fntype, uaddr, FUTEX_WAIT_REQUEUE_PI, val,
+ tst_ts_get(timeout), uaddr2, 0, opflags);
}
/**
@@ -232,11 +258,12 @@ futex_wait_requeue_pi(futex_t *uaddr, futex_t val, futex_t *uaddr2,
* @nr_requeue: requeue up to this many tasks
*/
static inline int
-futex_cmp_requeue_pi(futex_t *uaddr, futex_t val, futex_t *uaddr2, int nr_wake,
- int nr_requeue, int opflags)
+futex_cmp_requeue_pi(enum futex_fn_type fntype, futex_t *uaddr, futex_t val,
+ futex_t *uaddr2, int nr_wake, int nr_requeue, int opflags)
{
- return futex(uaddr, FUTEX_CMP_REQUEUE_PI, nr_wake, nr_requeue, uaddr2, val,
- opflags);
+ return futex_syscall(fntype, uaddr, FUTEX_CMP_REQUEUE_PI, nr_wake,
+ (void *)((unsigned long)nr_requeue), uaddr2, val,
+ opflags);
}
/**
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 07/17] syscalls/io_pgetevents: Add support for time64 tests
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (5 preceding siblings ...)
2020-05-08 4:23 ` [LTP] [PATCH V2 06/17] syscalls/futex: Add support for time64 tests Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 08/17] syscalls/sigwaitinfo: Migrate to new test framework Viresh Kumar
` (9 subsequent siblings)
16 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
This adds support for time64 tests to the existing io_pgetevents()
syscall tests.
While at it, also pass mode to SAFE_OPEN() to avoid any errors when the
tests are run for multiple variants together.
Also get rid of an extra copy of local fd variable, which is incorrect.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
include/lapi/io_pgetevents.h | 15 +++--
.../syscalls/io_pgetevents/io_pgetevents01.c | 41 ++++++++++++-
.../syscalls/io_pgetevents/io_pgetevents02.c | 59 +++++++++++++++----
3 files changed, 97 insertions(+), 18 deletions(-)
diff --git a/include/lapi/io_pgetevents.h b/include/lapi/io_pgetevents.h
index 3c9d5b2d7618..5bb9a60c352a 100644
--- a/include/lapi/io_pgetevents.h
+++ b/include/lapi/io_pgetevents.h
@@ -16,15 +16,20 @@
#ifdef HAVE_LIBAIO
#include <libaio.h>
-#ifndef HAVE_IO_PGETEVENTS
-int io_pgetevents(io_context_t ctx, long min_nr, long max_nr,
- struct io_event *events, struct timespec *timeout,
- sigset_t *sigmask)
+static inline int sys_io_pgetevents(io_context_t ctx, long min_nr, long max_nr,
+ struct io_event *events, void *timeout, sigset_t *sigmask)
{
return tst_syscall(__NR_io_pgetevents, ctx, min_nr, max_nr, events,
timeout, sigmask);
}
-#endif /* HAVE_IO_PGETEVENTS */
+
+static inline int sys_io_pgetevents_time64(io_context_t ctx, long min_nr, long max_nr,
+ struct io_event *events, void *timeout, sigset_t *sigmask)
+{
+ return tst_syscall(__NR_io_pgetevents_time64, ctx, min_nr, max_nr,
+ events, timeout, sigmask);
+}
+
#endif /* HAVE_LIBAIO */
#endif /* IO_PGETEVENTS_H */
diff --git a/testcases/kernel/syscalls/io_pgetevents/io_pgetevents01.c b/testcases/kernel/syscalls/io_pgetevents/io_pgetevents01.c
index 50d2c5f9a3ab..f99879c07ea6 100644
--- a/testcases/kernel/syscalls/io_pgetevents/io_pgetevents01.c
+++ b/testcases/kernel/syscalls/io_pgetevents/io_pgetevents01.c
@@ -6,11 +6,38 @@
* Basic io_pgetevents() test to receive 1 event successfully.
*/
#include "tst_test.h"
+#include "tst_timer.h"
#include "lapi/io_pgetevents.h"
+#include "lapi/abisize.h"
#ifdef HAVE_LIBAIO
static int fd;
+static struct test_variants {
+ int (*io_pgetevents)(io_context_t ctx, long min_nr, long max_nr,
+ struct io_event *events, void *timeout, sigset_t *sigmask);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .io_pgetevents = sys_io_pgetevents, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+ { .io_pgetevents = sys_io_pgetevents, .type = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .io_pgetevents = sys_io_pgetevents, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_io_pgetevents_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .io_pgetevents = sys_io_pgetevents_time64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
+static void setup(void)
+{
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+}
+
static void cleanup(void)
{
if (fd > 0)
@@ -19,17 +46,23 @@ static void cleanup(void)
static void run(void)
{
+ struct test_variants *tv = &variants[tst_variant];
struct io_event events[1];
struct iocb cb, *cbs[1];
io_context_t ctx = 0;
+ struct tst_ts to;
sigset_t sigmask;
char data[4096];
- int ret, fd;
+ int ret;
+
+ to.type = tv->type;
+ tst_ts_set_sec(&to, 0);
+ tst_ts_set_nsec(&to, 10000);
cbs[0] = &cb;
sigemptyset(&sigmask);
- fd = SAFE_OPEN("io_pgetevents_file", O_RDWR | O_CREAT);
+ fd = SAFE_OPEN("io_pgetevents_file", O_RDWR | O_CREAT, 0644);
io_prep_pwrite(&cb, fd, data, 4096, 0);
ret = io_setup(1, &ctx);
@@ -41,7 +74,7 @@ static void run(void)
tst_brk(TBROK | TERRNO, "io_submit() failed");
/* get the reply */
- ret = io_pgetevents(ctx, 1, 1, events, NULL, &sigmask);
+ ret = tv->io_pgetevents(ctx, 1, 1, events, tst_ts_get(&to), &sigmask);
if (ret == 1)
tst_res(TPASS, "io_pgetevents() works as expected");
@@ -55,8 +88,10 @@ static void run(void)
static struct tst_test test = {
.min_kver = "4.18",
.test_all = run,
+ .test_variants = ARRAY_SIZE(variants),
.needs_tmpdir = 1,
.cleanup = cleanup,
+ .setup = setup,
};
#else
diff --git a/testcases/kernel/syscalls/io_pgetevents/io_pgetevents02.c b/testcases/kernel/syscalls/io_pgetevents/io_pgetevents02.c
index 7cca7fc08567..7c0c6806d005 100644
--- a/testcases/kernel/syscalls/io_pgetevents/io_pgetevents02.c
+++ b/testcases/kernel/syscalls/io_pgetevents/io_pgetevents02.c
@@ -6,7 +6,9 @@
* Basic io_pgetevents() test to check various failures.
*/
#include "tst_test.h"
+#include "tst_timer.h"
#include "lapi/io_pgetevents.h"
+#include "lapi/abisize.h"
#ifdef HAVE_LIBAIO
static sigset_t sigmask;
@@ -14,35 +16,65 @@ static struct io_event events[1];
static io_context_t ctx, invalid_ctx = 0;
static int fd, ctx_initialized;
+static struct tst_ts to;
+static void *bad_addr;
+
static struct tcase {
char *name;
io_context_t *ctx;
long min_nr;
long max_nr;
struct io_event *events;
- struct timespec *timeout;
+ struct tst_ts *timeout;
sigset_t *sigmask;
int exp_errno;
} tcases[] = {
- {"invalid ctx", &invalid_ctx, 1, 1, events, NULL, &sigmask, EINVAL},
- {"invalid min_nr", &ctx, -1, 1, events, NULL, &sigmask, EINVAL},
- {"invalid max_nr", &ctx, 1, -1, events, NULL, &sigmask, EINVAL},
- {"invalid events", &ctx, 1, 1, NULL, NULL, &sigmask, EFAULT},
- {"invalid timeout", &ctx, 1, 1, events, (void *)(0xDEAD), &sigmask, EFAULT},
- {"invalid sigmask", &ctx, 1, 1, events, NULL, (void *)(0xDEAD), EFAULT},
+ {"invalid ctx", &invalid_ctx, 1, 1, events, &to, &sigmask, EINVAL},
+ {"invalid min_nr", &ctx, -1, 1, events, &to, &sigmask, EINVAL},
+ {"invalid max_nr", &ctx, 1, -1, events, &to, &sigmask, EINVAL},
+ {"invalid events", &ctx, 1, 1, NULL, &to, &sigmask, EFAULT},
+ {"invalid timeout", &ctx, 1, 1, events, NULL, &sigmask, EFAULT},
+ {"invalid sigmask", &ctx, 1, 1, events, &to, NULL, EFAULT},
+};
+
+static struct test_variants {
+ int (*io_pgetevents)(io_context_t ctx, long min_nr, long max_nr,
+ struct io_event *events, void *timeout, sigset_t *sigmask);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .io_pgetevents = sys_io_pgetevents, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+ { .io_pgetevents = sys_io_pgetevents, .type = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .io_pgetevents = sys_io_pgetevents, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_io_pgetevents_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .io_pgetevents = sys_io_pgetevents_time64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
};
static void setup(void)
{
+ struct test_variants *tv = &variants[tst_variant];
struct iocb cb, *cbs[1];
char data[4096];
int ret;
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ bad_addr = tst_get_bad_addr(NULL);
+ to.type = tv->type;
+ tst_ts_set_sec(&to, 0);
+ tst_ts_set_nsec(&to, 10000);
+
cbs[0] = &cb;
sigemptyset(&sigmask);
- fd = SAFE_OPEN("io_pgetevents_file", O_RDWR | O_CREAT);
+ fd = SAFE_OPEN("io_pgetevents_file", O_RDWR | O_CREAT, 0644);
io_prep_pwrite(&cb, fd, data, 4096, 0);
ret = io_setup(1, &ctx);
@@ -69,10 +101,16 @@ static void cleanup(void)
static void run(unsigned int n)
{
+ struct test_variants *tv = &variants[tst_variant];
struct tcase *tc = &tcases[n];
+ struct timespec *to;
+ sigset_t *sigmask;
+
+ sigmask = tc->sigmask ? tc->sigmask : bad_addr;
+ to = tc->timeout ? tc->timeout : bad_addr;
- TEST(io_pgetevents(*tc->ctx, tc->min_nr, tc->max_nr, tc->events,
- tc->timeout, tc->sigmask));
+ TEST(tv->io_pgetevents(*tc->ctx, tc->min_nr, tc->max_nr, tc->events, to,
+ sigmask));
if (TST_RET == 1) {
tst_res(TFAIL, "%s: io_pgetevents() passed unexpectedly",
@@ -95,6 +133,7 @@ static struct tst_test test = {
.needs_tmpdir = 1,
.tcnt = ARRAY_SIZE(tcases),
.test = run,
+ .test_variants = ARRAY_SIZE(variants),
.setup = setup,
.cleanup = cleanup,
};
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 08/17] syscalls/sigwaitinfo: Migrate to new test framework
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (6 preceding siblings ...)
2020-05-08 4:23 ` [LTP] [PATCH V2 07/17] syscalls/io_pgetevents: " Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 09/17] syscalls/rt_sigtimedwait: Add support for time64 tests Viresh Kumar
` (8 subsequent siblings)
16 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
This migrates the sigwaitinfo tests to use the new test framework.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
.../syscalls/sigwaitinfo/sigwaitinfo01.c | 209 +++++++-----------
1 file changed, 84 insertions(+), 125 deletions(-)
diff --git a/testcases/kernel/syscalls/sigwaitinfo/sigwaitinfo01.c b/testcases/kernel/syscalls/sigwaitinfo/sigwaitinfo01.c
index 6a30c27f6f2f..1b976a271d28 100644
--- a/testcases/kernel/syscalls/sigwaitinfo/sigwaitinfo01.c
+++ b/testcases/kernel/syscalls/sigwaitinfo/sigwaitinfo01.c
@@ -1,93 +1,58 @@
-/*
- * Copyright (c) Jiri Palecek<jpalecek@web.de>, 2009
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of version 2 of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
- *
- * Further, this software is distributed without any warranty that it is
- * free of the rightful claim of any third person regarding infringement
- * or the like. Any license provided herein, whether implied or
- * otherwise, applies only to this software file. Patent licenses, if
- * any, provided herein do not apply to combinations of this program with
- * other software, or any other product whatsoever.
- *
- * You should have received a copy of the GNU General Public License along
- * with this program; if not, write the Free Software Foundation, Inc.,
- * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
- */
-#include "test.h"
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Copyright (c) Jiri Palecek<jpalecek@web.de>, 2009 */
+
+#include "tst_test.h"
#include <errno.h>
+#include <stdlib.h>
#include <signal.h>
-#include "../utils/include_j_h.h"
-#include "../utils/common_j_h.c"
#include <limits.h>
#include "lapi/syscalls.h"
+#include "tst_sig_proc.h"
#define SUCCEED_OR_DIE(syscall, message, ...) \
(errno = 0, \
({int ret=syscall(__VA_ARGS__); \
if (ret==-1) \
- tst_brkm(TBROK|TERRNO, cleanup, message);\
+ tst_brk(TBROK|TERRNO, message);\
ret;}))
-/* Report success iff TEST_RETURN and TEST_ERRNO are equal to
+/* Report success iff TST_RET and TST_ERR are equal to
exp_return and exp_errno, resp., and cond is true. If cond is not
true, report condition_errmsg
*/
-static void report_success_cond(const char *func, const char *file, int line,
+static void report_success_cond(const char *func, int line,
long exp_return, int exp_errno, int condition,
char *condition_errmsg)
{
- if (exp_return == TEST_RETURN
- && (exp_return != -1 || exp_errno == TEST_ERRNO))
+ if (exp_return == TST_RET
+ && (exp_return != -1 || exp_errno == TST_ERR))
if (condition)
- tst_resm(TPASS, "Test passed");
+ tst_res(TPASS, "%s (%d): Test passed", func, line);
else
- tst_resm(TFAIL, "%s (%s: %d): %s", func, file, line,
+ tst_res(TFAIL, "%s (%d): %s", func, line,
condition_errmsg);
- else if (TEST_RETURN != -1)
- tst_resm(TFAIL,
- "%s (%s: %d): Unexpected return value; expected %ld, got %ld",
- func, file, line, exp_return, TEST_RETURN);
+ else if (TST_RET != -1)
+ tst_res(TFAIL,
+ "%s (%d): Unexpected return value; expected %ld, got %ld",
+ func, line, exp_return, TST_RET);
else
- tst_resm(TFAIL | TTERRNO, "%s (%s: %d): Unexpected failure",
- func, file, line);
+ tst_res(TFAIL | TTERRNO, "%s (%d): Unexpected failure",
+ func, line);
}
#define REPORT_SUCCESS_COND(exp_return, exp_errno, condition, condition_errmsg) \
- report_success_cond(__FUNCTION__, __FILE__, __LINE__, exp_return, exp_errno, condition, condition_errmsg);
+ report_success_cond(__FUNCTION__, __LINE__, exp_return, exp_errno, condition, condition_errmsg);
-/* Report success iff TEST_RETURN and TEST_ERRNO are equal to
+/* Report success iff TST_RET and TST_ERR are equal to
exp_return and exp_errno, resp.
*/
#define REPORT_SUCCESS(exp_return, exp_errno) \
REPORT_SUCCESS_COND(exp_return, exp_errno, 1, "");
-static void cleanup(void);
-
static void empty_handler(int sig)
{
}
-static void setup(void)
-{
- tst_sig(FORK, DEF_HANDLER, cleanup);
- signal(SIGUSR1, empty_handler);
- signal(SIGALRM, empty_handler);
- signal(SIGUSR2, SIG_IGN);
-
- TEST_PAUSE;
-}
-
-static void cleanup(void)
-{
-}
-
typedef int (*swi_func) (const sigset_t * set, siginfo_t * info,
struct timespec * timeout);
typedef void (*test_func) (swi_func, int);
@@ -110,7 +75,6 @@ static int my_sigwait(const sigset_t * set, siginfo_t * info,
static int my_sigwaitinfo(const sigset_t * set, siginfo_t * info,
struct timespec *timeout)
{
-
return sigwaitinfo(set, info);
}
#endif
@@ -119,7 +83,6 @@ static int my_sigwaitinfo(const sigset_t * set, siginfo_t * info,
static int my_sigtimedwait(const sigset_t * set, siginfo_t * info,
struct timespec *timeout)
{
-
return sigtimedwait(set, info, timeout);
}
#endif
@@ -129,7 +92,7 @@ static int my_rt_sigtimedwait(const sigset_t * set, siginfo_t * info,
struct timespec *timeout)
{
/* _NSIG is always the right number of bits of signal map for all arches */
- return ltp_syscall(__NR_rt_sigtimedwait, set, info, timeout, _NSIG/8);
+ return tst_syscall(__NR_rt_sigtimedwait, set, info, timeout, _NSIG/8);
}
#endif
@@ -141,12 +104,13 @@ void test_empty_set(swi_func sigwaitinfo, int signo)
SUCCEED_OR_DIE(sigemptyset, "sigemptyset failed", &sigs);
/* Run a child that will wake us up */
- child = create_sig_proc(100000, signo, UINT_MAX);
+ child = create_sig_proc(signo, INT_MAX, 100000);
TEST(sigwaitinfo(&sigs, &si, NULL));
REPORT_SUCCESS(-1, EINTR);
- kill(child, SIGTERM);
+ SAFE_KILL(child, SIGTERM);
+ SAFE_WAIT(NULL);
}
void test_timeout(swi_func sigwaitinfo, int signo)
@@ -159,12 +123,13 @@ void test_timeout(swi_func sigwaitinfo, int signo)
SUCCEED_OR_DIE(sigemptyset, "sigemptyset failed", &sigs);
/* Run a child that will wake us up */
- child = create_sig_proc(100000, signo, UINT_MAX);
+ child = create_sig_proc(signo, INT_MAX, 100000);
TEST(sigwaitinfo(&sigs, &si, &ts));
REPORT_SUCCESS(-1, EAGAIN);
- kill(child, SIGTERM);
+ SAFE_KILL(child, SIGTERM);
+ SAFE_WAIT(NULL);
}
/* Note: sigwait-ing for a signal that is not blocked is unspecified
@@ -180,14 +145,15 @@ void test_unmasked_matching(swi_func sigwaitinfo, int signo)
SUCCEED_OR_DIE(sigaddset, "sigaddset failed", &sigs, signo);
/* Run a child that will wake us up */
- child = create_sig_proc(100000, signo, UINT_MAX);
+ child = create_sig_proc(signo, INT_MAX, 100000);
TEST(sigwaitinfo(&sigs, &si, NULL));
REPORT_SUCCESS_COND(signo, 0, si.si_pid == child
&& si.si_code == SI_USER
&& si.si_signo == signo, "Struct siginfo mismatch");
- kill(child, SIGTERM);
+ SAFE_KILL(child, SIGTERM);
+ SAFE_WAIT(NULL);
}
void test_unmasked_matching_noinfo(swi_func sigwaitinfo, int signo)
@@ -198,12 +164,13 @@ void test_unmasked_matching_noinfo(swi_func sigwaitinfo, int signo)
SUCCEED_OR_DIE(sigemptyset, "sigemptyset failed", &sigs);
SUCCEED_OR_DIE(sigaddset, "sigaddset failed", &sigs, signo);
/* Run a child that will wake us up */
- child = create_sig_proc(100000, signo, UINT_MAX);
+ child = create_sig_proc(signo, INT_MAX, 100000);
TEST(sigwaitinfo(&sigs, NULL, NULL));
REPORT_SUCCESS(signo, 0);
- kill(child, SIGTERM);
+ SAFE_KILL(child, SIGTERM);
+ SAFE_WAIT(NULL);
}
void test_masked_matching(swi_func sigwaitinfo, int signo)
@@ -224,7 +191,7 @@ void test_masked_matching(swi_func sigwaitinfo, int signo)
SUCCEED_OR_DIE(sigdelset, "sigaddset failed", &sigs, SIGCHLD);
/* Run a child that will wake us up */
- child = create_sig_proc(0, signo, 1);
+ child = create_sig_proc(signo, 1, 0);
TEST(sigwaitinfo(&sigs, &si, NULL));
REPORT_SUCCESS_COND(signo, 0, si.si_pid == child
@@ -234,13 +201,14 @@ void test_masked_matching(swi_func sigwaitinfo, int signo)
SUCCEED_OR_DIE(sigprocmask, "restoring original signal mask failed",
SIG_SETMASK, &oldmask, &oldmask);
- tst_count--;
-
if (sigismember(&oldmask, signo))
- tst_resm(TPASS, "sigwaitinfo restored the original mask");
+ tst_res(TPASS, "sigwaitinfo restored the original mask");
else
- tst_resm(TFAIL,
+ tst_res(TFAIL,
"sigwaitinfo failed to restore the original mask");
+
+ SAFE_KILL(child, SIGTERM);
+ SAFE_WAIT(NULL);
}
void test_masked_matching_rt(swi_func sigwaitinfo, int signo)
@@ -265,12 +233,12 @@ void test_masked_matching_rt(swi_func sigwaitinfo, int signo)
SUCCEED_OR_DIE(sigdelset, "sigdelset failed", &sigs, SIGCHLD);
/* Run a child that will wake us up */
- child[0] = create_sig_proc(0, signo, 1);
- child[1] = create_sig_proc(0, signo + 1, 1);
+ child[0] = create_sig_proc(signo, 1, 0);
+ child[1] = create_sig_proc(signo + 1, 1, 0);
/* Ensure that the signals have been sent */
- waitpid(child[0], &status, 0);
- waitpid(child[1], &status, 0);
+ SAFE_WAITPID(child[0], &status, 0);
+ SAFE_WAITPID(child[1], &status, 0);
TEST(sigwaitinfo(&sigs, &si, NULL));
REPORT_SUCCESS_COND(signo, 0, si.si_pid == child[0]
@@ -278,7 +246,6 @@ void test_masked_matching_rt(swi_func sigwaitinfo, int signo)
&& si.si_signo == signo, "Struct siginfo mismatch");
/* eat the other signal */
- tst_count--;
TEST(sigwaitinfo(&sigs, &si, NULL));
REPORT_SUCCESS_COND(signo + 1, 0, si.si_pid == child[1]
&& si.si_code == SI_USER
@@ -288,12 +255,10 @@ void test_masked_matching_rt(swi_func sigwaitinfo, int signo)
SUCCEED_OR_DIE(sigprocmask, "restoring original signal mask failed",
SIG_SETMASK, &oldmask, &oldmask);
- tst_count--;
-
if (sigismember(&oldmask, signo))
- tst_resm(TPASS, "sigwaitinfo restored the original mask");
+ tst_res(TPASS, "sigwaitinfo restored the original mask");
else
- tst_resm(TFAIL,
+ tst_res(TFAIL,
"sigwaitinfo failed to restore the original mask");
}
@@ -314,7 +279,7 @@ void test_masked_matching_noinfo(swi_func sigwaitinfo, int signo)
SUCCEED_OR_DIE(sigdelset, "sigaddset failed", &sigs, SIGCHLD);
/* Run a child that will wake us up */
- child = create_sig_proc(0, signo, 1);
+ child = create_sig_proc(signo, 1, 0);
TEST(sigwaitinfo(&sigs, NULL, NULL));
REPORT_SUCCESS(signo, 0);
@@ -322,14 +287,14 @@ void test_masked_matching_noinfo(swi_func sigwaitinfo, int signo)
SUCCEED_OR_DIE(sigprocmask, "restoring original signal mask failed",
SIG_SETMASK, &oldmask, &oldmask);
- tst_count--;
-
if (sigismember(&oldmask, signo))
- tst_resm(TPASS, "sigwaitinfo restored the original mask");
+ tst_res(TPASS, "sigwaitinfo restored the original mask");
else
- tst_resm(TFAIL,
+ tst_res(TFAIL,
"sigwaitinfo failed to restore the original mask");
+ SAFE_KILL(child, SIGTERM);
+ SAFE_WAIT(NULL);
}
void test_bad_address(swi_func sigwaitinfo, int signo)
@@ -349,7 +314,7 @@ void test_bad_address(swi_func sigwaitinfo, int signo)
SUCCEED_OR_DIE(sigdelset, "sigaddset failed", &sigs, SIGCHLD);
/* Run a child that will wake us up */
- child = create_sig_proc(0, signo, 1);
+ child = create_sig_proc(signo, 1, 0);
TEST(sigwaitinfo(&sigs, (void *)1, NULL));
REPORT_SUCCESS(-1, EFAULT);
@@ -357,7 +322,8 @@ void test_bad_address(swi_func sigwaitinfo, int signo)
SUCCEED_OR_DIE(sigprocmask, "sigprocmask failed", SIG_SETMASK, &oldmask,
&oldmask);
- kill(child, SIGTERM);
+ SAFE_KILL(child, SIGTERM);
+ SAFE_WAIT(NULL);
}
void test_bad_address2(swi_func sigwaitinfo, int signo)
@@ -367,7 +333,7 @@ void test_bad_address2(swi_func sigwaitinfo, int signo)
switch (pid = fork()) {
case -1:
- tst_brkm(TBROK | TERRNO, NULL, "fork() failed");
+ tst_brk(TBROK | TERRNO, "fork() failed");
case 0:
signal(SIGSEGV, SIG_DFL);
@@ -377,31 +343,31 @@ void test_bad_address2(swi_func sigwaitinfo, int signo)
*/
TEST(sigwaitinfo((void *)1, NULL, NULL));
- if (TEST_RETURN == -1 && TEST_ERRNO == EFAULT)
+ if (TST_RET == -1 && TST_ERR == EFAULT)
_exit(0);
- tst_resm(TINFO | TTERRNO, "swi_func returned: %ld",
- TEST_RETURN);
+ tst_res(TINFO | TTERRNO, "swi_func returned: %ld",
+ TST_RET);
_exit(1);
break;
default:
break;
}
- SUCCEED_OR_DIE(waitpid, "waitpid failed", pid, &status, 0);
+ SAFE_WAITPID(pid, &status, 0);
if ((WIFSIGNALED(status) && WTERMSIG(status) == SIGSEGV)
|| (WIFEXITED(status) && WEXITSTATUS(status) == 0)) {
- tst_resm(TPASS, "Test passed");
+ tst_res(TPASS, "Test passed");
return;
}
if (WIFEXITED(status)) {
- tst_resm(TFAIL, "Unrecognised child exit code: %d",
+ tst_res(TFAIL, "Unrecognised child exit code: %d",
WEXITSTATUS(status));
}
if (WIFSIGNALED(status)) {
- tst_resm(TFAIL, "Unrecognised child termsig: %d",
+ tst_res(TFAIL, "Unrecognised child termsig: %d",
WTERMSIG(status));
}
}
@@ -474,38 +440,31 @@ struct test_desc {
#endif
};
-#if defined TEST_SIGWAITINFO
-const char *TCID = "sigwaitinfo01";
-#elif defined TEST_RT_SIGTIMEDWAIT
-const char *TCID = "rt_sigtimedwait01";
-#elif defined TEST_SIGTIMEDWAIT
-const char *TCID = "sigtimedwait01";
-#elif defined TEST_SIGWAIT
-const char *TCID = "sigwait01";
-#endif
-
-int TST_TOTAL = ARRAY_SIZE(tests);
-
-int main(int argc, char **argv)
+static void run(unsigned int i)
{
- unsigned i;
- int lc;
-
- tst_parse_opts(argc, argv, NULL, NULL);
+ struct test_desc *tc = &tests[i];
- setup();
+ tc->tf(tc->swi, tc->signo);
+}
- for (lc = 0; TEST_LOOPING(lc); ++lc) {
- tst_count = 0;
+static void setup(void)
+{
+ signal(SIGUSR1, empty_handler);
+ signal(SIGALRM, empty_handler);
+ signal(SIGUSR2, SIG_IGN);
- for (i = 0; i < ARRAY_SIZE(tests); i++) {
- alarm(10); /* arrange a 10 second timeout */
- tst_resm(TINFO, "%p, %d", tests[i].swi, tests[i].signo);
- tests[i].tf(tests[i].swi, tests[i].signo);
- }
- alarm(0);
- }
+ alarm(10); /* arrange a 10 second timeout */
+}
- cleanup();
- tst_exit();
+static void cleanup(void)
+{
+ alarm(0);
}
+
+static struct tst_test test = {
+ .test= run,
+ .tcnt = ARRAY_SIZE(tests),
+ .setup = setup,
+ .cleanup = cleanup,
+ .forks_child = 1,
+};
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 09/17] syscalls/rt_sigtimedwait: Add support for time64 tests
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (7 preceding siblings ...)
2020-05-08 4:23 ` [LTP] [PATCH V2 08/17] syscalls/sigwaitinfo: Migrate to new test framework Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 10/17] syscalls/mq_timed{send|receive}: " Viresh Kumar
` (7 subsequent siblings)
16 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
This adds support for time64 tests to the existing rt_sigtimedwait()
syscall tests.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
.../syscalls/sigwaitinfo/sigwaitinfo01.c | 118 +++++++++++++-----
1 file changed, 86 insertions(+), 32 deletions(-)
diff --git a/testcases/kernel/syscalls/sigwaitinfo/sigwaitinfo01.c b/testcases/kernel/syscalls/sigwaitinfo/sigwaitinfo01.c
index 1b976a271d28..60d0a2b7cf04 100644
--- a/testcases/kernel/syscalls/sigwaitinfo/sigwaitinfo01.c
+++ b/testcases/kernel/syscalls/sigwaitinfo/sigwaitinfo01.c
@@ -2,11 +2,13 @@
/* Copyright (c) Jiri Palecek<jpalecek@web.de>, 2009 */
#include "tst_test.h"
+#include "tst_timer.h"
#include <errno.h>
#include <stdlib.h>
#include <signal.h>
#include <limits.h>
#include "lapi/syscalls.h"
+#include "lapi/abisize.h"
#include "tst_sig_proc.h"
#define SUCCEED_OR_DIE(syscall, message, ...) \
@@ -54,12 +56,12 @@ static void empty_handler(int sig)
}
typedef int (*swi_func) (const sigset_t * set, siginfo_t * info,
- struct timespec * timeout);
-typedef void (*test_func) (swi_func, int);
+ void * timeout);
+typedef void (*test_func) (swi_func, int, enum tst_ts_type type);
#ifdef TEST_SIGWAIT
static int my_sigwait(const sigset_t * set, siginfo_t * info,
- struct timespec *timeout)
+ void *timeout)
{
int ret;
int err = sigwait(set, &ret);
@@ -73,7 +75,7 @@ static int my_sigwait(const sigset_t * set, siginfo_t * info,
#ifdef TEST_SIGWAITINFO
static int my_sigwaitinfo(const sigset_t * set, siginfo_t * info,
- struct timespec *timeout)
+ void *timeout)
{
return sigwaitinfo(set, info);
}
@@ -81,7 +83,7 @@ static int my_sigwaitinfo(const sigset_t * set, siginfo_t * info,
#ifdef TEST_SIGTIMEDWAIT
static int my_sigtimedwait(const sigset_t * set, siginfo_t * info,
- struct timespec *timeout)
+ void *timeout)
{
return sigtimedwait(set, info, timeout);
}
@@ -89,14 +91,23 @@ static int my_sigtimedwait(const sigset_t * set, siginfo_t * info,
#ifdef TEST_RT_SIGTIMEDWAIT
static int my_rt_sigtimedwait(const sigset_t * set, siginfo_t * info,
- struct timespec *timeout)
+ void *timeout)
{
/* _NSIG is always the right number of bits of signal map for all arches */
return tst_syscall(__NR_rt_sigtimedwait, set, info, timeout, _NSIG/8);
}
+
+#if (__NR_rt_sigtimedwait_time64 != __LTP__NR_INVALID_SYSCALL)
+static int my_rt_sigtimedwait_time64(const sigset_t * set, siginfo_t * info,
+ void *timeout)
+{
+ /* _NSIG is always the right number of bits of signal map for all arches */
+ return tst_syscall(__NR_rt_sigtimedwait_time64, set, info, timeout, _NSIG/8);
+}
+#endif
#endif
-void test_empty_set(swi_func sigwaitinfo, int signo)
+void test_empty_set(swi_func sigwaitinfo, int signo, enum tst_ts_type type)
{
sigset_t sigs;
siginfo_t si;
@@ -113,19 +124,23 @@ void test_empty_set(swi_func sigwaitinfo, int signo)
SAFE_WAIT(NULL);
}
-void test_timeout(swi_func sigwaitinfo, int signo)
+void test_timeout(swi_func sigwaitinfo, int signo, enum tst_ts_type type)
{
sigset_t sigs;
siginfo_t si;
pid_t child;
- struct timespec ts = {.tv_sec = 1 };
+ struct tst_ts ts;
+
+ ts.type = type;
+ tst_ts_set_sec(&ts, 1);
+ tst_ts_set_nsec(&ts, 0);
SUCCEED_OR_DIE(sigemptyset, "sigemptyset failed", &sigs);
/* Run a child that will wake us up */
child = create_sig_proc(signo, INT_MAX, 100000);
- TEST(sigwaitinfo(&sigs, &si, &ts));
+ TEST(sigwaitinfo(&sigs, &si, tst_ts_get(&ts)));
REPORT_SUCCESS(-1, EAGAIN);
SAFE_KILL(child, SIGTERM);
@@ -135,7 +150,8 @@ void test_timeout(swi_func sigwaitinfo, int signo)
/* Note: sigwait-ing for a signal that is not blocked is unspecified
* by POSIX; but works for non-ignored signals under Linux
*/
-void test_unmasked_matching(swi_func sigwaitinfo, int signo)
+void test_unmasked_matching(swi_func sigwaitinfo, int signo,
+ enum tst_ts_type type)
{
sigset_t sigs;
siginfo_t si;
@@ -156,7 +172,8 @@ void test_unmasked_matching(swi_func sigwaitinfo, int signo)
SAFE_WAIT(NULL);
}
-void test_unmasked_matching_noinfo(swi_func sigwaitinfo, int signo)
+void test_unmasked_matching_noinfo(swi_func sigwaitinfo, int signo,
+ enum tst_ts_type type)
{
sigset_t sigs;
pid_t child;
@@ -173,7 +190,8 @@ void test_unmasked_matching_noinfo(swi_func sigwaitinfo, int signo)
SAFE_WAIT(NULL);
}
-void test_masked_matching(swi_func sigwaitinfo, int signo)
+void test_masked_matching(swi_func sigwaitinfo, int signo,
+ enum tst_ts_type type)
{
sigset_t sigs, oldmask;
siginfo_t si;
@@ -211,7 +229,8 @@ void test_masked_matching(swi_func sigwaitinfo, int signo)
SAFE_WAIT(NULL);
}
-void test_masked_matching_rt(swi_func sigwaitinfo, int signo)
+void test_masked_matching_rt(swi_func sigwaitinfo, int signo,
+ enum tst_ts_type type)
{
sigset_t sigs, oldmask;
siginfo_t si;
@@ -262,7 +281,8 @@ void test_masked_matching_rt(swi_func sigwaitinfo, int signo)
"sigwaitinfo failed to restore the original mask");
}
-void test_masked_matching_noinfo(swi_func sigwaitinfo, int signo)
+void test_masked_matching_noinfo(swi_func sigwaitinfo, int signo,
+ enum tst_ts_type type)
{
sigset_t sigs, oldmask;
pid_t child;
@@ -297,7 +317,7 @@ void test_masked_matching_noinfo(swi_func sigwaitinfo, int signo)
SAFE_WAIT(NULL);
}
-void test_bad_address(swi_func sigwaitinfo, int signo)
+void test_bad_address(swi_func sigwaitinfo, int signo, enum tst_ts_type type)
{
sigset_t sigs, oldmask;
pid_t child;
@@ -326,7 +346,7 @@ void test_bad_address(swi_func sigwaitinfo, int signo)
SAFE_WAIT(NULL);
}
-void test_bad_address2(swi_func sigwaitinfo, int signo)
+void test_bad_address2(swi_func sigwaitinfo, int signo, enum tst_ts_type type)
{
pid_t pid;
int status;
@@ -372,7 +392,7 @@ void test_bad_address2(swi_func sigwaitinfo, int signo)
}
}
-void test_bad_address3(swi_func sigwaitinfo, int signo)
+void test_bad_address3(swi_func sigwaitinfo, int signo, enum tst_ts_type type)
{
sigset_t sigs;
SUCCEED_OR_DIE(sigemptyset, "sigemptyset failed", &sigs);
@@ -388,28 +408,28 @@ struct test_desc {
} tests[] = {
#ifdef TEST_RT_SIGTIMEDWAIT
{
- test_empty_set, my_rt_sigtimedwait, SIGUSR1}, {
- test_unmasked_matching, my_rt_sigtimedwait, SIGUSR1}, {
- test_masked_matching, my_rt_sigtimedwait, SIGUSR1}, {
- test_unmasked_matching_noinfo, my_rt_sigtimedwait, SIGUSR1}, {
- test_masked_matching_noinfo, my_rt_sigtimedwait, SIGUSR1}, {
- test_bad_address, my_rt_sigtimedwait, SIGUSR1}, {
- test_bad_address2, my_rt_sigtimedwait, SIGUSR1}, {
- test_bad_address3, my_rt_sigtimedwait, SIGUSR1}, {
- test_timeout, my_rt_sigtimedwait, 0},
+ test_empty_set, NULL, SIGUSR1}, {
+ test_unmasked_matching, NULL, SIGUSR1}, {
+ test_masked_matching, NULL, SIGUSR1}, {
+ test_unmasked_matching_noinfo, NULL, SIGUSR1}, {
+ test_masked_matching_noinfo, NULL, SIGUSR1}, {
+ test_bad_address, NULL, SIGUSR1}, {
+ test_bad_address2, NULL, SIGUSR1}, {
+ test_bad_address3, NULL, SIGUSR1}, {
+ test_timeout, NULL, 0},
/* Special cases */
/* 1: sigwaitinfo does respond to ignored signal */
{
- test_masked_matching, my_rt_sigtimedwait, SIGUSR2},
+ test_masked_matching, NULL, SIGUSR2},
/* 2: An ignored signal doesn't cause sigwaitinfo to return EINTR */
{
- test_timeout, my_rt_sigtimedwait, SIGUSR2},
+ test_timeout, NULL, SIGUSR2},
/* 3: The handler is not called when the signal is waited for by sigwaitinfo */
{
- test_masked_matching, my_rt_sigtimedwait, SIGTERM},
+ test_masked_matching, NULL, SIGTERM},
/* 4: Simultaneous realtime signals are delivered in the order of increasing signal number */
{
- test_masked_matching_rt, my_rt_sigtimedwait, -1},
+ test_masked_matching_rt, NULL, -1},
#endif
#if defined TEST_SIGWAIT
{
@@ -440,15 +460,48 @@ struct test_desc {
#endif
};
+static struct test_variants {
+ swi_func swi;
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#ifdef TEST_RT_SIGTIMEDWAIT
+
+#if defined(TST_ABI32)
+ { .swi = my_rt_sigtimedwait, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+ { .swi = my_rt_sigtimedwait, .type = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .swi = my_rt_sigtimedwait, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_rt_sigtimedwait_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .swi = my_rt_sigtimedwait_time64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+
+#else /* !TEST_RT_SIGTIMEDWAIT */
+
+ { .swi = NULL, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+
+#endif /* TEST_RT_SIGTIMEDWAIT */
+};
+
static void run(unsigned int i)
{
+ struct test_variants *tv = &variants[tst_variant];
struct test_desc *tc = &tests[i];
+ swi_func swi;
- tc->tf(tc->swi, tc->signo);
+ swi = tv->swi ? tv->swi : tc->swi;
+
+ tc->tf(swi, tc->signo, tv->type);
}
static void setup(void)
{
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+
signal(SIGUSR1, empty_handler);
signal(SIGALRM, empty_handler);
signal(SIGUSR2, SIG_IGN);
@@ -464,6 +517,7 @@ static void cleanup(void)
static struct tst_test test = {
.test= run,
.tcnt = ARRAY_SIZE(tests),
+ .test_variants = ARRAY_SIZE(variants),
.setup = setup,
.cleanup = cleanup,
.forks_child = 1,
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 10/17] syscalls/mq_timed{send|receive}: Add support for time64 tests
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (8 preceding siblings ...)
2020-05-08 4:23 ` [LTP] [PATCH V2 09/17] syscalls/rt_sigtimedwait: Add support for time64 tests Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 11/17] syscalls/recvmmsg: " Viresh Kumar
` (6 subsequent siblings)
16 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
This adds support for time64 tests to the existing
mq_timed{send|receive}() syscall tests.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
include/tst_timer.h | 41 +++++++
.../mq_timedreceive/mq_timedreceive01.c | 97 ++++++++++++-----
.../syscalls/mq_timedsend/mq_timedsend01.c | 101 ++++++++++++------
testcases/kernel/syscalls/utils/mq_timed.h | 42 +++-----
4 files changed, 192 insertions(+), 89 deletions(-)
diff --git a/include/tst_timer.h b/include/tst_timer.h
index b4280042c77f..28f11c7376b4 100644
--- a/include/tst_timer.h
+++ b/include/tst_timer.h
@@ -14,6 +14,7 @@
#include <sched.h>
#include <sys/time.h>
+#include <mqueue.h>
#include <time.h>
#include "tst_test.h"
#include "lapi/common_timers.h"
@@ -255,6 +256,46 @@ static inline int sys_futex_time64(int *uaddr, int futex_op, int val, void *to,
return tst_syscall(__NR_futex_time64, uaddr, futex_op, val, to, uaddr2, val3);
}
+static inline int libc_mq_timedsend(mqd_t mqdes, const char *msg_ptr,
+ size_t msg_len, unsigned int msg_prio, void *abs_timeout)
+{
+ return mq_timedsend(mqdes, msg_ptr, msg_len, msg_prio, abs_timeout);
+}
+
+static inline int sys_mq_timedsend(mqd_t mqdes, const char *msg_ptr,
+ size_t msg_len, unsigned int msg_prio, void *abs_timeout)
+{
+ return tst_syscall(__NR_mq_timedsend, mqdes, msg_ptr, msg_len, msg_prio,
+ abs_timeout);
+}
+
+static inline int sys_mq_timedsend64(mqd_t mqdes, const char *msg_ptr,
+ size_t msg_len, unsigned int msg_prio, void *abs_timeout)
+{
+ return tst_syscall(__NR_mq_timedsend_time64, mqdes, msg_ptr, msg_len,
+ msg_prio, abs_timeout);
+}
+
+static inline ssize_t libc_mq_timedreceive(mqd_t mqdes, char *msg_ptr,
+ size_t msg_len, unsigned int *msg_prio, void *abs_timeout)
+{
+ return mq_timedreceive(mqdes, msg_ptr, msg_len, msg_prio, abs_timeout);
+}
+
+static inline ssize_t sys_mq_timedreceive(mqd_t mqdes, char *msg_ptr,
+ size_t msg_len, unsigned int *msg_prio, void *abs_timeout)
+{
+ return tst_syscall(__NR_mq_timedreceive, mqdes, msg_ptr, msg_len,
+ msg_prio, abs_timeout);
+}
+
+static inline ssize_t sys_mq_timedreceive64(mqd_t mqdes, char *msg_ptr,
+ size_t msg_len, unsigned int *msg_prio, void *abs_timeout)
+{
+ return tst_syscall(__NR_mq_timedreceive_time64, mqdes, msg_ptr, msg_len,
+ msg_prio, abs_timeout);
+}
+
static inline int libc_sched_rr_get_interval(pid_t pid, void *ts)
{
return sched_rr_get_interval(pid, ts);
diff --git a/testcases/kernel/syscalls/mq_timedreceive/mq_timedreceive01.c b/testcases/kernel/syscalls/mq_timedreceive/mq_timedreceive01.c
index 1be515ced671..f2a4c5c0f8f1 100644
--- a/testcases/kernel/syscalls/mq_timedreceive/mq_timedreceive01.c
+++ b/testcases/kernel/syscalls/mq_timedreceive/mq_timedreceive01.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (c) Crackerjack Project., 2007-2008, Hitachi, Ltd
* Copyright (c) 2017 Petr Vorel <pvorel@suse.cz>
@@ -6,29 +7,17 @@
* Takahiro Yasui <takahiro.yasui.mp@hitachi.com>,
* Yumiko Sugita <yumiko.sugita.yf@hitachi.com>,
* Satoshi Fujiwara <sa-fuji@sdl.hitachi.co.jp>
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation; either version 2 of
- * the License, or (at your option) any later version.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <errno.h>
#include <limits.h>
+#include "lapi/abisize.h"
static int fd, fd_root, fd_nonblock, fd_maxint = INT_MAX - 1, fd_invalid = -1;
#include "mq_timed.h"
-static struct timespec ts;
+static struct tst_ts ts;
static struct test_case tcase[] = {
{
@@ -95,21 +84,27 @@ static struct test_case tcase[] = {
{
.fd = &fd,
.len = 16,
- .rq = &(struct timespec) {.tv_sec = -1, .tv_nsec = 0},
+ .tv_sec = -1,
+ .tv_nsec = 0,
+ .rq = &ts,
.ret = -1,
.err = EINVAL,
},
{
.fd = &fd,
.len = 16,
- .rq = &(struct timespec) {.tv_sec = 0, .tv_nsec = -1},
+ .tv_sec = 0,
+ .tv_nsec = -1,
+ .rq = &ts,
.ret = -1,
.err = EINVAL,
},
{
.fd = &fd,
.len = 16,
- .rq = &(struct timespec) {.tv_sec = 0, .tv_nsec = 1000000000},
+ .tv_sec = 0,
+ .tv_nsec = 1000000000,
+ .rq = &ts,
.ret = -1,
.err = EINVAL,
},
@@ -131,8 +126,44 @@ static struct test_case tcase[] = {
},
};
+static struct test_variants {
+ int (*send)(mqd_t mqdes, const char *msg_ptr, size_t msg_len,
+ unsigned int msg_prio, void *abs_timeout);
+ ssize_t (*receive)(mqd_t mqdes, char *msg_ptr, size_t msg_len,
+ unsigned int *msg_prio, void *abs_timeout);
+
+ int (*gettime)(clockid_t clk_id, void *ts);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .gettime = libc_clock_gettime, .send = libc_mq_timedsend, .receive = libc_mq_timedreceive, .type = TST_LIBC_TIMESPEC, .desc = "vDSO or syscall with libc spec"},
+ { .gettime = sys_clock_gettime, .send = sys_mq_timedsend, .receive = sys_mq_timedreceive, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+ { .gettime = sys_clock_gettime, .send = sys_mq_timedsend, .receive = sys_mq_timedreceive, .type = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .gettime = sys_clock_gettime, .send = sys_mq_timedsend, .receive = sys_mq_timedreceive, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_mq_timedsend_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .gettime = sys_clock_gettime64, .send = sys_mq_timedsend64, .receive = sys_mq_timedreceive64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
+static void setup(void)
+{
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ ts.type = tv->type;
+
+ setup_common();
+}
+
static void do_test(unsigned int i)
{
+ struct test_variants *tv = &variants[tst_variant];
const struct test_case *tc = &tcase[i];
unsigned int j;
unsigned int prio;
@@ -140,19 +171,26 @@ static void do_test(unsigned int i)
char rmsg[len];
pid_t pid = -1;
+ tst_ts_set_sec(&ts, tc->tv_sec);
+ tst_ts_set_nsec(&ts, tc->tv_nsec);
+
if (tc->signal)
- pid = set_sig(tc->rq);
+ pid = set_sig(tc->rq, tv->gettime);
if (tc->timeout)
- set_timeout(tc->rq);
+ set_timeout(tc->rq, tv->gettime);
- if (tc->send)
- send_msg(*tc->fd, tc->len, tc->prio);
+ if (tc->send) {
+ if (tv->send(*tc->fd, smsg, tc->len, tc->prio, NULL) < 0) {
+ tst_res(TFAIL | TTERRNO, "mq_timedsend() failed");
+ return;
+ }
+ }
if (tc->invalid_msg)
len -= 1;
- TEST(mq_timedreceive(*tc->fd, rmsg, len, &prio, tc->rq));
+ TEST(tv->receive(*tc->fd, rmsg, len, &prio, tst_ts_get(tc->rq)));
if (pid > 0)
kill_pid(pid);
@@ -163,22 +201,22 @@ static void do_test(unsigned int i)
if (TST_RET < 0) {
if (tc->err != TST_ERR)
tst_res(TFAIL | TTERRNO,
- "mq_timedreceive failed unexpectedly, expected %s",
+ "mq_timedreceive() failed unexpectedly, expected %s",
tst_strerrno(tc->err));
else
- tst_res(TPASS | TTERRNO, "mq_timedreceive failed expectedly");
+ tst_res(TPASS | TTERRNO, "mq_timedreceive() failed expectedly");
return;
}
if (tc->len != TST_RET) {
- tst_res(TFAIL, "mq_timedreceive wrong length %ld, expected %zu",
+ tst_res(TFAIL, "mq_timedreceive() wrong length %ld, expected %u",
TST_RET, tc->len);
return;
}
if (tc->prio != prio) {
- tst_res(TFAIL, "mq_timedreceive wrong prio %d, expected %d",
+ tst_res(TFAIL, "mq_timedreceive() wrong prio %d, expected %d",
prio, tc->prio);
return;
}
@@ -186,20 +224,21 @@ static void do_test(unsigned int i)
for (j = 0; j < tc->len; j++) {
if (rmsg[j] != smsg[j]) {
tst_res(TFAIL,
- "mq_timedreceive wrong data %d in %u, expected %d",
+ "mq_timedreceive() wrong data %d in %u, expected %d",
rmsg[j], i, smsg[j]);
return;
}
}
- tst_res(TPASS, "mq_timedreceive returned %ld, priority %u, length: %zu",
+ tst_res(TPASS, "mq_timedreceive() returned %ld, priority %u, length: %zu",
TST_RET, prio, len);
}
static struct tst_test test = {
.tcnt = ARRAY_SIZE(tcase),
.test = do_test,
- .setup = setup_common,
+ .test_variants = ARRAY_SIZE(variants),
+ .setup = setup,
.cleanup = cleanup_common,
.forks_child = 1,
};
diff --git a/testcases/kernel/syscalls/mq_timedsend/mq_timedsend01.c b/testcases/kernel/syscalls/mq_timedsend/mq_timedsend01.c
index 8f209d396040..292b8ab5b429 100644
--- a/testcases/kernel/syscalls/mq_timedsend/mq_timedsend01.c
+++ b/testcases/kernel/syscalls/mq_timedsend/mq_timedsend01.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (c) Crackerjack Project., 2007-2008, Hitachi, Ltd
* Copyright (c) 2017 Petr Vorel <pvorel@suse.cz>
@@ -6,29 +7,18 @@
* Takahiro Yasui <takahiro.yasui.mp@hitachi.com>,
* Yumiko Sugita <yumiko.sugita.yf@hitachi.com>,
* Satoshi Fujiwara <sa-fuji@sdl.hitachi.co.jp>
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation; either version 2 of
- * the License, or (at your option) any later version.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <errno.h>
#include <limits.h>
+#include "lapi/abisize.h"
static int fd, fd_root, fd_nonblock, fd_maxint = INT_MAX - 1, fd_invalid = -1;
-static struct timespec ts;
#include "mq_timed.h"
+static struct tst_ts ts;
+
static struct test_case tcase[] = {
{
.fd = &fd,
@@ -95,7 +85,9 @@ static struct test_case tcase[] = {
{
.fd = &fd,
.len = 16,
- .rq = &(struct timespec) {.tv_sec = -1, .tv_nsec = 0},
+ .tv_sec = -1,
+ .tv_nsec = 0,
+ .rq = &ts,
.send = 1,
.ret = -1,
.err = EINVAL,
@@ -103,7 +95,9 @@ static struct test_case tcase[] = {
{
.fd = &fd,
.len = 16,
- .rq = &(struct timespec) {.tv_sec = 0, .tv_nsec = -1},
+ .tv_sec = 0,
+ .tv_nsec = -1,
+ .rq = &ts,
.send = 1,
.ret = -1,
.err = EINVAL,
@@ -111,7 +105,9 @@ static struct test_case tcase[] = {
{
.fd = &fd,
.len = 16,
- .rq = &(struct timespec) {.tv_sec = 0, .tv_nsec = 1000000000},
+ .tv_sec = 0,
+ .tv_nsec = 1000000000,
+ .rq = &ts,
.send = 1,
.ret = -1,
.err = EINVAL,
@@ -136,8 +132,44 @@ static struct test_case tcase[] = {
},
};
+static struct test_variants {
+ int (*send)(mqd_t mqdes, const char *msg_ptr, size_t msg_len,
+ unsigned int msg_prio, void *abs_timeout);
+ ssize_t (*receive)(mqd_t mqdes, char *msg_ptr, size_t msg_len,
+ unsigned int *msg_prio, void *abs_timeout);
+
+ int (*gettime)(clockid_t clk_id, void *ts);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .gettime = libc_clock_gettime, .send = libc_mq_timedsend, .receive = libc_mq_timedreceive, .type = TST_LIBC_TIMESPEC, .desc = "vDSO or syscall with libc spec"},
+ { .gettime = sys_clock_gettime, .send = sys_mq_timedsend, .receive = sys_mq_timedreceive, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+ { .gettime = sys_clock_gettime, .send = sys_mq_timedsend, .receive = sys_mq_timedreceive, .type = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .gettime = sys_clock_gettime, .send = sys_mq_timedsend, .receive = sys_mq_timedreceive, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_mq_timedsend_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .gettime = sys_clock_gettime64, .send = sys_mq_timedsend64, .receive = sys_mq_timedreceive64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
+static void setup(void)
+{
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ ts.type = tv->type;
+
+ setup_common();
+}
+
static void do_test(unsigned int i)
{
+ struct test_variants *tv = &variants[tst_variant];
const struct test_case *tc = &tcase[i];
unsigned int j;
unsigned int prio;
@@ -145,18 +177,24 @@ static void do_test(unsigned int i)
char rmsg[len];
pid_t pid = -1;
+ tst_ts_set_sec(&ts, tc->tv_sec);
+ tst_ts_set_nsec(&ts, tc->tv_nsec);
+
if (tc->signal)
- pid = set_sig(tc->rq);
+ pid = set_sig(tc->rq, tv->gettime);
if (tc->timeout)
- set_timeout(tc->rq);
+ set_timeout(tc->rq, tv->gettime);
if (tc->send) {
for (j = 0; j < MSG_LENGTH; j++)
- send_msg(*tc->fd, tc->len, tc->prio);
+ if (tv->send(*tc->fd, smsg, tc->len, tc->prio, NULL) < 0) {
+ tst_res(TFAIL | TTERRNO, "mq_timedsend() failed");
+ return;
+ }
}
- TEST(mq_timedsend(*tc->fd, smsg, tc->len, tc->prio, tc->rq));
+ TEST(tv->send(*tc->fd, smsg, tc->len, tc->prio, tst_ts_get(tc->rq)));
if (pid > 0)
kill_pid(pid);
@@ -164,10 +202,10 @@ static void do_test(unsigned int i)
if (TST_RET < 0) {
if (tc->err != TST_ERR)
tst_res(TFAIL | TTERRNO,
- "mq_timedsend failed unexpectedly, expected %s",
+ "mq_timedsend() failed unexpectedly, expected %s",
tst_strerrno(tc->err));
else
- tst_res(TPASS | TTERRNO, "mq_timedreceive failed expectedly");
+ tst_res(TPASS | TTERRNO, "mq_timedreceive() failed expectedly");
if (*tc->fd == fd)
cleanup_queue(fd);
@@ -175,7 +213,7 @@ static void do_test(unsigned int i)
return;
}
- TEST(mq_timedreceive(*tc->fd, rmsg, len, &prio, tc->rq));
+ TEST(tv->receive(*tc->fd, rmsg, len, &prio, tst_ts_get(tc->rq)));
if (*tc->fd == fd)
cleanup_queue(fd);
@@ -183,26 +221,26 @@ static void do_test(unsigned int i)
if (TST_RET < 0) {
if (tc->err != TST_ERR) {
tst_res(TFAIL | TTERRNO,
- "mq_timedreceive failed unexpectedly, expected %s",
+ "mq_timedreceive() failed unexpectedly, expected %s",
tst_strerrno(tc->err));
return;
}
if (tc->ret >= 0) {
- tst_res(TFAIL | TTERRNO, "mq_timedreceive returned %ld, expected %d",
+ tst_res(TFAIL | TTERRNO, "mq_timedreceive() returned %ld, expected %d",
TST_RET, tc->ret);
return;
}
}
if (tc->len != TST_RET) {
- tst_res(TFAIL, "mq_timedreceive wrong length %ld, expected %d",
+ tst_res(TFAIL, "mq_timedreceive() wrong length %ld, expected %u",
TST_RET, tc->len);
return;
}
if (tc->prio != prio) {
- tst_res(TFAIL, "mq_timedreceive wrong prio %d, expected %d",
+ tst_res(TFAIL, "mq_timedreceive() wrong prio %d, expected %d",
prio, tc->prio);
return;
}
@@ -210,20 +248,21 @@ static void do_test(unsigned int i)
for (j = 0; j < tc->len; j++) {
if (rmsg[j] != smsg[j]) {
tst_res(TFAIL,
- "mq_timedreceive wrong data %d in %u, expected %d",
+ "mq_timedreceive() wrong data %d in %u, expected %d",
rmsg[j], i, smsg[j]);
return;
}
}
- tst_res(TPASS, "mq_timedreceive returned %ld, priority %u, length: %zu",
+ tst_res(TPASS, "mq_timedreceive() returned %ld, priority %u, length: %zu",
TST_RET, prio, len);
}
static struct tst_test test = {
.tcnt = ARRAY_SIZE(tcase),
.test = do_test,
- .setup = setup_common,
+ .test_variants = ARRAY_SIZE(variants),
+ .setup = setup,
.cleanup = cleanup_common,
.forks_child = 1,
};
diff --git a/testcases/kernel/syscalls/utils/mq_timed.h b/testcases/kernel/syscalls/utils/mq_timed.h
index a163ddcee3af..1d5515f9d629 100644
--- a/testcases/kernel/syscalls/utils/mq_timed.h
+++ b/testcases/kernel/syscalls/utils/mq_timed.h
@@ -1,30 +1,21 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Copyright (c) 2017 Petr Vorel <pvorel@suse.cz>
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation; either version 2 of
- * the License, or (at your option) any later version.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#ifndef MQ_TIMED_H
#define MQ_TIMED_H
#include "mq.h"
+#include "tst_timer.h"
struct test_case {
int *fd;
unsigned int len;
unsigned int prio;
- struct timespec *rq;
+ struct tst_ts *rq;
+ long tv_sec;
+ long tv_nsec;
int invalid_msg;
int send;
int signal;
@@ -33,27 +24,20 @@ struct test_case {
int err;
};
-static pid_t set_sig(struct timespec *ts)
+static pid_t set_sig(struct tst_ts *ts,
+ int (*gettime)(clockid_t clk_id, void *ts))
{
- clock_gettime(CLOCK_REALTIME, ts);
- ts->tv_sec += 3;
+ gettime(CLOCK_REALTIME, tst_ts_get(ts));
+ *ts = tst_ts_add_us(*ts, 3000000);
return create_sig_proc(SIGINT, 40, 200000);
}
-static void set_timeout(struct timespec *ts)
+static void set_timeout(struct tst_ts *ts,
+ int (*gettime)(clockid_t clk_id, void *ts))
{
- clock_gettime(CLOCK_REALTIME, ts);
- ts->tv_nsec += 50000000;
- ts->tv_sec += ts->tv_nsec / 1000000000;
- ts->tv_nsec %= 1000000000;
-}
-
-static void send_msg(int fd, int len, int prio)
-{
- if (mq_timedsend(fd, smsg, len, prio,
- &((struct timespec){0})) < 0)
- tst_brk(TBROK | TERRNO, "mq_timedsend failed");
+ gettime(CLOCK_REALTIME, tst_ts_get(ts));
+ *ts = tst_ts_add_us(*ts, 50000);
}
static void kill_pid(pid_t pid)
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 11/17] syscalls/recvmmsg: Add support for time64 tests
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (9 preceding siblings ...)
2020-05-08 4:23 ` [LTP] [PATCH V2 10/17] syscalls/mq_timed{send|receive}: " Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 12/17] syscalls/ppoll: " Viresh Kumar
` (5 subsequent siblings)
16 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
This adds support for time64 tests to the existing recvmmsg() syscall
tests.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
.../kernel/syscalls/sendmmsg/sendmmsg01.c | 45 +++++++++++----
.../kernel/syscalls/sendmmsg/sendmmsg_var.h | 55 ++++++++-----------
2 files changed, 58 insertions(+), 42 deletions(-)
diff --git a/testcases/kernel/syscalls/sendmmsg/sendmmsg01.c b/testcases/kernel/syscalls/sendmmsg/sendmmsg01.c
index 37084102ee09..5d1fd1b6b9fa 100644
--- a/testcases/kernel/syscalls/sendmmsg/sendmmsg01.c
+++ b/testcases/kernel/syscalls/sendmmsg/sendmmsg01.c
@@ -14,8 +14,8 @@
#include "tst_test.h"
#include "lapi/socket.h"
#include "tst_safe_macros.h"
-
#include "sendmmsg_var.h"
+#include "lapi/abisize.h"
#define BUFSIZE 16
#define VLEN 2
@@ -25,27 +25,52 @@ static int receive_sockfd;
static struct mmsghdr *snd_msg, *rcv_msg;
static struct iovec *snd1, *snd2, *rcv1, *rcv2;
+static struct test_variants {
+ int (*receive)(int sockfd, struct mmsghdr *msgvec, unsigned int vlen,
+ unsigned int flags, void *timeout);
+ int (*send)(int sockfd, struct mmsghdr *msgvec, unsigned int vlen,
+ unsigned int flags);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .receive = libc_recvmmsg, .send = libc_sendmmsg, .type = TST_LIBC_TIMESPEC, .desc = "vDSO or syscall with libc spec"},
+ { .receive = sys_recvmmsg, .send = sys_sendmmsg, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+ { .receive = sys_recvmmsg, .send = sys_sendmmsg, .type = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .receive = sys_recvmmsg, .send = sys_sendmmsg, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_recvmmsg_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .receive = sys_recvmmsg64, .send = sys_sendmmsg, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
static void run(void)
{
- struct timespec timeout;
+ struct test_variants *tv = &variants[tst_variant];
+ struct tst_ts timeout;
int retval;
- retval = do_sendmmsg(send_sockfd, snd_msg, VLEN, 0);
+ retval = tv->send(send_sockfd, snd_msg, VLEN, 0);
if (retval < 0 || snd_msg[0].msg_len != 6 || snd_msg[1].msg_len != 6) {
- tst_res(TFAIL|TTERRNO, "sendmmsg failed");
+ tst_res(TFAIL|TTERRNO, "sendmmsg() failed");
return;
}
memset(rcv1->iov_base, 0, rcv1->iov_len);
memset(rcv2->iov_base, 0, rcv2->iov_len);
- timeout.tv_sec = 1;
- timeout.tv_nsec = 0;
+ timeout.type = tv->type;
+ tst_ts_set_sec(&timeout, 1);
+ tst_ts_set_nsec(&timeout, 0);
- retval = do_recvmmsg(receive_sockfd, rcv_msg, VLEN, 0, &timeout);
+ retval = tv->receive(receive_sockfd, rcv_msg, VLEN, 0, tst_ts_get(&timeout));
if (retval == -1) {
- tst_res(TFAIL | TTERRNO, "recvmmsg failed");
+ tst_res(TFAIL | TTERRNO, "recvmmsg() failed");
return;
}
if (retval != 2) {
@@ -96,7 +121,7 @@ static void setup(void)
rcv_msg[1].msg_hdr.msg_iov = rcv2;
rcv_msg[1].msg_hdr.msg_iovlen = 1;
- test_info();
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
}
static void cleanup(void)
@@ -111,7 +136,7 @@ static struct tst_test test = {
.test_all = run,
.setup = setup,
.cleanup = cleanup,
- .test_variants = TEST_VARIANTS,
+ .test_variants = ARRAY_SIZE(variants),
.bufs = (struct tst_buffers []) {
{&snd1, .iov_sizes = (int[]){3, 3, -1}},
{&snd2, .iov_sizes = (int[]){6, -1}},
diff --git a/testcases/kernel/syscalls/sendmmsg/sendmmsg_var.h b/testcases/kernel/syscalls/sendmmsg/sendmmsg_var.h
index f00cf056a747..a142b17f9718 100644
--- a/testcases/kernel/syscalls/sendmmsg/sendmmsg_var.h
+++ b/testcases/kernel/syscalls/sendmmsg/sendmmsg_var.h
@@ -6,55 +6,46 @@
#ifndef SENDMMSG_VAR__
#define SENDMMSG_VAR__
+#include "tst_timer.h"
#include "lapi/syscalls.h"
-static int do_sendmmsg(int sockfd, struct mmsghdr *msgvec, unsigned int vlen,
- int flags)
+static inline int libc_sendmmsg(int sockfd, struct mmsghdr *msgvec,
+ unsigned int vlen, unsigned int flags)
{
- switch (tst_variant) {
- case 0:
- return tst_syscall(__NR_sendmmsg, sockfd, msgvec, vlen, flags);
- case 1:
#ifdef HAVE_SENDMMSG
- return sendmmsg(sockfd, msgvec, vlen, flags);
+ return sendmmsg(sockfd, msgvec, vlen, flags);
#else
- tst_brk(TCONF, "libc sendmmsg not present");
+ tst_brk(TCONF, "libc sendmmsg not present");
#endif
- }
+}
- return -1;
+static inline int sys_sendmmsg(int sockfd, struct mmsghdr *msgvec,
+ unsigned int vlen, unsigned int flags)
+{
+ return tst_syscall(__NR_sendmmsg, sockfd, msgvec, vlen, flags);
}
-static int do_recvmmsg(int sockfd, struct mmsghdr *msgvec, unsigned int vlen,
- int flags, struct timespec *timeout)
+static inline int libc_recvmmsg(int sockfd, struct mmsghdr *msgvec,
+ unsigned int vlen, unsigned int flags, void *timeout)
{
- switch (tst_variant) {
- case 0:
- return tst_syscall(__NR_recvmmsg, sockfd, msgvec, vlen, flags,
- timeout);
- case 1:
#ifdef HAVE_RECVMMSG
- return recvmmsg(sockfd, msgvec, vlen, flags, timeout);
+ return recvmmsg(sockfd, msgvec, vlen, flags, timeout);
#else
- tst_brk(TCONF, "libc recvmmsg not present");
+ tst_brk(TCONF, "libc recvmmsg not present");
#endif
- }
-
- return -1;
}
-static void test_info(void)
+static inline int sys_recvmmsg(int sockfd, struct mmsghdr *msgvec,
+ unsigned int vlen, unsigned int flags, void *timeout)
{
- switch (tst_variant) {
- case 0:
- tst_res(TINFO, "Testing direct sendmmsg and recvmmsg syscalls");
- break;
- case 1:
- tst_res(TINFO, "Testing libc sendmmsg and recvmmsg syscalls");
- break;
- }
+ return tst_syscall(__NR_recvmmsg, sockfd, msgvec, vlen, flags, timeout);
}
-#define TEST_VARIANTS 2
+static inline int sys_recvmmsg64(int sockfd, struct mmsghdr *msgvec,
+ unsigned int vlen, unsigned int flags, void *timeout)
+{
+ return tst_syscall(__NR_recvmmsg_time64, sockfd, msgvec, vlen, flags,
+ timeout);
+}
#endif /* SENDMMSG_VAR__ */
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 12/17] syscalls/ppoll: Add support for time64 tests
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (10 preceding siblings ...)
2020-05-08 4:23 ` [LTP] [PATCH V2 11/17] syscalls/recvmmsg: " Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 13/17] syscalls/select6: " Viresh Kumar
` (4 subsequent siblings)
16 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
This adds support for time64 tests to the existing ppoll() syscall
tests.
Note that the O_EXCL flag is removed from SAFE_OPEN() calls as it made
the tests failed when run for the second variant as the file existed.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
testcases/kernel/syscalls/ppoll/ppoll01.c | 76 ++++++++++++++++++-----
1 file changed, 60 insertions(+), 16 deletions(-)
diff --git a/testcases/kernel/syscalls/ppoll/ppoll01.c b/testcases/kernel/syscalls/ppoll/ppoll01.c
index 2fadd0653948..eb8a14534e2e 100644
--- a/testcases/kernel/syscalls/ppoll/ppoll01.c
+++ b/testcases/kernel/syscalls/ppoll/ppoll01.c
@@ -21,6 +21,8 @@
#include "ltp_signal.h"
#include "tst_sig_proc.h"
#include "tst_test.h"
+#include "tst_timer.h"
+#include "lapi/abisize.h"
/* Older versions of glibc don't publish this constant's value. */
#ifndef POLLRDHUP
@@ -38,7 +40,7 @@ struct test_case {
unsigned int nfds; /* nfds ppoll parameter */
sigset_t *sigmask; /* sigmask ppoll parameter */
sigset_t *sigmask_cur; /* sigmask set for current process */
- struct timespec *ts; /* ts ppoll parameter */
+ struct tst_ts *ts; /* ts ppoll parameter */
struct pollfd *fds; /* fds ppoll parameter */
int sigint_count; /* if > 0, spawn process to send SIGINT */
/* 'count' times to current process */
@@ -60,14 +62,7 @@ static int fd1 = -1;
static sigset_t sigmask_empty, sigmask_sigint;
static struct pollfd fds_good[1], fds_already_closed[1];
-static struct timespec ts_short = {
- .tv_sec = 0,
- .tv_nsec = 200000000,
-};
-static struct timespec ts_long = {
- .tv_sec = 2,
- .tv_nsec = 0,
-};
+static struct tst_ts ts_short, ts_long;
/* Test cases
*
@@ -160,14 +155,57 @@ static struct test_case tcase[] = {
},
};
+static inline int libc_ppoll(struct pollfd *fds, nfds_t nfds, void *tmo_p,
+ const sigset_t *sigmask, size_t sigsetsize)
+{
+ return ppoll(fds, nfds, tmo_p, sigmask);
+}
+
+static inline int sys_ppoll(struct pollfd *fds, nfds_t nfds, void *tmo_p,
+ const sigset_t *sigmask, size_t sigsetsize)
+{
+ return tst_syscall(__NR_ppoll, fds, nfds, tmo_p, sigmask, sigsetsize);
+}
+
+static inline int sys_ppoll_time64(struct pollfd *fds, nfds_t nfds, void *tmo_p,
+ const sigset_t *sigmask, size_t sigsetsize)
+{
+ return tst_syscall(__NR_ppoll_time64, fds, nfds, tmo_p, sigmask,
+ sigsetsize);
+}
+
+static struct test_variants {
+ int (*ppoll)(struct pollfd *fds, nfds_t nfds, void *tmo_p,
+ const sigset_t *sigmask, size_t sigsetsize);
+
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .ppoll = libc_ppoll, .type = TST_LIBC_TIMESPEC, .desc = "vDSO or syscall with libc spec"},
+ { .ppoll = sys_ppoll, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+ { .ppoll = sys_ppoll, .type = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .ppoll = sys_ppoll, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_ppoll_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .ppoll = sys_ppoll_time64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
static void sighandler(int sig LTP_ATTRIBUTE_UNUSED)
{
}
static void setup(void)
{
+ struct test_variants *tv = &variants[tst_variant];
int fd2;
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
SAFE_SIGNAL(SIGINT, sighandler);
if (sigemptyset(&sigmask_empty) == -1)
@@ -177,18 +215,22 @@ static void setup(void)
if (sigaddset(&sigmask_sigint, SIGINT) == -1)
tst_brk(TBROK | TERRNO, "sigaddset");
- fd1 = SAFE_OPEN("testfile1", O_CREAT | O_EXCL | O_RDWR,
- S_IRUSR | S_IWUSR);
+ fd1 = SAFE_OPEN("testfile1", O_CREAT | O_RDWR, S_IRUSR | S_IWUSR);
fds_good[0].fd = fd1;
fds_good[0].events = POLLIN | POLLPRI | POLLOUT | POLLRDHUP;
fds_good[0].revents = 0;
- fd2 = SAFE_OPEN("testfile2", O_CREAT | O_EXCL | O_RDWR,
- S_IRUSR | S_IWUSR);
+ fd2 = SAFE_OPEN("testfile2", O_CREAT | O_RDWR, S_IRUSR | S_IWUSR);
fds_already_closed[0].fd = fd2;
fds_already_closed[0].events = POLLIN | POLLPRI | POLLOUT | POLLRDHUP;
fds_already_closed[0].revents = 0;
SAFE_CLOSE(fd2);
+
+ ts_short.type = ts_long.type = tv->type;
+ tst_ts_set_sec(&ts_short, 0);
+ tst_ts_set_nsec(&ts_short, 200000000);
+ tst_ts_set_sec(&ts_long, 2);
+ tst_ts_set_nsec(&ts_long, 0);
}
static void cleanup(void)
@@ -199,10 +241,11 @@ static void cleanup(void)
static void do_test(unsigned int i)
{
+ struct test_variants *tv = &variants[tst_variant];
pid_t pid = 0;
int sys_ret, sys_errno = 0, dummy;
struct test_case *tc = &tcase[i];
- struct timespec ts, *tsp = NULL;
+ struct tst_ts ts, *tsp = NULL;
if (tc->ts) {
memcpy(&ts, tc->ts, sizeof(ts));
@@ -223,8 +266,8 @@ static void do_test(unsigned int i)
/* test */
errno = 0;
- sys_ret = tst_syscall(__NR_ppoll, tc->fds, tc->nfds, tsp,
- tc->sigmask, SIGSETSIZE);
+ sys_ret = tv->ppoll(tc->fds, tc->nfds, tst_ts_get(tsp), tc->sigmask,
+ SIGSETSIZE);
sys_errno = errno;
/* cleanup */
@@ -261,6 +304,7 @@ static void do_test(unsigned int i)
static struct tst_test test = {
.tcnt = ARRAY_SIZE(tcase),
.test = do_test,
+ .test_variants = ARRAY_SIZE(variants),
.setup = setup,
.cleanup = cleanup,
.forks_child = 1,
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 13/17] syscalls/select6: Add support for time64 tests
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (11 preceding siblings ...)
2020-05-08 4:23 ` [LTP] [PATCH V2 12/17] syscalls/ppoll: " Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-08 7:21 ` Arnd Bergmann
2020-05-08 4:23 ` [LTP] [PATCH V2 14/17] syscalls/semop: Migrate to new test framework Viresh Kumar
` (3 subsequent siblings)
16 siblings, 1 reply; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
This adds support for time64 tests to the existing select6() syscall
tests.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
testcases/kernel/syscalls/select/select_var.h | 27 ++++++++++++++++---
1 file changed, 24 insertions(+), 3 deletions(-)
diff --git a/testcases/kernel/syscalls/select/select_var.h b/testcases/kernel/syscalls/select/select_var.h
index b19a1d1bf085..6e8d44088217 100644
--- a/testcases/kernel/syscalls/select/select_var.h
+++ b/testcases/kernel/syscalls/select/select_var.h
@@ -6,6 +6,8 @@
#define SELECT_VAR__
#include "lapi/syscalls.h"
+#include "tst_timer.h"
+#include "lapi/abisize.h"
struct compat_sel_arg_struct {
long _n;
@@ -47,7 +49,24 @@ static int do_select(int nfds, fd_set *readfds, fd_set *writefds, fd_set *except
timeout->tv_usec = ts.tv_nsec / 1000;
return ret;
}
- case 3:
+ case 3: {
+ int ret = 0;
+ struct __kernel_timespec ts = {
+ .tv_sec = timeout->tv_sec,
+ .tv_nsec = timeout->tv_usec * 1000,
+ };
+#if defined(TST_ABI64)
+ ret = tst_syscall(__NR_pselect6, nfds, readfds, writefds, exceptfds, &ts, NULL);
+#elif (__NR_clock_settime64 != __LTP__NR_INVALID_SYSCALL)
+ ret = tst_syscall(__NR_pselect6_time64, nfds, readfds, writefds, exceptfds, &ts, NULL);
+#else
+ tst_brk(TCONF, "__NR_pselect6 time64 variant not supported");
+#endif
+ timeout->tv_sec = ts.tv_sec;
+ timeout->tv_usec = ts.tv_nsec / 1000;
+ return ret;
+ }
+ case 4:
#ifdef __NR__newselect
return tst_syscall(__NR__newselect, nfds, readfds, writefds, exceptfds, timeout);
#else
@@ -55,7 +74,6 @@ static int do_select(int nfds, fd_set *readfds, fd_set *writefds, fd_set *except
#endif
break;
}
-
return -1;
}
@@ -72,11 +90,14 @@ static void select_info(void)
tst_res(TINFO, "Testing SYS_pselect6 syscall");
break;
case 3:
+ tst_res(TINFO, "Testing SYS_pselect6 time64 syscall");
+ break;
+ case 4:
tst_res(TINFO, "Testing SYS__newselect syscall");
break;
}
}
-#define TEST_VARIANTS 4
+#define TEST_VARIANTS 5
#endif /* SELECT_VAR__ */
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 14/17] syscalls/semop: Migrate to new test framework
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (12 preceding siblings ...)
2020-05-08 4:23 ` [LTP] [PATCH V2 13/17] syscalls/select6: " Viresh Kumar
@ 2020-05-08 4:23 ` Viresh Kumar
2020-05-08 4:24 ` [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version Viresh Kumar
` (2 subsequent siblings)
16 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:23 UTC (permalink / raw)
To: ltp
This migrates the semop tests to the new test framework.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
testcases/kernel/syscalls/ipc/semop/Makefile | 2 +-
testcases/kernel/syscalls/ipc/semop/semop01.c | 140 ++++----
testcases/kernel/syscalls/ipc/semop/semop02.c | 145 +++------
testcases/kernel/syscalls/ipc/semop/semop03.c | 150 +++------
testcases/kernel/syscalls/ipc/semop/semop04.c | 165 ++++------
testcases/kernel/syscalls/ipc/semop/semop05.c | 303 ++++++++----------
6 files changed, 347 insertions(+), 558 deletions(-)
diff --git a/testcases/kernel/syscalls/ipc/semop/Makefile b/testcases/kernel/syscalls/ipc/semop/Makefile
index 401e5e8cbb9f..5e4c695dd310 100644
--- a/testcases/kernel/syscalls/ipc/semop/Makefile
+++ b/testcases/kernel/syscalls/ipc/semop/Makefile
@@ -4,5 +4,5 @@
top_srcdir ?= ../../../../..
include $(top_srcdir)/include/mk/testcases.mk
-include $(abs_srcdir)/../Makefile.inc
+include $(abs_srcdir)/../Makefile2.inc
include $(top_srcdir)/include/mk/generic_leaf_target.mk
diff --git a/testcases/kernel/syscalls/ipc/semop/semop01.c b/testcases/kernel/syscalls/ipc/semop/semop01.c
index ea05c53eb919..bcb45fa69320 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop01.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop01.c
@@ -1,21 +1,5 @@
-/*
- *
- * Copyright (c) International Business Machines Corp., 2001
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Copyright (c) International Business Machines Corp., 2001 */
/*
* NAME
@@ -52,92 +36,73 @@
* none
*/
-#include "ipcsem.h"
+#include <stdlib.h>
+#include <sys/sem.h>
+#include "tst_test.h"
+#include "libnewipc.h"
+#include "lapi/semun.h"
#define NSEMS 4 /* the number of primitive semaphores to test */
-char *TCID = "semop01";
-int TST_TOTAL = 1;
-
-int sem_id_1 = -1; /* a semaphore set with read & alter permissions */
+static int sem_id = -1; /* a semaphore set with read & alter permissions */
+static key_t semkey;
-union semun get_arr;
-struct sembuf sops[PSEMS];
+static union semun get_arr;
+static struct sembuf sops[PSEMS];
-int main(int ac, char **av)
+static void run(void)
{
- int lc;
- int i;
+ union semun arr = { .val = 0 };
int fail = 0;
+ int i;
- tst_parse_opts(ac, av, NULL, NULL);
-
- setup();
-
- for (lc = 0; TEST_LOOPING(lc); lc++) {
- tst_count = 0;
-
- TEST(semop(sem_id_1, sops, NSEMS));
-
- if (TEST_RETURN == -1) {
- tst_resm(TFAIL, "%s call failed - errno = %d : %s",
- TCID, TEST_ERRNO, strerror(TEST_ERRNO));
- } else {
- /* get the values and make sure they */
- /* are the same as what was set */
- if (semctl(sem_id_1, 0, GETALL, get_arr) == -1) {
- tst_brkm(TBROK, cleanup,
- "semctl() failed in functional test");
- }
-
- for (i = 0; i < NSEMS; i++) {
- if (get_arr.array[i] != i * i) {
- fail = 1;
- }
- }
- if (fail)
- tst_resm(TFAIL,
- "semaphore values are wrong");
- else
- tst_resm(TPASS,
- "semaphore values are correct");
- }
+ TEST(semop(sem_id, sops, NSEMS));
+ if (TST_RET == -1) {
+ tst_res(TFAIL | TTERRNO, "semop() failed");
+ } else {
/*
- * clean up things in case we are looping
+ * Get the values and make sure they are the same as what was
+ * set
*/
- union semun set_arr;
- set_arr.val = 0;
+ if (semctl(sem_id, 0, GETALL, get_arr) == -1) {
+ tst_brk(TBROK | TERRNO, "semctl() failed in functional test");
+ }
+
for (i = 0; i < NSEMS; i++) {
- if (semctl(sem_id_1, i, SETVAL, set_arr) == -1) {
- tst_brkm(TBROK, cleanup, "semctl failed");
+ if (get_arr.array[i] != i * i) {
+ fail = 1;
}
}
+ if (fail)
+ tst_res(TFAIL | TERRNO, "semaphore values are wrong");
+ else
+ tst_res(TPASS, "semaphore values are correct");
}
- cleanup();
- tst_exit();
+ /*
+ * clean up things in case we are looping
+ */
+ for (i = 0; i < NSEMS; i++) {
+ if (semctl(sem_id, i, SETVAL, arr) == -1) {
+ tst_brk(TBROK | TERRNO, "semctl failed");
+ }
+ }
}
-void setup(void)
+static void setup(void)
{
int i;
- tst_sig(NOFORK, DEF_HANDLER, cleanup);
-
- TEST_PAUSE;
-
- tst_tmpdir();
-
get_arr.array = malloc(sizeof(unsigned short int) * PSEMS);
if (get_arr.array == NULL)
- tst_brkm(TBROK, cleanup, "malloc failed");
+ tst_brk(TBROK, "malloc failed");
- semkey = getipckey();
+ semkey = GETIPCKEY();
- sem_id_1 = semget(semkey, PSEMS, IPC_CREAT | IPC_EXCL | SEM_RA);
- if (sem_id_1 == -1)
- tst_brkm(TBROK, cleanup, "couldn't create semaphore in setup");
+ sem_id = semget(semkey, PSEMS, IPC_CREAT | IPC_EXCL | SEM_RA);
+ if (sem_id == -1)
+ tst_brk(TBROK | TERRNO, "couldn't create semaphore in setup");
for (i = 0; i < NSEMS; i++) {
sops[i].sem_num = i;
@@ -146,11 +111,20 @@ void setup(void)
}
}
-void cleanup(void)
+static void cleanup(void)
{
- rm_sema(sem_id_1);
+ union semun arr;
+ if (sem_id != -1) {
+ if (semctl(sem_id, 0, IPC_RMID, arr) == -1)
+ tst_res(TINFO, "WARNING: semaphore deletion failed.");
+ }
free(get_arr.array);
-
- tst_rmdir();
}
+
+static struct tst_test test = {
+ .test_all = run,
+ .setup = setup,
+ .cleanup = cleanup,
+ .needs_tmpdir = 1,
+};
diff --git a/testcases/kernel/syscalls/ipc/semop/semop02.c b/testcases/kernel/syscalls/ipc/semop/semop02.c
index f067229b1cf5..f24d284776a4 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop02.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop02.c
@@ -1,21 +1,5 @@
-/*
- *
- * Copyright (c) International Business Machines Corp., 2001
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Copyright (c) International Business Machines Corp., 2001 */
/*
* DESCRIPTION
@@ -31,30 +15,27 @@
#define _GNU_SOURCE
#include <pwd.h>
-#include "test.h"
-#include "safe_macros.h"
-#include "ipcsem.h"
-
-char *TCID = "semop02";
-
-static void semop_verify(int i);
-int sem_id_1 = -1; /* a semaphore set with read & alter permissions */
-int sem_id_2 = -1; /* a semaphore set without read & alter permissions */
-int bad_id = -1;
-
-struct sembuf s_buf[PSEMS];
-
-int badbuf = -1;
+#include <sys/ipc.h>
+#include <sys/sem.h>
+#include "tst_test.h"
+#include "libnewipc.h"
+#include "lapi/semun.h"
+
+static int sem_id_1 = -1; /* a semaphore set with read & alter permissions */
+static int sem_id_2 = -1; /* a semaphore set without read & alter permissions */
+static int bad_id = -1;
+static key_t semkey;
+static struct sembuf s_buf[PSEMS];
#define NSOPS 5 /* a resonable number of operations */
#define BIGOPS 1024 /* a value that is too large for the number */
/* of semop operations that are permitted */
-struct test_case_t {
+static struct test_case_t {
int *semid;
struct sembuf *t_sbuf;
unsigned t_ops;
int error;
-} TC[] = {
+} tc[] = {
{&sem_id_1, (struct sembuf *)&s_buf, BIGOPS, E2BIG},
{&sem_id_2, (struct sembuf *)&s_buf, NSOPS, EACCES},
{&sem_id_1, (struct sembuf *)-1, NSOPS, EFAULT},
@@ -63,29 +44,7 @@ struct test_case_t {
{&sem_id_1, (struct sembuf *)&s_buf, 1, ERANGE}
};
-int TST_TOTAL = ARRAY_SIZE(TC);
-
-int main(int ac, char **av)
-{
- int lc;
- int i;
-
- tst_parse_opts(ac, av, NULL, NULL);
-
- setup();
-
- for (lc = 0; TEST_LOOPING(lc); lc++) {
- tst_count = 0;
-
- for (i = 0; i < TST_TOTAL; i++)
- semop_verify(i);
- }
-
- cleanup();
- tst_exit();
-}
-
-void setup(void)
+static void setup(void)
{
char nobody_uid[] = "nobody";
struct passwd *ltpuser;
@@ -93,71 +52,73 @@ void setup(void)
struct seminfo ipc_buf;
union semun arr;
- tst_require_root();
-
- ltpuser = SAFE_GETPWNAM(NULL, nobody_uid);
- SAFE_SETUID(NULL, ltpuser->pw_uid);
-
- tst_sig(NOFORK, DEF_HANDLER, cleanup);
-
- TEST_PAUSE;
-
- tst_tmpdir();
+ ltpuser = SAFE_GETPWNAM(nobody_uid);
+ SAFE_SETUID(ltpuser->pw_uid);
/* get an IPC resource key */
- semkey = getipckey();
+ semkey = GETIPCKEY();
/* create a semaphore set with read and alter permissions */
sem_id_1 = semget(semkey, PSEMS, IPC_CREAT | IPC_EXCL | SEM_RA);
- if (sem_id_1 == -1) {
- tst_brkm(TBROK | TERRNO, cleanup,
- "couldn't create semaphore in setup");
- }
+ if (sem_id_1 == -1)
+ tst_brk(TBROK | TERRNO, "couldn't create semaphore in setup");
/* Get an new IPC resource key. */
- semkey2 = getipckey();
+ semkey2 = GETIPCKEY();
/* create a semaphore set without read and alter permissions */
sem_id_2 = semget(semkey2, PSEMS, IPC_CREAT | IPC_EXCL);
- if (sem_id_2 == -1) {
- tst_brkm(TBROK | TERRNO, cleanup,
- "couldn't create semaphore in setup");
- }
+ if (sem_id_2 == -1)
+ tst_brk(TBROK | TERRNO, "couldn't create semaphore in setup");
arr.__buf = &ipc_buf;
if (semctl(sem_id_1, 0, IPC_INFO, arr) == -1)
- tst_brkm(TBROK | TERRNO, cleanup, "semctl() IPC_INFO failed");
+ tst_brk(TBROK | TERRNO, "semctl() IPC_INFO failed");
/* for ERANGE errno test */
arr.val = 1;
s_buf[0].sem_op = ipc_buf.semvmx;
if (semctl(sem_id_1, 0, SETVAL, arr) == -1)
- tst_brkm(TBROK | TERRNO, cleanup, "semctl() SETVAL failed");
+ tst_brk(TBROK | TERRNO, "semctl() SETVAL failed");
}
-static void semop_verify(int i)
+static void run(unsigned int i)
{
- TEST(semop(*(TC[i].semid), TC[i].t_sbuf, TC[i].t_ops));
+ TEST(semop(*(tc[i].semid), tc[i].t_sbuf, tc[i].t_ops));
- if (TEST_RETURN != -1) {
- tst_resm(TFAIL, "call succeeded unexpectedly");
+ if (TST_RET != -1) {
+ tst_res(TFAIL | TTERRNO, "call succeeded unexpectedly");
return;
}
- if (TEST_ERRNO == TC[i].error) {
- tst_resm(TPASS | TTERRNO, "semop failed as expected");
+ if (TST_ERR == tc[i].error) {
+ tst_res(TPASS | TTERRNO, "semop failed as expected");
} else {
- tst_resm(TFAIL | TTERRNO,
+ tst_res(TFAIL | TTERRNO,
"semop failed unexpectedly; expected: "
- "%d - %s", TC[i].error, strerror(TC[i].error));
+ "%d - %s", tc[i].error, strerror(tc[i].error));
}
}
-void cleanup(void)
+static void cleanup(void)
{
- /* if they exist, remove the semaphore resources */
- rm_sema(sem_id_1);
- rm_sema(sem_id_2);
+ union semun arr;
- tst_rmdir();
+ if (sem_id_1 != -1) {
+ if (semctl(sem_id_1, 0, IPC_RMID, arr) == -1)
+ tst_res(TINFO, "WARNING: semaphore deletion failed.");
+ }
+ if (sem_id_2 != -1) {
+ if (semctl(sem_id_2, 0, IPC_RMID, arr) == -1)
+ tst_res(TINFO, "WARNING: semaphore deletion failed.");
+ }
}
+
+static struct tst_test test = {
+ .test = run,
+ .tcnt = ARRAY_SIZE(tc),
+ .setup = setup,
+ .cleanup = cleanup,
+ .needs_tmpdir = 1,
+ .needs_root = 1,
+};
diff --git a/testcases/kernel/syscalls/ipc/semop/semop03.c b/testcases/kernel/syscalls/ipc/semop/semop03.c
index a904e99439e2..4f5f78eb6d8d 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop03.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop03.c
@@ -1,21 +1,5 @@
-/*
- *
- * Copyright (c) International Business Machines Corp., 2001
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Copyright (c) International Business Machines Corp., 2001 */
/*
* NAME
@@ -50,109 +34,75 @@
* none
*/
-#include "ipcsem.h"
-
-char *TCID = "semop03";
-int TST_TOTAL = 2;
+#include <sys/sem.h>
+#include "tst_test.h"
+#include "libnewipc.h"
+#include "lapi/semun.h"
-int sem_id_1 = -1;
+static key_t semkey;
+static int sem_id = -1;
+static struct sembuf s_buf;
-struct sembuf s_buf;
+static int tc[] = { -1, PSEMS + 1 }; /* negative and too many "primitive" semas */
-int TC[] = { -1, PSEMS + 1 }; /* negative and too many "primitive" semas */
-
-int main(int ac, char **av)
+static void run(unsigned int i)
{
- int lc;
- int i;
-
- tst_parse_opts(ac, av, NULL, NULL);
-
- setup(); /* global setup */
-
/* initialize two fields in the sembuf structure here */
s_buf.sem_op = 1; /* add this value to struct sem.semval */
s_buf.sem_flg = SEM_UNDO; /* undo when process exits */
- /* The following loop checks looping state if -i option given */
-
- for (lc = 0; TEST_LOOPING(lc); lc++) {
- /* reset tst_count in case we are looping */
- tst_count = 0;
-
- for (i = 0; i < TST_TOTAL; i++) {
-
- /* initialize the last field in the sembuf */
- /* structure to the test dependent value */
- s_buf.sem_num = TC[i];
-
- /*
- * use the TEST macro to make the call
- */
+ /*
+ * initialize the last field in the sembuf structure to the test
+ * dependent value.
+ */
+ s_buf.sem_num = tc[i];
- TEST(semop(sem_id_1, &s_buf, 1));
+ /*
+ * use the TEST macro to make the call
+ */
- if (TEST_RETURN != -1) {
- tst_resm(TFAIL, "call succeeded unexpectedly");
- continue;
- }
+ TEST(semop(sem_id, &s_buf, 1));
- switch (TEST_ERRNO) {
- case EFBIG:
- tst_resm(TPASS, "expected failure - errno = "
- "%d : %s", TEST_ERRNO,
- strerror(TEST_ERRNO));
- break;
- default:
- tst_resm(TFAIL, "unexpected error - "
- "%d : %s", TEST_ERRNO,
- strerror(TEST_ERRNO));
- break;
- }
- }
+ if (TST_RET != -1) {
+ tst_res(TFAIL | TTERRNO, "call succeeded unexpectedly");
+ return;
}
- cleanup();
-
- tst_exit();
+ switch (TST_ERR) {
+ case EFBIG:
+ tst_res(TPASS | TTERRNO, "expected failure");
+ break;
+ default:
+ tst_res(TFAIL | TTERRNO, "unexpected failure");
+ break;
+ }
}
-/*
- * setup() - performs all the ONE TIME setup for this test.
- */
-void setup(void)
+static void setup(void)
{
-
- tst_sig(NOFORK, DEF_HANDLER, cleanup);
-
- TEST_PAUSE;
-
- /*
- * Create a temporary directory and cd into it.
- * This helps to ensure that a unique msgkey is created.
- * See ../lib/libipc.c for more information.
- */
- tst_tmpdir();
-
/* get an IPC resource key */
- semkey = getipckey();
+ semkey = GETIPCKEY();
/* create a semaphore with read and alter permissions */
- if ((sem_id_1 =
- semget(semkey, PSEMS, IPC_CREAT | IPC_EXCL | SEM_RA)) == -1) {
- tst_brkm(TBROK, cleanup, "couldn't create semaphore in setup");
- }
+ if ((sem_id = semget(semkey, PSEMS, IPC_CREAT | IPC_EXCL | SEM_RA)) ==
+ -1)
+ tst_brk(TBROK | TERRNO, "couldn't create semaphore in setup");
}
-/*
- * cleanup() - performs all the ONE TIME cleanup for this test@completion
- * or premature exit.
- */
-void cleanup(void)
+static void cleanup(void)
{
- /* if it exists, remove the semaphore resource */
- rm_sema(sem_id_1);
-
- tst_rmdir();
+ union semun arr;
+ if (sem_id != -1) {
+ if (semctl(sem_id, 0, IPC_RMID, arr) == -1)
+ tst_res(TINFO, "WARNING: semaphore deletion failed.");
+ }
}
+
+static struct tst_test test = {
+ .test = run,
+ .tcnt = ARRAY_SIZE(tc),
+ .setup = setup,
+ .cleanup = cleanup,
+ .needs_tmpdir = 1,
+};
diff --git a/testcases/kernel/syscalls/ipc/semop/semop04.c b/testcases/kernel/syscalls/ipc/semop/semop04.c
index bf425b54bb4c..0faf00a3585f 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop04.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop04.c
@@ -1,21 +1,5 @@
-/*
- *
- * Copyright (c) International Business Machines Corp., 2001
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Copyright (c) International Business Machines Corp., 2001 */
/*
* NAME
@@ -50,22 +34,24 @@
* none
*/
-#include "ipcsem.h"
-
-char *TCID = "semop04";
-int TST_TOTAL = 2;
+#include <sys/sem.h>
+#include "tst_test.h"
+#include "libnewipc.h"
+#include "lapi/semun.h"
-int sem_id_1 = -1;
+static int sem_id = -1;
+static int val; /* value for SETVAL */
-struct sembuf s_buf;
+static key_t semkey;
+static struct sembuf s_buf;
-struct test_case_t {
+static struct test_case_t {
union semun get_arr;
short op;
short flg;
short num;
int error;
-} TC[] = {
+} tc[] = {
/* EAGAIN sem_op = 0 */
{ {
1}, 0, IPC_NOWAIT, 2, EAGAIN},
@@ -74,104 +60,61 @@ struct test_case_t {
0}, -1, IPC_NOWAIT, 2, EAGAIN}
};
-int main(int ac, char **av)
+static void run(unsigned int i)
{
- int lc;
- int val; /* value for SETVAL */
-
- int i;
-
- tst_parse_opts(ac, av, NULL, NULL);
-
- setup(); /* global setup */
-
- /* The following loop checks looping state if -i option given */
-
- for (lc = 0; TEST_LOOPING(lc); lc++) {
- /* reset tst_count in case we are looping */
- tst_count = 0;
-
- val = 1;
- for (i = 0; i < TST_TOTAL; i++) {
-
- /* initialize the s_buf buffer */
- s_buf.sem_op = TC[i].op;
- s_buf.sem_flg = TC[i].flg;
- s_buf.sem_num = TC[i].num;
-
- /* initialize all the primitive semaphores */
- TC[i].get_arr.val = val--;
- if (semctl(sem_id_1, TC[i].num, SETVAL, TC[i].get_arr)
- == -1) {
- tst_brkm(TBROK, cleanup, "semctl() failed");
- }
-
- /*
- * make the call with the TEST macro
- */
-
- TEST(semop(sem_id_1, &s_buf, 1));
-
- if (TEST_RETURN != -1) {
- tst_resm(TFAIL, "call succeeded unexpectedly");
- continue;
- }
-
- if (TEST_ERRNO == TC[i].error) {
- tst_resm(TPASS,
- "expected failure - errno = %d"
- " : %s", TEST_ERRNO,
- strerror(TEST_ERRNO));
- } else {
- tst_resm(TFAIL, "unexpected error - "
- "%d : %s", TEST_ERRNO,
- strerror(TEST_ERRNO));
- }
- }
+ /* initialize the s_buf buffer */
+ s_buf.sem_op = tc[i].op;
+ s_buf.sem_flg = tc[i].flg;
+ s_buf.sem_num = tc[i].num;
+
+ /* initialize all the primitive semaphores */
+ tc[i].get_arr.val = val--;
+ if (semctl(sem_id, tc[i].num, SETVAL, tc[i].get_arr) == -1)
+ tst_brk(TBROK | TERRNO, "semctl() failed");
+
+ TEST(semop(sem_id, &s_buf, 1));
+ if (TST_RET != -1) {
+ tst_res(TFAIL, "call succeeded unexpectedly");
+ return;
}
- cleanup();
-
- tst_exit();
+ if (TST_ERR == tc[i].error)
+ tst_res(TPASS | TTERRNO, "expected failure");
+ else
+ tst_res(TFAIL | TTERRNO, "unexpected failure");
}
-/*
- * setup() - performs all the ONE TIME setup for this test.
- */
-void setup(void)
+static void setup(void)
{
+ val = 1;
- tst_sig(NOFORK, DEF_HANDLER, cleanup);
-
- TEST_PAUSE;
+ /* get an IPC resource key */
+ semkey = GETIPCKEY();
/*
- * Create a temporary directory and cd into it.
- * This helps to ensure that a unique msgkey is created.
- * See ../lib/libipc.c for more information.
+ * create a semaphore set with read and alter permissions and PSEMS
+ * "primitive" semaphores.
*/
- tst_tmpdir();
-
- /* get an IPC resource key */
- semkey = getipckey();
-
- /* create a semaphore set with read and alter permissions */
- /* and PSEMS "primitive" semaphores */
- if ((sem_id_1 =
- semget(semkey, PSEMS, IPC_CREAT | IPC_EXCL | SEM_RA)) == -1) {
- tst_brkm(TBROK, cleanup, "couldn't create semaphore in setup");
+ if ((sem_id = semget(semkey, PSEMS, IPC_CREAT | IPC_EXCL | SEM_RA)) ==
+ -1) {
+ tst_brk(TBROK | TERRNO, "couldn't create semaphore in setup");
}
}
-/*
- * cleanup() - performs all the ONE TIME cleanup for this test@completion
- * or premature exit.
- */
-void cleanup(void)
+static void cleanup(void)
{
- /* if it exists, remove the semaphore resource */
- rm_sema(sem_id_1);
-
- tst_rmdir();
+ union semun arr;
+ if (sem_id != -1) {
+ if (semctl(sem_id, 0, IPC_RMID, arr) == -1)
+ tst_res(TINFO, "WARNING: semaphore deletion failed.");
+ }
}
+
+static struct tst_test test = {
+ .test = run,
+ .tcnt = ARRAY_SIZE(tc),
+ .setup = setup,
+ .cleanup = cleanup,
+ .needs_tmpdir = 1,
+};
diff --git a/testcases/kernel/syscalls/ipc/semop/semop05.c b/testcases/kernel/syscalls/ipc/semop/semop05.c
index 6ed25014c7af..9e8e040b0b19 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop05.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop05.c
@@ -1,21 +1,5 @@
-/*
- *
- * Copyright (c) International Business Machines Corp., 2001
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program 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 General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Copyright (c) International Business Machines Corp., 2001 */
/*
* NAME
@@ -60,18 +44,18 @@
* none
*/
-#include "ipcsem.h"
-
+#include <stdio.h>
+#include <stdlib.h>
#include <sys/types.h>
#include <sys/wait.h>
-#include "safe_macros.h"
-
-char *TCID = "semop05";
-int TST_TOTAL = 4;
-
-int sem_id_1 = -1;
+#include <sys/sem.h>
+#include "tst_test.h"
+#include "libnewipc.h"
+#include "lapi/semun.h"
-struct sembuf s_buf;
+static key_t semkey;
+static int sem_id = -1;
+static struct sembuf s_buf;
struct test_case_t {
union semun semunptr;
@@ -79,7 +63,7 @@ struct test_case_t {
short flg;
short num;
int error;
-} TC[] = {
+} tc[] = {
/* EIRDM sem_op = 0 */
{ {
1}, 0, 0, 2, EIDRM},
@@ -96,189 +80,166 @@ struct test_case_t {
#ifdef UCLINUX
#define PIPE_NAME "semop05"
-void do_child_uclinux();
+static void do_child_uclinux();
static int i_uclinux;
#endif
-int main(int ac, char **av)
+static inline int process_state_wait2(pid_t pid, const char state)
{
- int lc;
- int i;
- pid_t pid;
- void do_child();
+ char proc_path[128], cur_state;
- tst_parse_opts(ac, av, NULL, NULL);
+ snprintf(proc_path, sizeof(proc_path), "/proc/%i/stat", pid);
-#ifdef UCLINUX
- maybe_run_child(&do_child_uclinux, "dd", &i_uclinux, &sem_id_1);
-#endif
-
- setup(); /* global setup */
-
- /* The following loop checks looping state if -i option given */
-
- for (lc = 0; TEST_LOOPING(lc); lc++) {
- /* reset tst_count in case we are looping */
- tst_count = 0;
-
- for (i = 0; i < TST_TOTAL; i++) {
-
- /* initialize the s_buf buffer */
- s_buf.sem_op = TC[i].op;
- s_buf.sem_flg = TC[i].flg;
- s_buf.sem_num = TC[i].num;
-
- /* initialize all of the primitive semaphores */
- if (semctl(sem_id_1, TC[i].num, SETVAL, TC[i].semunptr)
- == -1) {
- tst_brkm(TBROK, cleanup, "semctl() failed");
- }
-
- if ((pid = FORK_OR_VFORK()) == -1) {
- tst_brkm(TBROK, cleanup, "could not fork");
- }
-
- if (pid == 0) { /* child */
+ for (;;) {
+ FILE *f = fopen(proc_path, "r");
+ if (!f) {
+ tst_res(TFAIL, "Failed to open '%s': %s\n", proc_path,
+ strerror(errno));
+ return 1;
+ }
-#ifdef UCLINUX
- if (self_exec(av[0], "dd", i, sem_id_1) < 0) {
- tst_brkm(TBROK, cleanup,
- "could not self_exec");
- }
-#else
- do_child(i);
-#endif
- } else {
- TST_PROCESS_STATE_WAIT(cleanup, pid, 'S');
-
- /*
- * If we are testing for EIDRM then remove
- * the semaphore, else send a signal that
- * must be caught as we are testing for
- * EINTR.
- */
- if (TC[i].error == EIDRM) {
- /* remove the semaphore resource */
- rm_sema(sem_id_1);
- } else {
- SAFE_KILL(cleanup, pid, SIGHUP);
- }
-
- /* let the child carry on */
- waitpid(pid, NULL, 0);
- }
-
- /*
- * recreate the semaphore resource if needed
- */
- if (TC[i].error == EINTR) {
- continue;
- }
-
- if ((sem_id_1 = semget(semkey, PSEMS, IPC_CREAT |
- IPC_EXCL | SEM_RA)) == -1) {
- tst_brkm(TBROK, cleanup, "couldn't recreate "
- "semaphore");
- }
+ if (fscanf(f, "%*i %*s %c", &cur_state) != 1) {
+ fclose(f);
+ tst_res(TFAIL, "Failed to read '%s': %s\n", proc_path,
+ strerror(errno));
+ return 1;
}
- }
+ fclose(f);
- cleanup();
+ if (state == cur_state)
+ return 0;
- tst_exit();
+ usleep(10000);
+ }
}
-/*
- * do_child()
- */
-void do_child(int i)
+static void do_child(int i)
{
- /*
- * make the call with the TEST macro
- */
-
- TEST(semop(sem_id_1, &s_buf, 1));
-
- if (TEST_RETURN != -1) {
- tst_resm(TFAIL, "call succeeded when error expected");
+ TEST(semop(sem_id, &s_buf, 1));
+ if (TST_RET != -1) {
+ tst_res(TFAIL, "call succeeded when error expected");
exit(-1);
}
- if (TEST_ERRNO == TC[i].error) {
- tst_resm(TPASS, "expected failure - errno = %d"
- " : %s", TEST_ERRNO, strerror(TEST_ERRNO));
- } else {
- tst_resm(TFAIL, "unexpected error - "
- "%d : %s", TEST_ERRNO, strerror(TEST_ERRNO));
- }
+ if (TST_ERR == tc[i].error)
+ tst_res(TPASS | TTERRNO, "expected failure");
+ else
+ tst_res(TFAIL | TTERRNO, "unexpected failure");
exit(0);
}
-void sighandler(int sig)
+static void sighandler(int sig)
{
- if (sig == SIGHUP)
- return;
- else
- tst_brkm(TBROK, NULL, "unexpected signal %d received", sig);
+ if (sig != SIGHUP)
+ tst_brk(TBROK, "unexpected signal %d received", sig);
}
-#ifdef UCLINUX
-/*
- * do_child_uclinux() - capture signals, re-initialize s_buf then call do_child
- * with the appropriate argument
- */
-void do_child_uclinux(void)
+static void setup(void)
{
- int i = i_uclinux;
+ SAFE_SIGNAL(SIGHUP, sighandler);
- tst_sig(FORK, sighandler, cleanup);
+ /* get an IPC resource key */
+ semkey = GETIPCKEY();
- /* initialize the s_buf buffer */
- s_buf.sem_op = TC[i].op;
- s_buf.sem_flg = TC[i].flg;
- s_buf.sem_num = TC[i].num;
+ /*
+ * create a semaphore set with read and alter permissions and PSEMS
+ * "primitive" semaphores.
+ */
+ if ((sem_id = semget(semkey, PSEMS, IPC_CREAT | IPC_EXCL | SEM_RA)) ==
+ -1)
+ tst_brk(TBROK | TERRNO, "couldn't create semaphore in setup");
+}
- do_child(i);
+static void cleanup(void)
+{
+ union semun arr;
+
+ if (sem_id != -1) {
+ if (semctl(sem_id, 0, IPC_RMID, arr) == -1)
+ tst_res(TINFO, "WARNING: semaphore deletion failed.");
+ }
}
-#endif
-/*
- * setup() - performs all the ONE TIME setup for this test.
- */
-void setup(void)
+static void run(unsigned int i)
{
+ pid_t pid;
- tst_sig(FORK, sighandler, cleanup);
+#ifdef UCLINUX
+ maybe_run_child(&do_child_uclinux, "dd", &i_uclinux, &sem_id);
+#endif
+ /* initialize the s_buf buffer */
+ s_buf.sem_op = tc[i].op;
+ s_buf.sem_flg = tc[i].flg;
+ s_buf.sem_num = tc[i].num;
- TEST_PAUSE;
+ /* initialize all of the primitive semaphores */
+ if (semctl(sem_id, tc[i].num, SETVAL, tc[i].semunptr) == -1)
+ tst_brk(TBROK | TERRNO, "semctl() failed");
- /*
- * Create a temporary directory and cd into it.
- * This helps to ensure that a unique msgkey is created.
- * See ../lib/libipc.c for more information.
- */
- tst_tmpdir();
+ pid = SAFE_FORK();
- /* get an IPC resource key */
- semkey = getipckey();
+ if (pid == 0) { /* child */
+#ifdef UCLINUX
+ if (self_exec(av[0], "dd", i, sem_id) < 0)
+ tst_brk(TBROK, "could not self_exec");
+#else
+ do_child(i);
+#endif
+ } else {
+ process_state_wait2(pid, 'S');
+
+ /*
+ * If we are testing for EIDRM then remove
+ * the semaphore, else send a signal that
+ * must be caught as we are testing for
+ * EINTR.
+ */
+ if (tc[i].error == EIDRM) {
+ /* remove the semaphore resource */
+ cleanup();
+ } else {
+ SAFE_KILL(pid, SIGHUP);
+ }
- /* create a semaphore set with read and alter permissions */
- /* and PSEMS "primitive" semaphores */
- if ((sem_id_1 =
- semget(semkey, PSEMS, IPC_CREAT | IPC_EXCL | SEM_RA)) == -1) {
- tst_brkm(TBROK, cleanup, "couldn't create semaphore in setup");
+ /* let the child carry on */
+ waitpid(pid, NULL, 0);
}
+
+ /*
+ * recreate the semaphore resource if needed
+ */
+ if (tc[i].error == EINTR)
+ return;
+
+ if ((sem_id = semget(semkey, PSEMS, IPC_CREAT | IPC_EXCL | SEM_RA)) ==
+ -1)
+ tst_brk(TBROK | TERRNO, "couldn't recreate semaphore");
}
+#ifdef UCLINUX
/*
- * cleanup() - performs all the ONE TIME cleanup for this test@completion
- * or premature exit.
+ * do_child_uclinux() - capture signals, re-initialize s_buf then call do_child
+ * with the appropriate argument
*/
-void cleanup(void)
+static void do_child_uclinux(void)
{
- /* if it exists, remove the semaphore resource */
- rm_sema(sem_id_1);
+ int i = i_uclinux;
+
+ /* initialize the s_buf buffer */
+ s_buf.sem_op = tc[i].op;
+ s_buf.sem_flg = tc[i].flg;
+ s_buf.sem_num = tc[i].num;
- tst_rmdir();
+ do_child(i);
}
+#endif
+
+static struct tst_test test = {
+ .test = run,
+ .tcnt = ARRAY_SIZE(tc),
+ .setup = setup,
+ .cleanup = cleanup,
+ .needs_tmpdir = 1,
+ .forks_child = 1,
+};
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (13 preceding siblings ...)
2020-05-08 4:23 ` [LTP] [PATCH V2 14/17] syscalls/semop: Migrate to new test framework Viresh Kumar
@ 2020-05-08 4:24 ` Viresh Kumar
2020-05-08 7:18 ` Arnd Bergmann
2020-05-09 4:21 ` Li Wang
2020-05-08 4:24 ` [LTP] [PATCH V2 16/17] syscalls/utimensat: Migrate to new test framework Viresh Kumar
2020-05-08 4:24 ` [LTP] [PATCH V2 17/17] syscalls/utimensat: Add support for time64 tests Viresh Kumar
16 siblings, 2 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:24 UTC (permalink / raw)
To: ltp
This adds support for semtimedop() and its time64 variant to the
existing semop() syscall tests.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
testcases/kernel/syscalls/ipc/semop/semop.h | 52 +++++++++++++++++++
testcases/kernel/syscalls/ipc/semop/semop01.c | 11 +++-
testcases/kernel/syscalls/ipc/semop/semop02.c | 12 ++++-
testcases/kernel/syscalls/ipc/semop/semop03.c | 12 ++++-
testcases/kernel/syscalls/ipc/semop/semop04.c | 12 ++++-
testcases/kernel/syscalls/ipc/semop/semop05.c | 12 ++++-
6 files changed, 106 insertions(+), 5 deletions(-)
create mode 100644 testcases/kernel/syscalls/ipc/semop/semop.h
diff --git a/testcases/kernel/syscalls/ipc/semop/semop.h b/testcases/kernel/syscalls/ipc/semop/semop.h
new file mode 100644
index 000000000000..8d1245b65ec0
--- /dev/null
+++ b/testcases/kernel/syscalls/ipc/semop/semop.h
@@ -0,0 +1,52 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#ifndef SEMOP_VAR__
+#define SEMOP_VAR__
+
+#include <sys/sem.h>
+#include "tst_timer.h"
+#include "lapi/abisize.h"
+
+static inline int sys_semtimedop(int semid, struct sembuf *sops, size_t nsops,
+ void *timeout)
+{
+ return tst_syscall(__NR_semtimedop, semid, sops, nsops, timeout);
+}
+
+static inline int sys_semtimedop_time64(int semid, struct sembuf *sops,
+ size_t nsops, void *timeout)
+{
+ return tst_syscall(__NR_semtimedop_time64, semid, sops, nsops, timeout);
+}
+
+struct test_variants {
+ int (*semop)(int semid, struct sembuf *sops, size_t nsops);
+ int (*semtimedop)(int semid, struct sembuf *sops, size_t nsops, void *timeout);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+ { .semop = semop, .type = TST_LIBC_TIMESPEC, .desc = "semop: vDSO or syscall"},
+#if defined(TST_ABI32)
+ { .semtimedop = sys_semtimedop, .type = TST_LIBC_TIMESPEC, .desc = "semtimedop: syscall with libc spec"},
+ { .semtimedop = sys_semtimedop, .type = TST_KERN_OLD_TIMESPEC, .desc = "semtimedop: syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .semtimedop = sys_semtimedop, .type = TST_KERN_TIMESPEC, .desc = "semtimedop: syscall with kernel spec64"},
+#endif
+
+#if (__NR_semtimedop_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .semtimedop = sys_semtimedop_time64, .type = TST_KERN_TIMESPEC, .desc = "semtimedop: syscall time64 with kernel spec64"},
+#endif
+};
+
+static inline int call_semop(struct test_variants *tv, int semid,
+ struct sembuf *sops, size_t nsops, struct tst_ts *timeout)
+{
+ if (tv->semop)
+ return tv->semop(semid, sops, nsops);
+
+ return tv->semtimedop(semid, sops, nsops, tst_ts_get(timeout));
+}
+
+#endif /* SEMOP_VAR__ */
diff --git a/testcases/kernel/syscalls/ipc/semop/semop01.c b/testcases/kernel/syscalls/ipc/semop/semop01.c
index bcb45fa69320..a4924376e32c 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop01.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop01.c
@@ -41,6 +41,7 @@
#include "tst_test.h"
#include "libnewipc.h"
#include "lapi/semun.h"
+#include "semop.h"
#define NSEMS 4 /* the number of primitive semaphores to test */
@@ -52,11 +53,16 @@ static struct sembuf sops[PSEMS];
static void run(void)
{
+ struct test_variants *tv = &variants[tst_variant];
union semun arr = { .val = 0 };
+ struct tst_ts timeout;
int fail = 0;
int i;
- TEST(semop(sem_id, sops, NSEMS));
+ timeout.type = tv->type;
+ tst_ts_set_nsec(&timeout, 10000);
+
+ TEST(call_semop(tv, sem_id, sops, NSEMS, &timeout));
if (TST_RET == -1) {
tst_res(TFAIL | TTERRNO, "semop() failed");
@@ -94,6 +100,8 @@ static void setup(void)
{
int i;
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+
get_arr.array = malloc(sizeof(unsigned short int) * PSEMS);
if (get_arr.array == NULL)
tst_brk(TBROK, "malloc failed");
@@ -124,6 +132,7 @@ static void cleanup(void)
static struct tst_test test = {
.test_all = run,
+ .test_variants = ARRAY_SIZE(variants),
.setup = setup,
.cleanup = cleanup,
.needs_tmpdir = 1,
diff --git a/testcases/kernel/syscalls/ipc/semop/semop02.c b/testcases/kernel/syscalls/ipc/semop/semop02.c
index f24d284776a4..a51906340400 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop02.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop02.c
@@ -20,6 +20,7 @@
#include "tst_test.h"
#include "libnewipc.h"
#include "lapi/semun.h"
+#include "semop.h"
static int sem_id_1 = -1; /* a semaphore set with read & alter permissions */
static int sem_id_2 = -1; /* a semaphore set without read & alter permissions */
@@ -52,6 +53,8 @@ static void setup(void)
struct seminfo ipc_buf;
union semun arr;
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+
ltpuser = SAFE_GETPWNAM(nobody_uid);
SAFE_SETUID(ltpuser->pw_uid);
@@ -84,7 +87,13 @@ static void setup(void)
static void run(unsigned int i)
{
- TEST(semop(*(tc[i].semid), tc[i].t_sbuf, tc[i].t_ops));
+ struct test_variants *tv = &variants[tst_variant];
+ struct tst_ts timeout;
+
+ timeout.type = tv->type;
+ tst_ts_set_nsec(&timeout, 10000);
+
+ TEST(call_semop(tv, *(tc[i].semid), tc[i].t_sbuf, tc[i].t_ops, &timeout));
if (TST_RET != -1) {
tst_res(TFAIL | TTERRNO, "call succeeded unexpectedly");
@@ -117,6 +126,7 @@ static void cleanup(void)
static struct tst_test test = {
.test = run,
.tcnt = ARRAY_SIZE(tc),
+ .test_variants = ARRAY_SIZE(variants),
.setup = setup,
.cleanup = cleanup,
.needs_tmpdir = 1,
diff --git a/testcases/kernel/syscalls/ipc/semop/semop03.c b/testcases/kernel/syscalls/ipc/semop/semop03.c
index 4f5f78eb6d8d..d36194ce1f41 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop03.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop03.c
@@ -38,6 +38,7 @@
#include "tst_test.h"
#include "libnewipc.h"
#include "lapi/semun.h"
+#include "semop.h"
static key_t semkey;
static int sem_id = -1;
@@ -47,6 +48,12 @@ static int tc[] = { -1, PSEMS + 1 }; /* negative and too many "primitive" semas
static void run(unsigned int i)
{
+ struct test_variants *tv = &variants[tst_variant];
+ struct tst_ts timeout;
+
+ timeout.type = tv->type;
+ tst_ts_set_nsec(&timeout, 10000);
+
/* initialize two fields in the sembuf structure here */
s_buf.sem_op = 1; /* add this value to struct sem.semval */
s_buf.sem_flg = SEM_UNDO; /* undo when process exits */
@@ -61,7 +68,7 @@ static void run(unsigned int i)
* use the TEST macro to make the call
*/
- TEST(semop(sem_id, &s_buf, 1));
+ TEST(call_semop(tv, sem_id, &s_buf, 1, &timeout));
if (TST_RET != -1) {
tst_res(TFAIL | TTERRNO, "call succeeded unexpectedly");
@@ -80,6 +87,8 @@ static void run(unsigned int i)
static void setup(void)
{
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+
/* get an IPC resource key */
semkey = GETIPCKEY();
@@ -102,6 +111,7 @@ static void cleanup(void)
static struct tst_test test = {
.test = run,
.tcnt = ARRAY_SIZE(tc),
+ .test_variants = ARRAY_SIZE(variants),
.setup = setup,
.cleanup = cleanup,
.needs_tmpdir = 1,
diff --git a/testcases/kernel/syscalls/ipc/semop/semop04.c b/testcases/kernel/syscalls/ipc/semop/semop04.c
index 0faf00a3585f..033065e7bbfc 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop04.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop04.c
@@ -38,6 +38,7 @@
#include "tst_test.h"
#include "libnewipc.h"
#include "lapi/semun.h"
+#include "semop.h"
static int sem_id = -1;
static int val; /* value for SETVAL */
@@ -62,6 +63,12 @@ static struct test_case_t {
static void run(unsigned int i)
{
+ struct test_variants *tv = &variants[tst_variant];
+ struct tst_ts timeout;
+
+ timeout.type = tv->type;
+ tst_ts_set_nsec(&timeout, 10000);
+
/* initialize the s_buf buffer */
s_buf.sem_op = tc[i].op;
s_buf.sem_flg = tc[i].flg;
@@ -72,7 +79,7 @@ static void run(unsigned int i)
if (semctl(sem_id, tc[i].num, SETVAL, tc[i].get_arr) == -1)
tst_brk(TBROK | TERRNO, "semctl() failed");
- TEST(semop(sem_id, &s_buf, 1));
+ TEST(call_semop(tv, sem_id, &s_buf, 1, &timeout));
if (TST_RET != -1) {
tst_res(TFAIL, "call succeeded unexpectedly");
return;
@@ -86,6 +93,8 @@ static void run(unsigned int i)
static void setup(void)
{
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+
val = 1;
/* get an IPC resource key */
@@ -114,6 +123,7 @@ static void cleanup(void)
static struct tst_test test = {
.test = run,
.tcnt = ARRAY_SIZE(tc),
+ .test_variants = ARRAY_SIZE(variants),
.setup = setup,
.cleanup = cleanup,
.needs_tmpdir = 1,
diff --git a/testcases/kernel/syscalls/ipc/semop/semop05.c b/testcases/kernel/syscalls/ipc/semop/semop05.c
index 9e8e040b0b19..f892cd519ecf 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop05.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop05.c
@@ -52,6 +52,7 @@
#include "tst_test.h"
#include "libnewipc.h"
#include "lapi/semun.h"
+#include "semop.h"
static key_t semkey;
static int sem_id = -1;
@@ -115,7 +116,13 @@ static inline int process_state_wait2(pid_t pid, const char state)
static void do_child(int i)
{
- TEST(semop(sem_id, &s_buf, 1));
+ struct test_variants *tv = &variants[tst_variant];
+ struct tst_ts timeout;
+
+ timeout.type = tv->type;
+ tst_ts_set_nsec(&timeout, 10000);
+
+ TEST(call_semop(tv, sem_id, &s_buf, 1, &timeout));
if (TST_RET != -1) {
tst_res(TFAIL, "call succeeded when error expected");
exit(-1);
@@ -137,6 +144,8 @@ static void sighandler(int sig)
static void setup(void)
{
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+
SAFE_SIGNAL(SIGHUP, sighandler);
/* get an IPC resource key */
@@ -238,6 +247,7 @@ static void do_child_uclinux(void)
static struct tst_test test = {
.test = run,
.tcnt = ARRAY_SIZE(tc),
+ .test_variants = ARRAY_SIZE(variants),
.setup = setup,
.cleanup = cleanup,
.needs_tmpdir = 1,
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 16/17] syscalls/utimensat: Migrate to new test framework
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (14 preceding siblings ...)
2020-05-08 4:24 ` [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version Viresh Kumar
@ 2020-05-08 4:24 ` Viresh Kumar
2020-05-09 3:50 ` Li Wang
2020-05-11 9:33 ` Cyril Hrubis
2020-05-08 4:24 ` [LTP] [PATCH V2 17/17] syscalls/utimensat: Add support for time64 tests Viresh Kumar
16 siblings, 2 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:24 UTC (permalink / raw)
To: ltp
This gets rid of the test script and updates the utimensat01.c file to
handle most of the tests covered earlier (apart from testing the same
for files owned by root, as the test is run as root right now).
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
.../kernel/syscalls/utimensat/utimensat01.c | 411 ++++++--------
.../syscalls/utimensat/utimensat_tests.sh | 517 ------------------
2 files changed, 167 insertions(+), 761 deletions(-)
delete mode 100755 testcases/kernel/syscalls/utimensat/utimensat_tests.sh
diff --git a/testcases/kernel/syscalls/utimensat/utimensat01.c b/testcases/kernel/syscalls/utimensat/utimensat01.c
index 264ca1e15120..2a029ecb17e9 100644
--- a/testcases/kernel/syscalls/utimensat/utimensat01.c
+++ b/testcases/kernel/syscalls/utimensat/utimensat01.c
@@ -1,282 +1,205 @@
-/*************************************************************************************/
-/* */
-/* Copyright (C) 2008, Michael Kerrisk <mtk.manpages@gmail.com>, */
-/* Copyright (C) 2008, Linux Foundation */
-/* */
-/* This program is free software; you can redistribute it and/or modify */
-/* it under the terms of the GNU General Public License as published by */
-/* the Free Software Foundation; either version 2 of the License, or */
-/* (at your option) any later version. */
-/* */
-/* This program 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 General Public License for more details. */
-/* */
-/* You should have received a copy of the GNU General Public License */
-/* along with this program; if not, write to the Free Software */
-/* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */
-/*************************************************************************************/
-/* */
-/* File: utimnsat01.c */
-/* Description: A command-line interface for testing the utimensat() system call. */
-/* Author: Michael Kerrisk <mtk.manpages@gmail.com> */
-/* History: */
-/* 17 Mar 2008 Initial creation, */
-/* 31 May 2008 Reworked for easier test automation, */
-/* 2 June 2008 Renamed from t_utimensat.c to test_utimensat.c, */
-/* 05 June 2008 Submitted to LTP by Subrata Modak <subrata@linux.vnet.ibm.com> */
-/*************************************************************************************/
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2008 Michael Kerrisk <mtk.manpages@gmail.com>
+ * Copyright (c) 2008 Subrata Modak <subrata@linux.vnet.ibm.com>
+ * Copyright (c) 2020 Viresh Kumar <viresh.kumar@linaro.org>
+ *
+ * Basic utimnsat() test.
+ */
-#define _GNU_SOURCE
-#define _ATFILE_SOURCE
#include <stdio.h>
#include <time.h>
#include <errno.h>
#include <stdlib.h>
#include <unistd.h>
-#include <sys/syscall.h>
#include <fcntl.h>
#include <string.h>
#include <sys/stat.h>
-#include "test.h"
+#include "tst_test.h"
#include "lapi/syscalls.h"
-char *TCID = "utimensat01";
-int TST_TOTAL = 0;
-
-#define cleanup tst_exit
-
-/* We use EXIT_FAILURE for an expected failure from utimensat()
- (e.g., EACCES and EPERM), and one of the following for unexpected
- failures (i.e., something broke in our test setup). */
-
-#ifndef AT_FDCWD
-#define AT_FDCWD -100
-#endif
-#ifndef AT_SYMLINK_NOFOLLOW
-#define AT_SYMLINK_NOFOLLOW 0x100
-#endif
-
-#define EXIT_bad_usage 3
-#define EXIT_failed_syscall 3
-
-#define errExit(msg) do { perror(msg); exit(EXIT_failed_syscall); \
- } while (0)
-
-#define UTIME_NOW ((1l << 30) - 1l)
-#define UTIME_OMIT ((1l << 30) - 2l)
-
-static inline int
-utimensat_sc(int dirfd, const char *pathname,
- const struct timespec times[2], int flags)
+#define UTIME_NOW ((1l << 30) - 1l)
+#define UTIME_OMIT ((1l << 30) - 2l)
+
+#define TEST_FILE "test_file"
+#define TEST_DIR "test_dir"
+
+static void *bad_addr;
+
+struct mytime {
+ long access_tv_sec;
+ long access_tv_nsec;
+ long mod_tv_sec;
+ long mod_tv_nsec;
+};
+
+static struct mytime tnn = {0, UTIME_NOW, 0, UTIME_NOW}, *time_nn = &tnn;
+static struct mytime too = {0, UTIME_OMIT, 0, UTIME_OMIT}, *time_oo = &too;
+static struct mytime tno = {0, UTIME_NOW, 0, UTIME_OMIT}, *time_no = &tno;
+static struct mytime ton = {0, UTIME_OMIT, 0, UTIME_NOW}, *time_on = &ton;
+static struct mytime t11 = {1, 1, 1, 1}, *time_11 = &t11;
+
+struct test_case {
+ int dirfd;
+ char *pathname;
+ struct mytime **mytime;
+ int flags;
+ int oflags;
+ char *attr;
+ char *rattr;
+ int mode;
+ int exp_err;
+} tcase[] = {
+ /* Testing read-only file */
+ {AT_FDCWD, TEST_FILE, NULL, 0, O_RDONLY, NULL, NULL, 0400, 0},
+ {AT_FDCWD, TEST_FILE, &time_nn, 0, O_RDONLY, NULL, NULL, 0400, 0},
+ {AT_FDCWD, TEST_FILE, &time_oo, 0, O_RDONLY, NULL, NULL, 0400, 0},
+ {AT_FDCWD, TEST_FILE, &time_no, 0, O_RDONLY, NULL, NULL, 0400, 0},
+ {AT_FDCWD, TEST_FILE, &time_on, 0, O_RDONLY, NULL, NULL, 0400, 0},
+ {AT_FDCWD, TEST_FILE, &time_11, 0, O_RDONLY, NULL, NULL, 0400, 0},
+
+ /* Testing writable file */
+ {AT_FDCWD, TEST_FILE, NULL, 0, O_RDONLY, NULL, NULL, 0666, 0},
+ {AT_FDCWD, TEST_FILE, &time_nn, 0, O_RDONLY, NULL, NULL, 0666, 0},
+ {AT_FDCWD, TEST_FILE, &time_oo, 0, O_RDONLY, NULL, NULL, 0666, 0},
+ {AT_FDCWD, TEST_FILE, &time_no, 0, O_RDONLY, NULL, NULL, 0666, 0},
+ {AT_FDCWD, TEST_FILE, &time_on, 0, O_RDONLY, NULL, NULL, 0666, 0},
+ {AT_FDCWD, TEST_FILE, &time_11, 0, O_RDONLY, NULL, NULL, 0666, 0},
+
+ /* Testing append-only file */
+ {AT_FDCWD, TEST_FILE, NULL, 0, O_RDONLY, "+a", "-a", 0600, 0},
+ {AT_FDCWD, TEST_FILE, &time_nn, 0, O_RDONLY, "+a", "-a", 0600, 0},
+ {AT_FDCWD, TEST_FILE, &time_oo, 0, O_RDONLY, "+a", "-a", 0600, 0},
+ {AT_FDCWD, TEST_FILE, &time_no, 0, O_RDONLY, "+a", "-a", 0600, EPERM},
+ {AT_FDCWD, TEST_FILE, &time_on, 0, O_RDONLY, "+a", "-a", 0600, EPERM},
+ {AT_FDCWD, TEST_FILE, &time_11, 0, O_RDONLY, "+a", "-a", 0600, EPERM},
+
+ /* Testing immutable file */
+ {AT_FDCWD, TEST_FILE, NULL, 0, O_RDONLY, "+i", "-i", 0600, -1},
+ {AT_FDCWD, TEST_FILE, &time_nn, 0, O_RDONLY, "+i", "-i", 0600, -1},
+ {AT_FDCWD, TEST_FILE, &time_oo, 0, O_RDONLY, "+i", "-i", 0600, 0},
+ {AT_FDCWD, TEST_FILE, &time_no, 0, O_RDONLY, "+i", "-i", 0600, EPERM},
+ {AT_FDCWD, TEST_FILE, &time_on, 0, O_RDONLY, "+i", "-i", 0600, EPERM},
+ {AT_FDCWD, TEST_FILE, &time_11, 0, O_RDONLY, "+i", "-i", 0600, EPERM},
+
+ /* Testing immutable-append-only file */
+ {AT_FDCWD, TEST_FILE, NULL, 0, O_RDONLY, "+ai", "-ai", 0600, -1},
+ {AT_FDCWD, TEST_FILE, &time_nn, 0, O_RDONLY, "+ai", "-ai", 0600, -1},
+ {AT_FDCWD, TEST_FILE, &time_oo, 0, O_RDONLY, "+ai", "-ai", 0600, 0},
+ {AT_FDCWD, TEST_FILE, &time_no, 0, O_RDONLY, "+ai", "-ai", 0600, EPERM},
+ {AT_FDCWD, TEST_FILE, &time_on, 0, O_RDONLY, "+ai", "-ai", 0600, EPERM},
+ {AT_FDCWD, TEST_FILE, &time_11, 0, O_RDONLY, "+ai", "-ai", 0600, EPERM},
+
+ /* Other failure tests */
+ {AT_FDCWD, TEST_FILE, NULL, 0, O_RDONLY, NULL, NULL, 0400, EFAULT},
+ {AT_FDCWD, NULL, &time_nn, 0, O_RDONLY, NULL, NULL, 0400, EFAULT},
+ {-1, NULL, &time_nn, AT_SYMLINK_NOFOLLOW, O_RDONLY, NULL, NULL, 0400, EINVAL},
+ {-1, TEST_FILE, &time_nn, 0, O_RDONLY, NULL, NULL, 0400, ENOENT},
+};
+
+static inline int sys_utimensat(int dirfd, const char *pathname,
+ const struct timespec times[2], int flags)
{
- return ltp_syscall(__NR_utimensat, dirfd, pathname, times, flags);
+ return tst_syscall(__NR_utimensat, dirfd, pathname, times, flags);
}
-static void usageError(char *progName)
+static void setup(void)
{
- fprintf(stderr, "Usage: %s pathname [atime-sec "
- "atime-nsec mtime-sec mtime-nsec]\n\n", progName);
- fprintf(stderr, "Permitted options are:\n");
- fprintf(stderr, " [-d path] "
- "open a directory file descriptor"
- " (instead of using AT_FDCWD)\n");
- fprintf(stderr, " -q Quiet\n");
- fprintf(stderr, " -w Open directory file "
- "descriptor with O_RDWR|O_APPEND\n"
- " (instead of O_RDONLY)\n");
- fprintf(stderr, " -n Use AT_SYMLINK_NOFOLLOW\n");
- fprintf(stderr, "\n");
-
- fprintf(stderr, "pathname can be \"NULL\" to use NULL "
- "argument in call\n");
- fprintf(stderr, "\n");
+ bad_addr = tst_get_bad_addr(NULL);
+}
- fprintf(stderr, "Either nsec field can be\n");
- fprintf(stderr, " 'n' for UTIME_NOW\n");
- fprintf(stderr, " 'o' for UTIME_OMIT\n");
- fprintf(stderr, "\n");
+static void update_error(struct test_case *tc)
+{
+ if (tc->exp_err != -1)
+ return;
+
+ /*
+ * Starting with 4.8.0 operations on immutable files return EPERM
+ * instead of EACCES.
+ * This patch has also been merged to stable 4.4 with
+ * b3b4283 ("vfs: move permission checking into notify_change() for utimes(NULL)")
+ */
+ if (tst_kvercmp(4, 4, 27) < 0)
+ tc->exp_err = EACCES;
+ else
+ tc->exp_err = EPERM;
+}
- fprintf(stderr, "If the time fields are omitted, "
- "then a NULL 'times' argument is used\n");
- fprintf(stderr, "\n");
+static int run_command(char *command, char *option, char *file)
+{
+ const char *const cmd[] = {command, option, file, NULL};
- exit(EXIT_bad_usage);
+ return tst_cmd(cmd, NULL, NULL, TST_CMD_PASS_RETVAL);
}
-int main(int argc, char *argv[])
+static void run(unsigned int i)
{
- int flags, dirfd, opt, oflag;
+ struct test_case *tc = &tcase[i];
struct timespec ts[2];
- struct timespec *tsp;
- char *pathname, *dirfdPath;
+ void *tsp = NULL;
+ char *pathname = NULL;
+ int fd = AT_FDCWD;
struct stat sb;
- int verbose;
-
- /* Command-line argument parsing */
-
- flags = 0;
- verbose = 1;
- dirfd = AT_FDCWD;
- dirfdPath = NULL;
- oflag = O_RDONLY;
-
- while ((opt = getopt(argc, argv, "d:nqw")) != -1) {
- switch (opt) {
- case 'd':
- dirfdPath = optarg;
- break;
-
- case 'n':
- flags |= AT_SYMLINK_NOFOLLOW;
- if (verbose)
- printf("Not following symbolic links\n");
- break;
-
- case 'q':
- verbose = 0;
- break;
-
- case 'w':
- oflag = O_RDWR | O_APPEND;
- break;
-
- default:
- usageError(argv[0]);
- }
- }
-
- if ((optind + 5 != argc) && (optind + 1 != argc))
- usageError(argv[0]);
-
- if (dirfdPath != NULL) {
- dirfd = open(dirfdPath, oflag);
- if (dirfd == -1)
- errExit("open");
-
- if (verbose) {
- printf("Opened dirfd %d", oflag);
- if ((oflag & O_ACCMODE) == O_RDWR)
- printf(" O_RDWR");
- if (oflag & O_APPEND)
- printf(" O_APPEND");
- printf(": %s\n", dirfdPath);
- }
- }
- pathname = (strcmp(argv[optind], "NULL") == 0) ? NULL : argv[optind];
+ SAFE_MKDIR(TEST_DIR, 0700);
+ SAFE_TOUCH(TEST_FILE, 0666, NULL);
+ update_error(tc);
- /* Either, we get no values for 'times' fields, in which case
- we give a NULL pointer to utimensat(), or we get four values,
- for secs+nsecs for each of atime and mtime. The special
- values 'n' and 'o' can be used for tv_nsec settings of
- UTIME_NOW and UTIME_OMIT, respectively. */
+ if (tc->dirfd != AT_FDCWD)
+ fd = SAFE_OPEN(TEST_DIR, tc->oflags);
- if (argc == optind + 1) {
- tsp = NULL;
-
- } else {
- ts[0].tv_sec = atoi(argv[optind + 1]);
- if (argv[optind + 2][0] == 'n') {
- ts[0].tv_nsec = UTIME_NOW;
- } else if (argv[optind + 2][0] == 'o') {
- ts[0].tv_nsec = UTIME_OMIT;
- } else {
- ts[0].tv_nsec = atoi(argv[optind + 2]);
- }
-
- ts[1].tv_sec = atoi(argv[optind + 3]);
- if (argv[optind + 4][0] == 'n') {
- ts[1].tv_nsec = UTIME_NOW;
- } else if (argv[optind + 4][0] == 'o') {
- ts[1].tv_nsec = UTIME_OMIT;
- } else {
- ts[1].tv_nsec = atoi(argv[optind + 4]);
- }
+ if (tc->mytime) {
+ struct mytime *mytime = *tc->mytime;
+ ts[0].tv_sec = mytime->access_tv_sec;
+ ts[0].tv_nsec = mytime->access_tv_nsec;
+ ts[1].tv_sec = mytime->mod_tv_sec;
+ ts[1].tv_nsec = mytime->mod_tv_nsec;
tsp = ts;
+ } else if (tc->exp_err == EFAULT) {
+ tsp = bad_addr;
}
- /* For testing purposes, it may have been useful to run this program
- as set-user-ID-root so that a directory file descriptor could be
- opened as root. (This allows us to obtain a file descriptor even
- if normal user doesn't have permissions on the file.) Now we
- reset to the real UID before making the utimensat() call, so that
- the permission checking for the utimensat() call is performed
- under that UID. */
-
- if (geteuid() == 0) {
- uid_t u;
-
- u = getuid();
-
- if (verbose)
- printf("Resetting UIDs to %ld\n", (long)u);
-
- if (setresuid(u, u, u) == -1)
- errExit("setresuid");
- }
-
- /* Display information allowing user to verify arguments for call */
-
- if (verbose) {
- printf("dirfd is %d\n", dirfd);
- printf("pathname is %s\n", pathname);
- printf("tsp is %p", tsp);
- if (tsp != NULL) {
- printf("; struct = { %ld, %ld } { %ld, %ld }",
- (long)tsp[0].tv_sec, (long)tsp[0].tv_nsec,
- (long)tsp[1].tv_sec, (long)tsp[1].tv_nsec);
- }
- printf("\n");
- printf("flags is %d\n", flags);
+ if (tc->pathname) {
+ pathname = tc->pathname;
+ SAFE_CHMOD(tc->pathname, tc->mode);
+ if (tc->attr)
+ run_command("chattr", tc->attr, tc->pathname);
+ } else if (tc->exp_err == EFAULT) {
+ pathname = bad_addr;
}
- /* Make the call and see what happened */
-
- if (utimensat_sc(dirfd, pathname, tsp, flags) == -1) {
- if (errno == EPERM) {
- if (verbose)
- printf("utimensat() failed with EPERM\n");
- else
- printf("EPERM\n");
- exit(EXIT_FAILURE);
-
- } else if (errno == EACCES) {
- if (verbose)
- printf("utimensat() failed with EACCES\n");
- else
- printf("EACCES\n");
- exit(EXIT_FAILURE);
-
- } else if (errno == EINVAL) {
- if (verbose)
- printf("utimensat() failed with EINVAL\n");
- else
- printf("EINVAL\n");
- exit(EXIT_FAILURE);
-
- } else { /* Unexpected failure case from utimensat() */
- errExit("utimensat");
+ TEST(sys_utimensat(fd, pathname, tsp, tc->flags));
+ if (TST_RET) {
+ if (!tc->exp_err) {
+ tst_res(TFAIL | TTERRNO, "%2d: utimensat() failed", i);
+ } else if (tc->exp_err == TST_ERR) {
+ tst_res(TPASS | TTERRNO, "%2d: utimensat() failed expectedly", i);
+ } else {
+ tst_res(TFAIL | TTERRNO, "%2d: utimensat() failed with incorrect error, expected %s",
+ i, tst_strerrno(tc->exp_err));
}
+ } else if (tc->exp_err) {
+ tst_res(TFAIL, "%2d: utimensat() passed unexpectedly", i);
+ } else {
+ TEST(stat(tc->pathname ? tc->pathname : TEST_DIR, &sb));
+ tst_res(TPASS, "%2d: utimensat() passed: access time: %lu %lu", i,
+ sb.st_atime, sb.st_mtime);
}
- if (verbose)
- printf("utimensat() succeeded\n");
-
- if (stat((pathname != NULL) ? pathname : dirfdPath, &sb) == -1)
- errExit("stat");
-
- if (verbose) {
- printf("Last file access: %s", ctime(&sb.st_atime));
- printf("Last file modification: %s", ctime(&sb.st_mtime));
- printf("Last status change: %s", ctime(&sb.st_ctime));
+ if (tc->rattr != NULL)
+ run_command("chattr", tc->rattr, tc->pathname);
- } else {
- printf("SUCCESS %ld %ld\n", (long)sb.st_atime,
- (long)sb.st_mtime);
- }
+ if (fd != AT_FDCWD)
+ SAFE_CLOSE(fd);
- exit(EXIT_SUCCESS);
+ SAFE_RMDIR(TEST_DIR);
}
+
+static struct tst_test test = {
+ .test = run,
+ .tcnt = ARRAY_SIZE(tcase),
+ .setup = setup,
+ .needs_root = 1,
+ .needs_tmpdir = 1,
+};
diff --git a/testcases/kernel/syscalls/utimensat/utimensat_tests.sh b/testcases/kernel/syscalls/utimensat/utimensat_tests.sh
deleted file mode 100755
index 46a0cc64bf90..000000000000
--- a/testcases/kernel/syscalls/utimensat/utimensat_tests.sh
+++ /dev/null
@@ -1,517 +0,0 @@
-#!/bin/sh
-#
-# Automated tests for utimensat()
-#
-# Copyright (C) 2008, Linux Foundation
-# Written by Michael Kerrisk <mtk.manpages@gmail.com>
-# Licensed under GPLv2 or later
-#
-# Not (yet) included in this automated test set:
-# * AT_SYMLINK_NOFOLLOW in flags: If pathname specifies a symbolic link,
-# then update the timestamps of the link, rather than the file to which
-# it refers.
-# * Setting of nanosecond components of timestamps (support for
-# nanosecond timestamps is file-system-dependent)
-# * "Updated file timestamps are set to the greatest value supported
-# by the file system that is not greater than the specified time."
-# (i.e., if we set timestamp to {0, 999999999}, then the setting
-# is rounded down, rather than up, to unit of timestamp resolution.
-# * Privileged processes should be able to bypass permission checks.
-# (except when file is marked with the "Immutable" EFA).
-
-#=====================================================================
-
-export TCID=utimensat01
-export TST_TOTAL=99
-export TST_COUNT=0
-. test.sh
-
-if tst_kvcmp -lt "2.6.22"; then
- tst_brkm TCONF "System kernel version is less than 2.6.22,cannot execute test"
-fi
-
-# Starting with 4.8.0 operations on immutable files return EPERM instead of
-# EACCES.
-# This patch has also been merged to stable 4.4 with
-# b3b4283 ("vfs: move permission checking into notify_change() for utimes(NULL)")
-if tst_kvcmp -ge "4.4.27" -a -lt "4.5.0"; then
- imaccess=EPERM
-elif tst_kvcmp -lt "4.4.27"; then
- imaccess=EACCES
-else
- imaccess=EPERM
-fi
-
-
-RESULT_FILE=$TMPDIR/utimensat.result
-
-TEST_DIR=$TMPDIR/utimensat_tests
-FILE=$TEST_DIR/utimensat.test_file
-
-TEST_PROG=utimensat01
-
-if [ ! -f $LTPROOT/testcases/bin/$TEST_PROG ]; then
- tst_brkm TBROK "$LTPROOT/testcases/bin/$TEST_PROG is missing (please check install)"
-fi
-
-# Summary counters of all test results
-
-test_num=0
-failed_cnt=0
-passed_cnt=0
-failed_list=""
-
-#=====================================================================
-
-setup_file()
-{
-# $1 is test file pathname
-# $2 is owner for test file (chown(1))
-# $3 is permissions for test file (chmod(1))
-# $4 is "ext2" extended file attributes for test file (chattr(1))
-
- FILE=$1
-
- # Make sure any old version of file is deleted
-
- if test -e $FILE; then
- sudo $s_arg chattr -ai $FILE || return $?
- sudo $s_arg rm -f $FILE || return $?
- fi
-
- # Create file and make atime and mtime zero.
-
- sudo $s_arg -u $test_user touch $FILE || return $?
- if ! $TEST_PROG -q $FILE 0 0 0 0 > $RESULT_FILE; then
- echo "Failed to set up test file $FILE" 1>&2
- exit 1
- fi
-
- read res atime mtime < $RESULT_FILE
- if test "X$res" != "XSUCCESS" ||
- test $atime -ne 0 || test $mtime != 0; then
- echo "Failed to set correct times on test file $FILE" 1>&2
- exit 1
- fi
-
- # Set owner, permissions, and EFAs for file.
-
- if test -n "$2"; then
- sudo $s_arg chown $2 $FILE || return $?
- fi
-
- sudo $s_arg chmod $3 $FILE || return $?
-
- if test -n "$4"; then
- sudo $s_arg chattr $4 $FILE || return $?
- fi
-
- # Display file setup, for visual verification
-
- ls -l $FILE | awk '{ printf "Owner=%s; perms=%s; ", $3, $1}'
- if ! sudo $s_arg lsattr -l $FILE | sed 's/, /,/g' | awk '{print "EFAs=" $2}'
- then
- return $?
- fi
-
-}
-
-test_failed()
-{
- tst_resm TFAIL "FAILED test $test_num"
-
- failed_cnt=$(expr $failed_cnt + 1)
- failed_list="$failed_list $test_num"
-}
-
-check_result()
-{
- STATUS=$1 # Exit status from test program
- EXPECTED_RESULT=$2 # SUCCESS / EACCES / EPERM / EINVAL
- EXPECT_ATIME_CHANGED=$3 # Should be 'y' or 'n' (only for SUCCESS)
- EXPECT_MTIME_CHANGED=$4 # Should be 'y' or 'n' (only for SUCCESS)
-
- test_num=$(expr $test_num + 1)
-
- # If our test setup failed, stop immediately
-
- if test $STATUS -gt 1; then
- echo "FAILED (bad test setup)"
- exit 1
- fi
-
- read res atime mtime < $RESULT_FILE
-
- echo "EXPECTED: $EXPECTED_RESULT $EXPECT_ATIME_CHANGED "\
- "$EXPECT_MTIME_CHANGED"
- echo "RESULT: $res $atime $mtime"
-
- if test "$res" != "$EXPECTED_RESULT"; then
- test_failed
- return
- fi
-
- passed=1
-
- # If the test program exited successfully, then check that atime and
- # and mtime were updated / not updated, as expected.
-
- if test $EXPECTED_RESULT = "SUCCESS"; then
- if test $EXPECT_ATIME_CHANGED = "y"; then
- if test $atime -eq 0; then
- echo "atime should have changed, but did not"
- passed=0
- fi
- else
- if test $atime -ne 0; then
- echo "atime should not have changed, but did"
- passed=0
- fi
- fi
-
- if test $EXPECT_MTIME_CHANGED = "y"; then
- if test $mtime -eq 0; then
- echo "mtime should have changed, but did not"
- passed=0
- fi
- else
- if test $mtime -ne 0; then
- echo "mtime should not have changed, but did"
- passed=0
- fi
- fi
-
- if test $passed -eq 0; then
- test_failed
- return
- fi
- fi
-
- passed_cnt=$(expr $passed_cnt + 1)
- tst_resm TPASS "PASSED test $test_num"
-}
-
-run_test()
-{
- # By default, we do three types of test:
- # a) pathname (pathname != NULL)
- # b) readable file descriptor (pathname == NULL, dirfd opened O_RDONLY)
- # c) writable file descriptor (pathname == NULL, dirfd opened O_RDWR).
- # For this case we also include O_APPEND in open flags, since that
- # is needed if testing with a file that has the Append-only
- # attribute enabled.
-
- # -R says don't do tests with readable file descriptor
- # -W says don't do tests with writable file descriptor
-
- OPTIND=1
-
- do_read_fd_test=1
- do_write_fd_test=1
- while getopts "RW" opt; do
- case "$opt" in
- R) do_read_fd_test=0
- ;;
- W) do_write_fd_test=0
- ;;
- *) echo "run_test: bad usage"
- exit 1
- ;;
- esac
- done
- shift `expr $OPTIND - 1`
-
- echo "Pathname test"
- setup_file $FILE "$1" "$2" "$3"
- cp $LTPROOT/testcases/bin/$TEST_PROG ./
- CMD="./$TEST_PROG -q $FILE $4"
- echo "$CMD"
- sudo $s_arg -u $test_user $CMD > $RESULT_FILE
- check_result $? $5 $6 $7
- echo
-
- if test $do_read_fd_test -ne 0; then
- echo "Readable file descriptor (futimens(3)) test"
- setup_file $FILE "$1" "$2" "$3"
- CMD="./$TEST_PROG -q -d $FILE NULL $4"
- echo "$CMD"
- sudo $s_arg -u $test_user $CMD > $RESULT_FILE
- check_result $? $5 $6 $7
- echo
- fi
-
- # Can't do the writable file descriptor test for immutable files
- # (even root can't open an immutable file for writing)
-
- if test $do_write_fd_test -ne 0; then
- echo "Writable file descriptor (futimens(3)) test"
- setup_file $FILE "$1" "$2" "$3"
- CMD="./$TEST_PROG -q -w -d $FILE NULL $4"
- echo "$CMD"
- sudo $s_arg -u $test_user $CMD > $RESULT_FILE
- check_result $? $5 $6 $7
- echo
- fi
-
- sudo $s_arg chattr -ai $FILE
- sudo $s_arg rm -f $FILE
-}
-
-#=====================================================================
-
-# Since some automated testing systems have no tty while testing,
-# comment this line in /etc/sudoers to avoid the error message:
-# `sudo: sorry, you must have a tty to run sudo'
-# Use trap to restore this line after program terminates.
-sudoers=/etc/sudoers
-if [ ! -r $sudoers ]; then
- tst_brkm TBROK "can't read $sudoers"
-fi
-pattern="[[:space:]]*Defaults[[:space:]]*requiretty.*"
-if grep -q "^${pattern}" $sudoers; then
- tst_resm TINFO "Comment requiretty in $sudoers for automated testing systems"
- if ! sed -r -i.$$ -e "s/^($pattern)/#\1/" $sudoers; then
- tst_brkm TBROK "failed to mangle $sudoers properly"
- fi
- trap 'trap "" EXIT; restore_sudoers' EXIT
-fi
-
-restore_sudoers()
-{
- tst_resm TINFO "Restore requiretty in $sudoers"
- mv /etc/sudoers.$$ /etc/sudoers
-}
-
-test_user=nobody
-echo "test sudo for -n option, non-interactive"
-if sudo -h | grep -q -- -n; then
- s_arg="-n"
- echo "sudo supports -n"
-else
- s_arg=
- echo "sudo does not support -n"
-fi
-
-if ! sudo $s_arg true; then
- tst_brkm TBROK "sudo cannot be run by user non-interactively"
-fi
-if test ! -f $sudoers
-then
- echo "root ALL=(ALL) ALL" > $sudoers || exit
- chmod 440 $sudoers
- trap 'trap "" EXIT; nuke_sudoers' EXIT
-fi
-
-nuke_sudoers()
-{
- sudo rm -f $sudoers
-}
-
-sudo $s_arg -u $test_user mkdir -p $TEST_DIR
-
-# Make sure chattr command is supported
-touch $TEST_DIR/tmp_file
-chattr +a $TEST_DIR/tmp_file
-if [ $? -ne 0 ] ; then
- rm -rf $TEST_DIR
- tst_brkm TCONF "chattr not supported"
-fi
-chattr -a $TEST_DIR/tmp_file
-
-cd $TEST_DIR
-chown root $LTPROOT/testcases/bin/$TEST_PROG
-chmod ugo+x,u+s $LTPROOT/testcases/bin/$TEST_PROG
-
-#=====================================================================
-
-
-echo "============================================================"
-
-echo
-echo "Testing read-only file, owned by self"
-echo
-
-echo "***** Testing times==NULL case *****"
-run_test -W "" 400 "" "" SUCCESS y y
-
-echo "***** Testing times=={ UTIME_NOW, UTIME_NOW } case *****"
-run_test -W "" 400 "" "0 n 0 n" SUCCESS y y
-
-echo "***** Testing times=={ UTIME_OMIT, UTIME_OMIT } case *****"
-run_test -W "" 400 "" "0 o 0 o" SUCCESS n n
-
-echo "***** Testing times=={ UTIME_NOW, UTIME_OMIT } case *****"
-run_test -W "" 400 "" "0 n 0 o" SUCCESS y n
-
-echo "***** Testing times=={ UTIME_OMIT, UTIME_NOW } case *****"
-run_test -W "" 400 "" "0 o 0 n" SUCCESS n y
-
-echo "***** Testing times=={ x, y } case *****"
-run_test -W "" 400 "" "1 1 1 1" SUCCESS y y
-
-echo "============================================================"
-
-echo
-echo "Testing read-only file, not owned by self"
-echo
-
-echo "***** Testing times==NULL case *****"
-run_test -RW root 400 "" "" EACCES
-
-echo "***** Testing times=={ UTIME_NOW, UTIME_NOW } case *****"
-run_test -RW root 400 "" "0 n 0 n" EACCES
-
-echo "***** Testing times=={ UTIME_OMIT, UTIME_OMIT } case *****"
-run_test -RW root 400 "" "0 o 0 o" SUCCESS n n
-
-echo "***** Testing times=={ UTIME_NOW, UTIME_OMIT } case *****"
-run_test -RW root 400 "" "0 n 0 o" EPERM
-
-echo "***** Testing times=={ UTIME_OMIT, UTIME_NOW } case *****"
-run_test -RW root 400 "" "0 o 0 n" EPERM
-
-echo "***** Testing times=={ x, y } case *****"
-run_test -RW root 400 "" "1 1 1 1" EPERM
-
-echo "============================================================"
-
-echo
-echo "Testing writable file, not owned by self"
-echo
-
-echo "***** Testing times==NULL case *****"
-run_test root 666 "" "" SUCCESS y y
-
-echo "***** Testing times=={ UTIME_NOW, UTIME_NOW } case *****"
-run_test root 666 "" "0 n 0 n" SUCCESS y y
-
-echo "***** Testing times=={ UTIME_OMIT, UTIME_OMIT } case *****"
-run_test root 666 "" "0 o 0 o" SUCCESS n n
-
-echo "***** Testing times=={ UTIME_NOW, UTIME_OMIT } case *****"
-run_test root 666 "" "0 n 0 o" EPERM
-
-echo "***** Testing times=={ UTIME_OMIT, UTIME_NOW } case *****"
-run_test root 666 "" "0 o 0 n" EPERM
-
-echo "***** Testing times=={ x, y } case *****"
-run_test root 666 "" "1 1 1 1" EPERM
-
-echo "============================================================"
-
-echo
-echo "Testing append-only file, owned by self"
-echo
-
-echo "***** Testing times==NULL case *****"
-run_test "" 600 "+a" "" SUCCESS y y
-
-echo "***** Testing times=={ UTIME_NOW, UTIME_NOW } case *****"
-run_test "" 600 "+a" "0 n 0 n" SUCCESS y y
-
-echo "***** Testing times=={ UTIME_OMIT, UTIME_OMIT } case *****"
-run_test "" 600 "+a" "0 o 0 o" SUCCESS n n
-
-echo "***** Testing times=={ UTIME_NOW, UTIME_OMIT } case *****"
-run_test "" 600 "+a" "0 n 0 o" EPERM
-
-echo "***** Testing times=={ UTIME_OMIT, UTIME_NOW } case *****"
-run_test "" 600 "+a" "0 o 0 n" EPERM
-
-echo "***** Testing times=={ x, y } case *****"
-run_test "" 600 "+a" "1 1 1 1" EPERM
-
-echo "============================================================"
-
-echo
-echo "Testing immutable file, owned by self"
-echo
-
-echo "***** Testing times==NULL case *****"
-run_test -W "" 600 "+i" "" $imaccess
-
-echo "***** Testing times=={ UTIME_NOW, UTIME_NOW } case *****"
-run_test -W "" 600 "+i" "0 n 0 n" $imaccess
-
-echo "***** Testing times=={ UTIME_OMIT, UTIME_OMIT } case *****"
-run_test -W "" 600 "+i" "0 o 0 o" SUCCESS n n
-
-echo "***** Testing times=={ UTIME_NOW, UTIME_OMIT } case *****"
-run_test -W "" 600 "+i" "0 n 0 o" EPERM
-
-echo "***** Testing times=={ UTIME_OMIT, UTIME_NOW } case *****"
-run_test -W "" 600 "+i" "0 o 0 n" EPERM
-
-echo "***** Testing times=={ x, y } case *****"
-run_test -W "" 600 "+i" "1 1 1 1" EPERM
-
-echo "============================================================"
-
-# Immutable+append-only should have same results as immutable
-
-echo
-echo "Testing immutable append-only file, owned by self"
-echo
-
-echo "***** Testing times==NULL case *****"
-run_test -W "" 600 "+ai" "" $imaccess
-
-echo "***** Testing times=={ UTIME_NOW, UTIME_NOW } case *****"
-run_test -W "" 600 "+ai" "0 n 0 n" $imaccess
-
-echo "***** Testing times=={ UTIME_OMIT, UTIME_OMIT } case *****"
-run_test -W "" 600 "+ai" "0 o 0 o" SUCCESS n n
-
-echo "***** Testing times=={ UTIME_NOW, UTIME_OMIT } case *****"
-run_test -W "" 600 "+ai" "0 n 0 o" EPERM
-
-echo "***** Testing times=={ UTIME_OMIT, UTIME_NOW } case *****"
-run_test -W "" 600 "+ai" "0 o 0 n" EPERM
-
-echo "***** Testing times=={ x, y } case *****"
-run_test -W "" 600 "+ai" "1 1 1 1" EPERM
-
-echo "============================================================"
-
-echo
-
-# EINVAL should result, if pathname is NULL, dirfd is not
-# AT_FDCWD, and flags contains AT_SYMLINK_NOFOLLOW.
-
-echo "***** Testing pathname==NULL, dirfd!=AT_FDCWD, flags has" \
- "AT_SYMLINK_NOFOLLOW *****"
-setup_file $FILE "" 600 ""
-CMD="$TEST_PROG -q -n -d $FILE NULL $4"
-echo "$CMD"
-$CMD > $RESULT_FILE
-check_result $? EINVAL
-echo
-
-echo "============================================================"
-
-echo
-
-# If UTIME_NOW / UTIME_OMIT in tv_nsec field, the tv_sec should
-# be ignored.
-
-echo "tv_sec should be ignored if tv_nsec is UTIME_OMIT or UTIME_NOW"
-
-echo "***** Testing times=={ UTIME_NOW, UTIME_NOW } case *****"
-run_test -RW "" 600 "" "1 n 1 n" SUCCESS y y
-
-echo "***** Testing times=={ UTIME_OMIT, UTIME_OMIT } case *****"
-run_test -RW "" 600 "" "1 o 1 o" SUCCESS n n
-
-echo "============================================================"
-
-echo
-
-rm -rf "$TEST_DIR"
-uname -a
-date
-echo "Total tests: $test_num; passed: $passed_cnt; failed: $failed_cnt"
-if test $failed_cnt -gt 0; then
- echo "Failed tests: $failed_list"
-fi
-
-tst_exit
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 17/17] syscalls/utimensat: Add support for time64 tests
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
` (15 preceding siblings ...)
2020-05-08 4:24 ` [LTP] [PATCH V2 16/17] syscalls/utimensat: Migrate to new test framework Viresh Kumar
@ 2020-05-08 4:24 ` Viresh Kumar
16 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 4:24 UTC (permalink / raw)
To: ltp
This adds support for time64 tests to the existing utimensat() syscall
tests.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
.../kernel/syscalls/utimensat/utimensat01.c | 80 ++++++++++++++++---
1 file changed, 70 insertions(+), 10 deletions(-)
diff --git a/testcases/kernel/syscalls/utimensat/utimensat01.c b/testcases/kernel/syscalls/utimensat/utimensat01.c
index 2a029ecb17e9..f29cfe5009eb 100644
--- a/testcases/kernel/syscalls/utimensat/utimensat01.c
+++ b/testcases/kernel/syscalls/utimensat/utimensat01.c
@@ -15,8 +15,8 @@
#include <fcntl.h>
#include <string.h>
#include <sys/stat.h>
-#include "tst_test.h"
-#include "lapi/syscalls.h"
+#include "tst_timer.h"
+#include "lapi/abisize.h"
#define UTIME_NOW ((1l << 30) - 1l)
#define UTIME_OMIT ((1l << 30) - 2l)
@@ -98,13 +98,73 @@ struct test_case {
};
static inline int sys_utimensat(int dirfd, const char *pathname,
- const struct timespec times[2], int flags)
+ void *times, int flags)
{
return tst_syscall(__NR_utimensat, dirfd, pathname, times, flags);
}
+static inline int sys_utimensat_time64(int dirfd, const char *pathname,
+ void *times, int flags)
+{
+ return tst_syscall(__NR_utimensat_time64, dirfd, pathname, times, flags);
+}
+
+static struct test_variants {
+ int (*utimensat)(int dirfd, const char *pathname, void *times,
+ int flags);
+ enum tst_ts_type type;
+ char *desc;
+} variants[] = {
+#if defined(TST_ABI32)
+ { .utimensat = sys_utimensat, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
+ { .utimensat = sys_utimensat, .type = TST_KERN_OLD_TIMESPEC, .desc = "syscall with kernel spec32"},
+#endif
+
+#if defined(TST_ABI64)
+ { .utimensat = sys_utimensat, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
+#endif
+
+#if (__NR_utimensat_time64 != __LTP__NR_INVALID_SYSCALL)
+ { .utimensat = sys_utimensat_time64, .type = TST_KERN_TIMESPEC, .desc = "syscall time64 with kernel spec64"},
+#endif
+};
+
+union tst_multi {
+ struct timespec libc_ts[2];
+ struct __kernel_old_timespec kern_old_ts[2];
+ struct __kernel_timespec kern_ts[2];
+} ts;
+
+static void tst_multi_set_time(enum tst_ts_type type, long access_tv_sec,
+ long access_tv_nsec, long mod_tv_sec, long mod_tv_nsec)
+{
+ switch (type) {
+ case TST_LIBC_TIMESPEC:
+ ts.libc_ts[0].tv_sec = access_tv_sec;
+ ts.libc_ts[0].tv_nsec = access_tv_nsec;
+ ts.libc_ts[1].tv_sec = mod_tv_sec;
+ ts.libc_ts[1].tv_nsec = mod_tv_nsec;
+ break;
+ case TST_KERN_OLD_TIMESPEC:
+ ts.kern_old_ts[0].tv_sec = access_tv_sec;
+ ts.kern_old_ts[0].tv_nsec = access_tv_nsec;
+ ts.kern_old_ts[1].tv_sec = mod_tv_sec;
+ ts.kern_old_ts[1].tv_nsec = mod_tv_nsec;
+ break;
+ case TST_KERN_TIMESPEC:
+ ts.kern_ts[0].tv_sec = access_tv_sec;
+ ts.kern_ts[0].tv_nsec = access_tv_nsec;
+ ts.kern_ts[1].tv_sec = mod_tv_sec;
+ ts.kern_ts[1].tv_nsec = mod_tv_nsec;
+ break;
+ default:
+ tst_brk(TBROK, "Invalid type: %d", type);
+ }
+}
+
static void setup(void)
{
+ tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
bad_addr = tst_get_bad_addr(NULL);
}
@@ -134,8 +194,8 @@ static int run_command(char *command, char *option, char *file)
static void run(unsigned int i)
{
+ struct test_variants *tv = &variants[tst_variant];
struct test_case *tc = &tcase[i];
- struct timespec ts[2];
void *tsp = NULL;
char *pathname = NULL;
int fd = AT_FDCWD;
@@ -151,11 +211,10 @@ static void run(unsigned int i)
if (tc->mytime) {
struct mytime *mytime = *tc->mytime;
- ts[0].tv_sec = mytime->access_tv_sec;
- ts[0].tv_nsec = mytime->access_tv_nsec;
- ts[1].tv_sec = mytime->mod_tv_sec;
- ts[1].tv_nsec = mytime->mod_tv_nsec;
- tsp = ts;
+ tst_multi_set_time(tv->type, mytime->access_tv_sec,
+ mytime->access_tv_nsec, mytime->mod_tv_sec,
+ mytime->mod_tv_nsec);
+ tsp = &ts.libc_ts;
} else if (tc->exp_err == EFAULT) {
tsp = bad_addr;
}
@@ -169,7 +228,7 @@ static void run(unsigned int i)
pathname = bad_addr;
}
- TEST(sys_utimensat(fd, pathname, tsp, tc->flags));
+ TEST(tv->utimensat(fd, pathname, tsp, tc->flags));
if (TST_RET) {
if (!tc->exp_err) {
tst_res(TFAIL | TTERRNO, "%2d: utimensat() failed", i);
@@ -199,6 +258,7 @@ static void run(unsigned int i)
static struct tst_test test = {
.test = run,
.tcnt = ARRAY_SIZE(tcase),
+ .test_variants = ARRAY_SIZE(variants),
.setup = setup,
.needs_root = 1,
.needs_tmpdir = 1,
--
2.25.0.rc1.19.g042ed3e048af
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-08 4:24 ` [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version Viresh Kumar
@ 2020-05-08 7:18 ` Arnd Bergmann
2020-05-08 8:56 ` Viresh Kumar
2020-05-11 6:49 ` Viresh Kumar
2020-05-09 4:21 ` Li Wang
1 sibling, 2 replies; 37+ messages in thread
From: Arnd Bergmann @ 2020-05-08 7:18 UTC (permalink / raw)
To: ltp
On Fri, May 8, 2020 at 6:24 AM Viresh Kumar <viresh.kumar@linaro.org> wrote:
> +
> +static inline int sys_semtimedop(int semid, struct sembuf *sops, size_t nsops,
> + void *timeout)
> +{
> + return tst_syscall(__NR_semtimedop, semid, sops, nsops, timeout);
> +}
> +
> +static inline int sys_semtimedop_time64(int semid, struct sembuf *sops,
> + size_t nsops, void *timeout)
> +{
> + return tst_syscall(__NR_semtimedop_time64, semid, sops, nsops, timeout);
> +}
> +
> +struct test_variants {
> + int (*semop)(int semid, struct sembuf *sops, size_t nsops);
> + int (*semtimedop)(int semid, struct sembuf *sops, size_t nsops, void *timeout);
> + enum tst_ts_type type;
> + char *desc;
> +} variants[] = {
> + { .semop = semop, .type = TST_LIBC_TIMESPEC, .desc = "semop: vDSO or syscall"},
> +#if defined(TST_ABI32)
> + { .semtimedop = sys_semtimedop, .type = TST_LIBC_TIMESPEC, .desc = "semtimedop: syscall with libc spec"},
> + { .semtimedop = sys_semtimedop, .type = TST_KERN_OLD_TIMESPEC, .desc = "semtimedop: syscall with kernel spec32"},
> +#endif
> +
> +#if defined(TST_ABI64)
> + { .semtimedop = sys_semtimedop, .type = TST_KERN_TIMESPEC, .desc = "semtimedop: syscall with kernel spec64"},
> +#endif
It feels like this is more complicated than it need to be. The line
semtimedop = sys_semtimedop, .type = TST_KERN_OLD_TIMESPEC, .desc =
"semtimedop: syscall with kernel spec32"},
should apply to any kernel that has "__NR_semtimedop !=
__LTP__NR_INVALID_SYSCALL",
regardless of any other macros set, and then you don't need the separate line
{ .semtimedop = sys_semtimedop, .type = TST_KERN_TIMESPEC, .desc =
"semtimedop: syscall with kernel spec64"},
which is not what the ABI is meant to be anyway (sys_semtimedop takes
a __kernel_old_timespec,
not a __kernel_timespec).
Similarly, the line
{ .semop = semop, .type = TST_LIBC_TIMESPEC, .desc = "semop: vDSO or syscall"},
should apply to both 32 and 64 bit machines
Arnd
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 13/17] syscalls/select6: Add support for time64 tests
2020-05-08 4:23 ` [LTP] [PATCH V2 13/17] syscalls/select6: " Viresh Kumar
@ 2020-05-08 7:21 ` Arnd Bergmann
0 siblings, 0 replies; 37+ messages in thread
From: Arnd Bergmann @ 2020-05-08 7:21 UTC (permalink / raw)
To: ltp
On Fri, May 8, 2020 at 6:24 AM Viresh Kumar <viresh.kumar@linaro.org> wrote:
> + };
> +#if defined(TST_ABI64)
> + ret = tst_syscall(__NR_pselect6, nfds, readfds, writefds, exceptfds, &ts, NULL);
Again, this is not the actual ABI: __NR_pselect6 should be called with
a __kernel_old_timespec when the macro is defined, but it may not be defined
on new 32-bit architectures.
Arnd
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 01/17] syscalls/timer_gettime: Add support for time64 tests
2020-05-08 4:23 ` [LTP] [PATCH V2 01/17] syscalls/timer_gettime: Add support for time64 tests Viresh Kumar
@ 2020-05-08 8:30 ` Arnd Bergmann
0 siblings, 0 replies; 37+ messages in thread
From: Arnd Bergmann @ 2020-05-08 8:30 UTC (permalink / raw)
To: ltp
On Fri, May 8, 2020 at 6:24 AM Viresh Kumar <viresh.kumar@linaro.org> wrote:
> +#if defined(TST_ABI32)
> + { .func = sys_timer_gettime, .type = TST_LIBC_TIMESPEC, .desc = "syscall with libc spec"},
> +#endif
> +#if defined(TST_ABI64)
> + { .func = sys_timer_gettime, .type = TST_KERN_TIMESPEC, .desc = "syscall with kernel spec64"},
> +#endif
I think these are both invalid: timer_gettime() does not take a libc timespec or
a __kernel_timespec but takes a __kernel_old_timespec. You should only
use the libc data structure with the libc timer_gettime, and the
__kernel_timespec
with the kernel's timer_gettime64().
Arnd
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-08 7:18 ` Arnd Bergmann
@ 2020-05-08 8:56 ` Viresh Kumar
2020-05-08 9:24 ` Arnd Bergmann
2020-05-11 6:49 ` Viresh Kumar
1 sibling, 1 reply; 37+ messages in thread
From: Viresh Kumar @ 2020-05-08 8:56 UTC (permalink / raw)
To: ltp
On 08-05-20, 09:18, Arnd Bergmann wrote:
> On Fri, May 8, 2020 at 6:24 AM Viresh Kumar <viresh.kumar@linaro.org> wrote:
>
> > +
> > +static inline int sys_semtimedop(int semid, struct sembuf *sops, size_t nsops,
> > + void *timeout)
> > +{
> > + return tst_syscall(__NR_semtimedop, semid, sops, nsops, timeout);
> > +}
> > +
> > +static inline int sys_semtimedop_time64(int semid, struct sembuf *sops,
> > + size_t nsops, void *timeout)
> > +{
> > + return tst_syscall(__NR_semtimedop_time64, semid, sops, nsops, timeout);
> > +}
> > +
> > +struct test_variants {
> > + int (*semop)(int semid, struct sembuf *sops, size_t nsops);
> > + int (*semtimedop)(int semid, struct sembuf *sops, size_t nsops, void *timeout);
> > + enum tst_ts_type type;
> > + char *desc;
> > +} variants[] = {
> > + { .semop = semop, .type = TST_LIBC_TIMESPEC, .desc = "semop: vDSO or syscall"},
> > +#if defined(TST_ABI32)
> > + { .semtimedop = sys_semtimedop, .type = TST_LIBC_TIMESPEC, .desc = "semtimedop: syscall with libc spec"},
> > + { .semtimedop = sys_semtimedop, .type = TST_KERN_OLD_TIMESPEC, .desc = "semtimedop: syscall with kernel spec32"},
> > +#endif
> > +
> > +#if defined(TST_ABI64)
> > + { .semtimedop = sys_semtimedop, .type = TST_KERN_TIMESPEC, .desc = "semtimedop: syscall with kernel spec64"},
> > +#endif
>
>
> It feels like this is more complicated than it need to be. The line
>
> semtimedop = sys_semtimedop, .type = TST_KERN_OLD_TIMESPEC, .desc =
> "semtimedop: syscall with kernel spec32"},
>
> should apply to any kernel that has "__NR_semtimedop !=
> __LTP__NR_INVALID_SYSCALL",
> regardless of any other macros set, and then you don't need the separate line
>
> { .semtimedop = sys_semtimedop, .type = TST_KERN_TIMESPEC, .desc =
> "semtimedop: syscall with kernel spec64"},
> which is not what the ABI is meant to be anyway (sys_semtimedop takes
> a __kernel_old_timespec,
> not a __kernel_timespec).
There is some misunderstanding here, surely from my side. The sys_
helpers here are the direct syscalls called from userspace with help
of tst_syscall().
AFAIU, on 32 bit systems we need to call __NR_semtimedop with the 32
bit and 64 bit timespec (both), and on 64 bit systems which don't
implement __NR_semtimedop_time64, we need to call __NR_semtimedop with
the 64 bit timespec only.
What you are telling now is very different from that and so I don't
get it.
> Similarly, the line
>
> { .semop = semop, .type = TST_LIBC_TIMESPEC, .desc = "semop: vDSO or syscall"},
>
> should apply to both 32 and 64 bit machines
Yes and so it is called without ifdef hackery. Isn't that correct ?
--
viresh
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-08 8:56 ` Viresh Kumar
@ 2020-05-08 9:24 ` Arnd Bergmann
2020-05-11 5:44 ` Viresh Kumar
0 siblings, 1 reply; 37+ messages in thread
From: Arnd Bergmann @ 2020-05-08 9:24 UTC (permalink / raw)
To: ltp
On Fri, May 8, 2020 at 10:57 AM Viresh Kumar <viresh.kumar@linaro.org> wrote:
>
> On 08-05-20, 09:18, Arnd Bergmann wrote:
> > On Fri, May 8, 2020 at 6:24 AM Viresh Kumar <viresh.kumar@linaro.org> wrote:
> >
> > > +
> > > +static inline int sys_semtimedop(int semid, struct sembuf *sops, size_t nsops,
> > > + void *timeout)
> > > +{
> > > + return tst_syscall(__NR_semtimedop, semid, sops, nsops, timeout);
> > > +}
> > > +
> > > +static inline int sys_semtimedop_time64(int semid, struct sembuf *sops,
> > > + size_t nsops, void *timeout)
> > > +{
> > > + return tst_syscall(__NR_semtimedop_time64, semid, sops, nsops, timeout);
> > > +}
> > > +
> > > +struct test_variants {
> > > + int (*semop)(int semid, struct sembuf *sops, size_t nsops);
> > > + int (*semtimedop)(int semid, struct sembuf *sops, size_t nsops, void *timeout);
> > > + enum tst_ts_type type;
> > > + char *desc;
> > > +} variants[] = {
> > > + { .semop = semop, .type = TST_LIBC_TIMESPEC, .desc = "semop: vDSO or syscall"},
> > > +#if defined(TST_ABI32)
> > > + { .semtimedop = sys_semtimedop, .type = TST_LIBC_TIMESPEC, .desc = "semtimedop: syscall with libc spec"},
> > > + { .semtimedop = sys_semtimedop, .type = TST_KERN_OLD_TIMESPEC, .desc = "semtimedop: syscall with kernel spec32"},
> > > +#endif
> > > +
> > > +#if defined(TST_ABI64)
> > > + { .semtimedop = sys_semtimedop, .type = TST_KERN_TIMESPEC, .desc = "semtimedop: syscall with kernel spec64"},
> > > +#endif
> >
> >
> > It feels like this is more complicated than it need to be. The line
> >
> > semtimedop = sys_semtimedop, .type = TST_KERN_OLD_TIMESPEC, .desc =
> > "semtimedop: syscall with kernel spec32"},
> >
> > should apply to any kernel that has "__NR_semtimedop !=
> > __LTP__NR_INVALID_SYSCALL",
> > regardless of any other macros set, and then you don't need the separate line
> >
> > { .semtimedop = sys_semtimedop, .type = TST_KERN_TIMESPEC, .desc =
> > "semtimedop: syscall with kernel spec64"},
>
> > which is not what the ABI is meant to be anyway (sys_semtimedop takes
> > a __kernel_old_timespec,
> > not a __kernel_timespec).
>
> There is some misunderstanding here, surely from my side. The sys_
> helpers here are the direct syscalls called from userspace with help
> of tst_syscall().
>
> AFAIU, on 32 bit systems we need to call __NR_semtimedop with the 32
> bit and 64 bit timespec (both), and on 64 bit systems which don't
> implement __NR_semtimedop_time64, we need to call __NR_semtimedop with
> the 64 bit timespec only.
>
> What you are telling now is very different from that and so I don't
> get it.
__NR_semtimedop can only be called with the 'old' timespec, which
may have either 32 or 64 members depending on the architecture.
On x32 it uses 64-bit members, and on riscv32 it does not exist at all.
I think you already have a correct __kernel_old_timespec definition,
so what I'd expect to see here is code that passes __kernel_old_timespec
into __NR_semtimedop whenever __NR_semtimedop is defined.
Passing the libc timespec into __kernel_old_timespec is a bug, as
the libc may be using either the old or the new (always 64-bit)
definition.
> > { .semop = semop, .type = TST_LIBC_TIMESPEC, .desc = "semop: vDSO or syscall"},
> >
> > should apply to both 32 and 64 bit machines
>
> Yes and so it is called without ifdef hackery. Isn't that correct ?
My mistake, I confused the lines. What I meant is that there should
be an unconditional test of the libc 'semtimedop' with the libc 'timespec'
definition.
Arnd
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 16/17] syscalls/utimensat: Migrate to new test framework
2020-05-08 4:24 ` [LTP] [PATCH V2 16/17] syscalls/utimensat: Migrate to new test framework Viresh Kumar
@ 2020-05-09 3:50 ` Li Wang
2020-05-11 9:33 ` Cyril Hrubis
1 sibling, 0 replies; 37+ messages in thread
From: Li Wang @ 2020-05-09 3:50 UTC (permalink / raw)
To: ltp
Hi Viresh,
On Fri, May 8, 2020 at 12:28 PM Viresh Kumar <viresh.kumar@linaro.org>
wrote:
> This gets rid of the test script and updates the utimensat01.c file to
> handle most of the tests covered earlier (apart from testing the same
> for files owned by root, as the test is run as root right now).
>
Since we have removed the utimensat01_test.sh script, the Makefile,
runtest/syscalls should be updated too.
--- a/runtest/syscalls
+++ b/runtest/syscalls
@@ -1600,7 +1600,7 @@ utime06 utime06
utimes01 utimes01
# Introduced from Kernel 2.6.22 onwards
-utimensat01 utimensat_tests.sh
+utimensat01 utimensat01
vfork01 vfork01
vfork02 vfork02
diff --git a/testcases/kernel/syscalls/utimensat/Makefile
b/testcases/kernel/syscalls/utimensat/Makefile
index 69cab8dbc..044619fb8 100644
--- a/testcases/kernel/syscalls/utimensat/Makefile
+++ b/testcases/kernel/syscalls/utimensat/Makefile
@@ -5,8 +5,4 @@ top_srcdir ?= ../../../..
include $(top_srcdir)/include/mk/testcases.mk
-vpath %.c $(srcdir):$(top_srcdir)/lib
-
-INSTALL_TARGETS := utimensat_tests.sh
-
include $(top_srcdir)/include/mk/generic_leaf_target.mk
--
Regards,
Li Wang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linux.it/pipermail/ltp/attachments/20200509/3d8982e1/attachment-0001.htm>
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-08 4:24 ` [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version Viresh Kumar
2020-05-08 7:18 ` Arnd Bergmann
@ 2020-05-09 4:21 ` Li Wang
2020-05-11 10:43 ` Viresh Kumar
1 sibling, 1 reply; 37+ messages in thread
From: Li Wang @ 2020-05-09 4:21 UTC (permalink / raw)
To: ltp
On Fri, May 8, 2020 at 12:28 PM Viresh Kumar <viresh.kumar@linaro.org>
wrote:
> This adds support for semtimedop() and its time64 variant to the
> existing semop() syscall tests.
>
> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
> ---
> testcases/kernel/syscalls/ipc/semop/semop.h | 52 +++++++++++++++++++
> testcases/kernel/syscalls/ipc/semop/semop01.c | 11 +++-
> testcases/kernel/syscalls/ipc/semop/semop02.c | 12 ++++-
> testcases/kernel/syscalls/ipc/semop/semop03.c | 12 ++++-
> testcases/kernel/syscalls/ipc/semop/semop04.c | 12 ++++-
> testcases/kernel/syscalls/ipc/semop/semop05.c | 12 ++++-
> 6 files changed, 106 insertions(+), 5 deletions(-)
> create mode 100644 testcases/kernel/syscalls/ipc/semop/semop.h
>
> diff --git a/testcases/kernel/syscalls/ipc/semop/semop.h
> b/testcases/kernel/syscalls/ipc/semop/semop.h
> new file mode 100644
> index 000000000000..8d1245b65ec0
> --- /dev/null
> +++ b/testcases/kernel/syscalls/ipc/semop/semop.h
> @@ -0,0 +1,52 @@
> +// SPDX-License-Identifier: GPL-2.0-or-later
> +
> +#ifndef SEMOP_VAR__
> +#define SEMOP_VAR__
> +
> +#include <sys/sem.h>
> +#include "tst_timer.h"
> +#include "lapi/abisize.h"
> +
> +static inline int sys_semtimedop(int semid, struct sembuf *sops, size_t
> nsops,
> + void *timeout)
> +{
> + return tst_syscall(__NR_semtimedop, semid, sops, nsops, timeout);
> +}
> +
> +static inline int sys_semtimedop_time64(int semid, struct sembuf *sops,
> + size_t nsops, void *timeout)
> +{
> + return tst_syscall(__NR_semtimedop_time64, semid, sops, nsops,
> timeout);
> +}
> +
> +struct test_variants {
> + int (*semop)(int semid, struct sembuf *sops, size_t nsops);
> + int (*semtimedop)(int semid, struct sembuf *sops, size_t nsops,
> void *timeout);
> + enum tst_ts_type type;
> + char *desc;
> +} variants[] = {
> + { .semop = semop, .type = TST_LIBC_TIMESPEC, .desc = "semop: vDSO
> or syscall"},
> +#if defined(TST_ABI32)
> + { .semtimedop = sys_semtimedop, .type = TST_LIBC_TIMESPEC, .desc =
> "semtimedop: syscall with libc spec"},
> + { .semtimedop = sys_semtimedop, .type = TST_KERN_OLD_TIMESPEC,
> .desc = "semtimedop: syscall with kernel spec32"},
> +#endif
> +
> +#if defined(TST_ABI64)
> + { .semtimedop = sys_semtimedop, .type = TST_KERN_TIMESPEC, .desc =
> "semtimedop: syscall with kernel spec64"},
> +#endif
> +
> +#if (__NR_semtimedop_time64 != __LTP__NR_INVALID_SYSCALL)
> + { .semtimedop = sys_semtimedop_time64, .type = TST_KERN_TIMESPEC,
> .desc = "semtimedop: syscall time64 with kernel spec64"},
> +#endif
> +};
> +
> +static inline int call_semop(struct test_variants *tv, int semid,
> + struct sembuf *sops, size_t nsops, struct tst_ts *timeout)
> +{
> + if (tv->semop)
> + return tv->semop(semid, sops, nsops);
> +
> + return tv->semtimedop(semid, sops, nsops, tst_ts_get(timeout));
> +}
> +
> +#endif /* SEMOP_VAR__ */
> diff --git a/testcases/kernel/syscalls/ipc/semop/semop01.c
> b/testcases/kernel/syscalls/ipc/semop/semop01.c
> index bcb45fa69320..a4924376e32c 100644
> --- a/testcases/kernel/syscalls/ipc/semop/semop01.c
> +++ b/testcases/kernel/syscalls/ipc/semop/semop01.c
> @@ -41,6 +41,7 @@
> #include "tst_test.h"
> #include "libnewipc.h"
> #include "lapi/semun.h"
> +#include "semop.h"
>
> #define NSEMS 4 /* the number of primitive semaphores to
> test */
>
> @@ -52,11 +53,16 @@ static struct sembuf sops[PSEMS];
>
> static void run(void)
> {
> + struct test_variants *tv = &variants[tst_variant];
> union semun arr = { .val = 0 };
> + struct tst_ts timeout;
> int fail = 0;
> int i;
>
> - TEST(semop(sem_id, sops, NSEMS));
> + timeout.type = tv->type;
> + tst_ts_set_nsec(&timeout, 10000);
> +
> + TEST(call_semop(tv, sem_id, sops, NSEMS, &timeout));
>
> if (TST_RET == -1) {
> tst_res(TFAIL | TTERRNO, "semop() failed");
> @@ -94,6 +100,8 @@ static void setup(void)
> {
> int i;
>
> + tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
> +
> get_arr.array = malloc(sizeof(unsigned short int) * PSEMS);
> if (get_arr.array == NULL)
> tst_brk(TBROK, "malloc failed");
> @@ -124,6 +132,7 @@ static void cleanup(void)
>
> static struct tst_test test = {
> .test_all = run,
> + .test_variants = ARRAY_SIZE(variants),
> .setup = setup,
> .cleanup = cleanup,
> .needs_tmpdir = 1,
> diff --git a/testcases/kernel/syscalls/ipc/semop/semop02.c
> b/testcases/kernel/syscalls/ipc/semop/semop02.c
> index f24d284776a4..a51906340400 100644
> --- a/testcases/kernel/syscalls/ipc/semop/semop02.c
> +++ b/testcases/kernel/syscalls/ipc/semop/semop02.c
> @@ -20,6 +20,7 @@
> #include "tst_test.h"
> #include "libnewipc.h"
> #include "lapi/semun.h"
> +#include "semop.h"
>
> static int sem_id_1 = -1; /* a semaphore set with read & alter
> permissions */
> static int sem_id_2 = -1; /* a semaphore set without read & alter
> permissions */
> @@ -52,6 +53,8 @@ static void setup(void)
> struct seminfo ipc_buf;
> union semun arr;
>
> + tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
> +
> ltpuser = SAFE_GETPWNAM(nobody_uid);
> SAFE_SETUID(ltpuser->pw_uid);
>
> @@ -84,7 +87,13 @@ static void setup(void)
>
> static void run(unsigned int i)
> {
> - TEST(semop(*(tc[i].semid), tc[i].t_sbuf, tc[i].t_ops));
> + struct test_variants *tv = &variants[tst_variant];
> + struct tst_ts timeout;
> +
> + timeout.type = tv->type;
> + tst_ts_set_nsec(&timeout, 10000);
> +
> + TEST(call_semop(tv, *(tc[i].semid), tc[i].t_sbuf, tc[i].t_ops,
> &timeout));
>
> if (TST_RET != -1) {
> tst_res(TFAIL | TTERRNO, "call succeeded unexpectedly");
> @@ -117,6 +126,7 @@ static void cleanup(void)
> static struct tst_test test = {
> .test = run,
> .tcnt = ARRAY_SIZE(tc),
> + .test_variants = ARRAY_SIZE(variants),
> .setup = setup,
> .cleanup = cleanup,
> .needs_tmpdir = 1,
> diff --git a/testcases/kernel/syscalls/ipc/semop/semop03.c
> b/testcases/kernel/syscalls/ipc/semop/semop03.c
> index 4f5f78eb6d8d..d36194ce1f41 100644
> --- a/testcases/kernel/syscalls/ipc/semop/semop03.c
> +++ b/testcases/kernel/syscalls/ipc/semop/semop03.c
> @@ -38,6 +38,7 @@
> #include "tst_test.h"
> #include "libnewipc.h"
> #include "lapi/semun.h"
> +#include "semop.h"
>
> static key_t semkey;
> static int sem_id = -1;
> @@ -47,6 +48,12 @@ static int tc[] = { -1, PSEMS + 1 }; /* negative and
> too many "primitive" semas
>
> static void run(unsigned int i)
> {
> + struct test_variants *tv = &variants[tst_variant];
> + struct tst_ts timeout;
> +
> + timeout.type = tv->type;
> + tst_ts_set_nsec(&timeout, 10000);
> +
> /* initialize two fields in the sembuf structure here */
> s_buf.sem_op = 1; /* add this value to struct sem.semval */
> s_buf.sem_flg = SEM_UNDO; /* undo when process exits */
> @@ -61,7 +68,7 @@ static void run(unsigned int i)
> * use the TEST macro to make the call
> */
>
> - TEST(semop(sem_id, &s_buf, 1));
> + TEST(call_semop(tv, sem_id, &s_buf, 1, &timeout));
>
> if (TST_RET != -1) {
> tst_res(TFAIL | TTERRNO, "call succeeded unexpectedly");
> @@ -80,6 +87,8 @@ static void run(unsigned int i)
>
> static void setup(void)
> {
> + tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
> +
> /* get an IPC resource key */
> semkey = GETIPCKEY();
>
> @@ -102,6 +111,7 @@ static void cleanup(void)
> static struct tst_test test = {
> .test = run,
> .tcnt = ARRAY_SIZE(tc),
> + .test_variants = ARRAY_SIZE(variants),
> .setup = setup,
> .cleanup = cleanup,
> .needs_tmpdir = 1,
> diff --git a/testcases/kernel/syscalls/ipc/semop/semop04.c
> b/testcases/kernel/syscalls/ipc/semop/semop04.c
> index 0faf00a3585f..033065e7bbfc 100644
> --- a/testcases/kernel/syscalls/ipc/semop/semop04.c
> +++ b/testcases/kernel/syscalls/ipc/semop/semop04.c
> @@ -38,6 +38,7 @@
> #include "tst_test.h"
> #include "libnewipc.h"
> #include "lapi/semun.h"
> +#include "semop.h"
>
> static int sem_id = -1;
> static int val; /* value for SETVAL */
> @@ -62,6 +63,12 @@ static struct test_case_t {
>
> static void run(unsigned int i)
> {
> + struct test_variants *tv = &variants[tst_variant];
> + struct tst_ts timeout;
> +
> + timeout.type = tv->type;
> + tst_ts_set_nsec(&timeout, 10000);
> +
> /* initialize the s_buf buffer */
> s_buf.sem_op = tc[i].op;
> s_buf.sem_flg = tc[i].flg;
> @@ -72,7 +79,7 @@ static void run(unsigned int i)
> if (semctl(sem_id, tc[i].num, SETVAL, tc[i].get_arr) == -1)
> tst_brk(TBROK | TERRNO, "semctl() failed");
>
> - TEST(semop(sem_id, &s_buf, 1));
> + TEST(call_semop(tv, sem_id, &s_buf, 1, &timeout));
> if (TST_RET != -1) {
> tst_res(TFAIL, "call succeeded unexpectedly");
> return;
> @@ -86,6 +93,8 @@ static void run(unsigned int i)
>
> static void setup(void)
> {
> + tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
> +
> val = 1;
>
> /* get an IPC resource key */
> @@ -114,6 +123,7 @@ static void cleanup(void)
> static struct tst_test test = {
> .test = run,
> .tcnt = ARRAY_SIZE(tc),
> + .test_variants = ARRAY_SIZE(variants),
> .setup = setup,
> .cleanup = cleanup,
> .needs_tmpdir = 1,
> diff --git a/testcases/kernel/syscalls/ipc/semop/semop05.c
> b/testcases/kernel/syscalls/ipc/semop/semop05.c
> index 9e8e040b0b19..f892cd519ecf 100644
> --- a/testcases/kernel/syscalls/ipc/semop/semop05.c
> +++ b/testcases/kernel/syscalls/ipc/semop/semop05.c
> @@ -52,6 +52,7 @@
> #include "tst_test.h"
> #include "libnewipc.h"
> #include "lapi/semun.h"
> +#include "semop.h"
>
> static key_t semkey;
> static int sem_id = -1;
> @@ -115,7 +116,13 @@ static inline int process_state_wait2(pid_t pid,
> const char state)
>
> static void do_child(int i)
> {
> - TEST(semop(sem_id, &s_buf, 1));
> + struct test_variants *tv = &variants[tst_variant];
> + struct tst_ts timeout;
> +
> + timeout.type = tv->type;
> + tst_ts_set_nsec(&timeout, 10000);
> +
> + TEST(call_semop(tv, sem_id, &s_buf, 1, &timeout));
> if (TST_RET != -1) {
> tst_res(TFAIL, "call succeeded when error expected");
> exit(-1);
> @@ -137,6 +144,8 @@ static void sighandler(int sig)
>
> static void setup(void)
> {
> + tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
> +
> SAFE_SIGNAL(SIGHUP, sighandler);
>
> /* get an IPC resource key */
> @@ -238,6 +247,7 @@ static void do_child_uclinux(void)
> static struct tst_test test = {
> .test = run,
> .tcnt = ARRAY_SIZE(tc),
> + .test_variants = ARRAY_SIZE(variants),
>
RHEL8.2(x84_64) failed as:
# ./semop05
tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
semop05.c:147: INFO: Testing variant: semop: vDSO or syscall
semop05.c:132: PASS: expected failure: EIDRM (43)
semop05.c:132: PASS: expected failure: EIDRM (43)
semop05.c:132: PASS: expected failure: EINTR (4)
semop05.c:132: PASS: expected failure: EINTR (4)
tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
semop05.c:147: INFO: Testing variant: semtimedop: syscall with kernel spec64
semop05.c:134: FAIL: unexpected failure: EAGAIN/EWOULDBLOCK (11)
Test timeouted, sending SIGKILL!
tst_test.c:1286: INFO: If you are running on slow machine, try exporting
LTP_TIMEOUT_MUL > 1
tst_test.c:1287: BROK: Test killed! (timeout?)
RHEL8.2(s390x) failed as:
tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
semop05.c:147: INFO: Testing variant: semop: vDSO or syscall
semop05.c:132: PASS: expected failure: EIDRM (43)
semop05.c:132: PASS: expected failure: EIDRM (43)
semop05.c:132: PASS: expected failure: EINTR (4)
semop05.c:132: PASS: expected failure: EINTR (4)
tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
semop05.c:147: INFO: Testing variant: semtimedop: syscall with kernel spec64
semop.h:13: CONF: syscall(392) __NR_semtimedop not supported
Test timeouted, sending SIGKILL!
tst_test.c:1286: INFO: If you are running on slow machine, try exporting
LTP_TIMEOUT_MUL > 1
tst_test.c:1287: BROK: Test killed! (timeout?)
> .setup = setup,
> .cleanup = cleanup,
> .needs_tmpdir = 1,
> --
> 2.25.0.rc1.19.g042ed3e048af
>
>
> --
> Mailing list info: https://lists.linux.it/listinfo/ltp
>
>
--
Regards,
Li Wang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linux.it/pipermail/ltp/attachments/20200509/01bea12c/attachment-0001.htm>
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 06/17] syscalls/futex: Add support for time64 tests
2020-05-08 4:23 ` [LTP] [PATCH V2 06/17] syscalls/futex: Add support for time64 tests Viresh Kumar
@ 2020-05-09 9:02 ` Li Wang
2020-05-11 11:08 ` Viresh Kumar
0 siblings, 1 reply; 37+ messages in thread
From: Li Wang @ 2020-05-09 9:02 UTC (permalink / raw)
To: ltp
On Fri, May 8, 2020 at 12:27 PM Viresh Kumar <viresh.kumar@linaro.org>
wrote:
> This adds support for time64 tests to the existing futex() syscall
> tests.
>
> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
> Reviewed-by: Petr Vorel <pvorel@suse.cz>
> ---
> include/tst_timer.h | 12 ++
> .../syscalls/futex/futex_cmp_requeue01.c | 35 +++-
> .../syscalls/futex/futex_cmp_requeue02.c | 25 ++-
> testcases/kernel/syscalls/futex/futex_utils.h | 52 ++++--
> .../kernel/syscalls/futex/futex_wait01.c | 83 +++++-----
> .../kernel/syscalls/futex/futex_wait02.c | 104 ++++++------
> .../kernel/syscalls/futex/futex_wait03.c | 91 +++++------
> .../kernel/syscalls/futex/futex_wait04.c | 85 +++++-----
> .../kernel/syscalls/futex/futex_wait05.c | 2 +-
> .../syscalls/futex/futex_wait_bitset01.c | 50 ++++--
> .../kernel/syscalls/futex/futex_wake01.c | 71 ++++-----
> .../kernel/syscalls/futex/futex_wake02.c | 99 ++++++------
> .../kernel/syscalls/futex/futex_wake03.c | 111 +++++++------
> .../kernel/syscalls/futex/futex_wake04.c | 150 +++++++++---------
> testcases/kernel/syscalls/futex/futextest.h | 111 ++++++++-----
> 15 files changed, 609 insertions(+), 472 deletions(-)
>
> diff --git a/include/tst_timer.h b/include/tst_timer.h
> index a726298b407d..b4280042c77f 100644
> --- a/include/tst_timer.h
> +++ b/include/tst_timer.h
> @@ -243,6 +243,18 @@ static inline int sys_clock_nanosleep64(clockid_t
> clk_id, int flags,
> request, remain);
> }
>
> +static inline int sys_futex(int *uaddr, int futex_op, int val, void *to,
> + int *uaddr2, int val3)
> +{
> + return tst_syscall(__NR_futex, uaddr, futex_op, val, to, uaddr2,
> val3);
> +}
> +
> +static inline int sys_futex_time64(int *uaddr, int futex_op, int val,
> void *to,
> + int *uaddr2, int val3)
> +{
> + return tst_syscall(__NR_futex_time64, uaddr, futex_op, val, to,
> uaddr2, val3);
> +}
> +
> static inline int libc_sched_rr_get_interval(pid_t pid, void *ts)
> {
> return sched_rr_get_interval(pid, ts);
> diff --git a/testcases/kernel/syscalls/futex/futex_cmp_requeue01.c
> b/testcases/kernel/syscalls/futex/futex_cmp_requeue01.c
> index c50fade9f969..bc9d2a6bee37 100644
> --- a/testcases/kernel/syscalls/futex/futex_cmp_requeue01.c
> +++ b/testcases/kernel/syscalls/futex/futex_cmp_requeue01.c
> @@ -18,6 +18,7 @@
> #include "tst_timer_test.h"
> #include "tst_test.h"
> #include "futextest.h"
> +#include "lapi/abisize.h"
>
> struct shared_data {
> futex_t futexes[2];
> @@ -42,14 +43,33 @@ static struct tcase {
> {1000, 300, 500},
> };
>
> +static struct test_variants {
> + enum futex_fn_type fntype;
> + enum tst_ts_type tstype;
> + char *desc;
> +} variants[] = {
> +#if defined(TST_ABI32)
> + { .fntype = FUTEX_FN_FUTEX, .tstype = TST_KERN_OLD_TIMESPEC, .desc
> = "syscall with kernel spec32"},
> +#endif
> +
> +#if defined(TST_ABI64)
> + { .fntype = FUTEX_FN_FUTEX, .tstype = TST_KERN_TIMESPEC, .desc =
> "syscall with kernel spec64"},
> +#endif
> +
> +#if (__NR_futex_time64 != __LTP__NR_INVALID_SYSCALL)
>
This might not safe to use it directly if only check __NR_futex_time64 being
defined. i.e Test it on the platform without achieving the system
syscall(but
define __NR_futex_time64), the test children exit with TCONF will cause
kind of issues.
RHEL-8.2(ppc64le) failed as:
--------------------------------------
tag=futex_cmp_requeue01 stime=1588947939
cmdline="futex_cmp_requeue01"
contacts=""
analysis=exit
<<<test_output>>>
tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
futex_cmp_requeue01.c:202: INFO: Testing variant: syscall with kernel spec64
futex_cmp_requeue01.c:114: INFO: Test 0: waiters: 10, wakes: 3, requeues: 7
futex_cmp_requeue01.c:136: INFO: futex_cmp_requeue() returned 10
futex_cmp_requeue01.c:152: INFO: children woken, futex0: 0, futex1: 7,
spurious wakeups: 0
futex_cmp_requeue01.c:197: PASS: futex_cmp_requeue()
futex_cmp_requeue01.c:114: INFO: Test 1: waiters: 10, wakes: 0, requeues: 10
futex_cmp_requeue01.c:136: INFO: futex_cmp_requeue() returned 10
futex_cmp_requeue01.c:152: INFO: children woken, futex0: 0, futex1:
10, spurious wakeups: 0
futex_cmp_requeue01.c:197: PASS: futex_cmp_requeue()
futex_cmp_requeue01.c:114: INFO: Test 2: waiters: 10, wakes: 2, requeues: 6
...
futex_cmp_requeue01.c:152: INFO: children woken, futex0: 200, futex1:
500, spurious wakeups: 0
futex_cmp_requeue01.c:197: PASS: futex_cmp_requeue()
tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
futex_cmp_requeue01.c:202: INFO: Testing variant: syscall time64 with
kernel spec64
../../../../include/tst_timer.h:256: CONF: syscall(422)
__NR_futex_time64 not supported
...
../../../../include/tst_timer.h:256: CONF: syscall(422)
__NR_futex_time64 not supported
../../../../include/tst_timer.h:256: CONF: syscall(422)
__NR_futex_time64 not supported
Test timeouted, sending SIGKILL!
tst_test.c:1286: INFO: If you are running on slow machine, try
exporting LTP_TIMEOUT_MUL > 1
tst_test.c:1287: BROK: Test killed! (timeout?)
cmdline="futex_wake02"
contacts=""
analysis=exit
<<<test_output>>>
tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
futex_wake02.c:155: INFO: Testing variant: syscall with kernel spec64
futex_wake02.c:105: PASS: futex_wake() woken up 1 threads
futex_wake02.c:105: PASS: futex_wake() woken up 2 threads
futex_wake02.c:105: PASS: futex_wake() woken up 3 threads
futex_wake02.c:105: PASS: futex_wake() woken up 4 threads
futex_wake02.c:105: PASS: futex_wake() woken up 5 threads
futex_wake02.c:105: PASS: futex_wake() woken up 6 threads
futex_wake02.c:105: PASS: futex_wake() woken up 7 threads
futex_wake02.c:105: PASS: futex_wake() woken up 8 threads
futex_wake02.c:105: PASS: futex_wake() woken up 9 threads
futex_wake02.c:105: PASS: futex_wake() woken up 10 threads
futex_wake02.c:118: PASS: futex_wake() woken up 0 threads
tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
futex_wake02.c:155: INFO: Testing variant: syscall time64 with kernel spec64
../../../../include/tst_timer.h:256: CONF: syscall(422)
__NR_futex_time64 not supported
../../../../include/tst_timer.h:256: CONF: syscall(422)
__NR_futex_time64 not supported
tst_test.c:1063: BROK: Test haven't reported results!
futex_wake03.c:111: INFO: Testing variant: syscall with kernel spec64
futex_wake03.c:71: PASS: futex_wake() woken up 1 childs
futex_wake03.c:71: PASS: futex_wake() woken up 2 childs
futex_wake03.c:71: PASS: futex_wake() woken up 3 childs
futex_wake03.c:71: PASS: futex_wake() woken up 4 childs
futex_wake03.c:71: PASS: futex_wake() woken up 5 childs
futex_wake03.c:71: PASS: futex_wake() woken up 6 childs
futex_wake03.c:71: PASS: futex_wake() woken up 7 childs
futex_wake03.c:71: PASS: futex_wake() woken up 8 childs
futex_wake03.c:71: PASS: futex_wake() woken up 9 childs
futex_wake03.c:71: PASS: futex_wake() woken up 10 childs
futex_wake03.c:105: PASS: futex_wake() woken up 0 children
tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
futex_wake03.c:111: INFO: Testing variant: syscall time64 with kernel spec64
../../../../include/tst_timer.h:256: CONF: syscall(422)
__NR_futex_time64 not supported
../../../../include/tst_timer.h:256: CONF: syscall(422)
__NR_futex_time64 not supported
../../../../include/tst_timer.h:256: CONF: syscall(422)
__NR_futex_time64 not supported...
../../../../include/tst_timer.h:256: CONF: syscall(422)
__NR_futex_time64 not supported
../../../../include/tst_timer.h:256: CONF: syscall(422)
__NR_futex_time64 not supported
Test timeouted, sending SIGKILL!
tst_test.c:1286: INFO: If you are running on slow machine, try
exporting LTP_TIMEOUT_MUL > 1
tst_test.c:1287: BROK: Test killed! (timeout?)
cmdline="futex_wake04"
contacts=""
analysis=exit
<<<test_output>>>
tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
futex_wake04.c:69: INFO: Testing variant: syscall with kernel spec64
futex_wake04.c:142: INFO: Hugepagesize 16777216
futex_wake04.c:127: PASS: Hi hydra, thread2 awake!
tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
futex_wake04.c:69: INFO: Testing variant: syscall time64 with kernel spec64
futex_wake04.c:142: INFO: Hugepagesize 16777216
../../../../include/tst_timer.h:256: CONF: syscall(422)
__NR_futex_time64 not supported
../../../../include/tst_timer.h:256: CONF: syscall(422)
__NR_futex_time64 not supported
futex_wake04.c:129: FAIL: Bug: wait_thread2 did not wake after 30
secs.: SUCCESS (0)
--
Regards,
Li Wang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linux.it/pipermail/ltp/attachments/20200509/a0d436a4/attachment-0001.htm>
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-08 9:24 ` Arnd Bergmann
@ 2020-05-11 5:44 ` Viresh Kumar
0 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-11 5:44 UTC (permalink / raw)
To: ltp
On 08-05-20, 11:24, Arnd Bergmann wrote:
> __NR_semtimedop can only be called with the 'old' timespec, which
> may have either 32 or 64 members depending on the architecture.
> On x32 it uses 64-bit members, and on riscv32 it does not exist at all.
>
> I think you already have a correct __kernel_old_timespec definition,
> so what I'd expect to see here is code that passes __kernel_old_timespec
> into __NR_semtimedop whenever __NR_semtimedop is defined.
Ahh, so on ARM64 and x86_64, fields will be 8 bytes long even with the old
timespec. I missed that and so the confusion.
--
viresh
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-08 7:18 ` Arnd Bergmann
2020-05-08 8:56 ` Viresh Kumar
@ 2020-05-11 6:49 ` Viresh Kumar
2020-05-11 7:20 ` Arnd Bergmann
1 sibling, 1 reply; 37+ messages in thread
From: Viresh Kumar @ 2020-05-11 6:49 UTC (permalink / raw)
To: ltp
On 08-05-20, 09:18, Arnd Bergmann wrote:
> It feels like this is more complicated than it need to be. The line
>
> semtimedop = sys_semtimedop, .type = TST_KERN_OLD_TIMESPEC, .desc =
> "semtimedop: syscall with kernel spec32"},
>
> should apply to any kernel that has "__NR_semtimedop !=
> __LTP__NR_INVALID_SYSCALL",
> regardless of any other macros set, and then you don't need the separate line
>
> { .semtimedop = sys_semtimedop, .type = TST_KERN_TIMESPEC, .desc =
> "semtimedop: syscall with kernel spec64"},
>
> which is not what the ABI is meant to be anyway (sys_semtimedop takes
> a __kernel_old_timespec,
> not a __kernel_timespec).
>
> Similarly, the line
>
> { .semop = semop, .type = TST_LIBC_TIMESPEC, .desc = "semop: vDSO or syscall"},
>
> should apply to both 32 and 64 bit machines
This is how I modified this and all other time64 tests, I hope this looks fine now ?
diff --git a/testcases/kernel/syscalls/ipc/semop/semop.h b/testcases/kernel/syscalls/ipc/semop/semop.h
index 8d1245b65ec0..8647fa686bd3 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop.h
+++ b/testcases/kernel/syscalls/ipc/semop/semop.h
@@ -26,17 +26,13 @@ struct test_variants {
char *desc;
} variants[] = {
{ .semop = semop, .type = TST_LIBC_TIMESPEC, .desc = "semop: vDSO or syscall"},
-#if defined(TST_ABI32)
- { .semtimedop = sys_semtimedop, .type = TST_LIBC_TIMESPEC, .desc = "semtimedop: syscall with libc spec"},
- { .semtimedop = sys_semtimedop, .type = TST_KERN_OLD_TIMESPEC, .desc = "semtimedop: syscall with kernel spec32"},
-#endif
-#if defined(TST_ABI64)
- { .semtimedop = sys_semtimedop, .type = TST_KERN_TIMESPEC, .desc = "semtimedop: syscall with kernel spec64"},
+#if (__NR_semtimedop != __LTP__NR_INVALID_SYSCALL)
+ { .semtimedop = sys_semtimedop, .type = TST_KERN_OLD_TIMESPEC, .desc = "semtimedop: syscall with old kernel spec"},
#endif
#if (__NR_semtimedop_time64 != __LTP__NR_INVALID_SYSCALL)
- { .semtimedop = sys_semtimedop_time64, .type = TST_KERN_TIMESPEC, .desc = "semtimedop: syscall time64 with kernel spec64"},
+ { .semtimedop = sys_semtimedop_time64, .type = TST_KERN_TIMESPEC, .desc = "semtimedop: syscall time64 with kernel spec"},
#endif
};
--
viresh
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-11 6:49 ` Viresh Kumar
@ 2020-05-11 7:20 ` Arnd Bergmann
2020-05-11 7:33 ` Viresh Kumar
0 siblings, 1 reply; 37+ messages in thread
From: Arnd Bergmann @ 2020-05-11 7:20 UTC (permalink / raw)
To: ltp
On Mon, May 11, 2020 at 8:49 AM Viresh Kumar <viresh.kumar@linaro.org> wrote:
> On 08-05-20, 09:18, Arnd Bergmann wrote:
> >
> > Similarly, the line
> >
> > { .semop = semop, .type = TST_LIBC_TIMESPEC, .desc = "semop: vDSO or syscall"},
> >
> > should apply to both 32 and 64 bit machines
>
> This is how I modified this and all other time64 tests, I hope this looks fine now ?
>
> diff --git a/testcases/kernel/syscalls/ipc/semop/semop.h b/testcases/kernel/syscalls/ipc/semop/semop.h
> index 8d1245b65ec0..8647fa686bd3 100644
> --- a/testcases/kernel/syscalls/ipc/semop/semop.h
> +++ b/testcases/kernel/syscalls/ipc/semop/semop.h
> @@ -26,17 +26,13 @@ struct test_variants {
> char *desc;
> } variants[] = {
> { .semop = semop, .type = TST_LIBC_TIMESPEC, .desc = "semop: vDSO or syscall"},
> -#if defined(TST_ABI32)
> - { .semtimedop = sys_semtimedop, .type = TST_LIBC_TIMESPEC, .desc = "semtimedop: syscall with libc spec"},
> - { .semtimedop = sys_semtimedop, .type = TST_KERN_OLD_TIMESPEC, .desc = "semtimedop: syscall with kernel spec32"},
> -#endif
>
> -#if defined(TST_ABI64)
> - { .semtimedop = sys_semtimedop, .type = TST_KERN_TIMESPEC, .desc = "semtimedop: syscall with kernel spec64"},
> +#if (__NR_semtimedop != __LTP__NR_INVALID_SYSCALL)
> + { .semtimedop = sys_semtimedop, .type = TST_KERN_OLD_TIMESPEC, .desc = "semtimedop: syscall with old kernel spec"},
> #endif
>
> #if (__NR_semtimedop_time64 != __LTP__NR_INVALID_SYSCALL)
> - { .semtimedop = sys_semtimedop_time64, .type = TST_KERN_TIMESPEC, .desc = "semtimedop: syscall time64 with kernel spec64"},
> + { .semtimedop = sys_semtimedop_time64, .type = TST_KERN_TIMESPEC, .desc = "semtimedop: syscall time64 with kernel spec"},
> #endif
> };
Yes, much better. Please do it like this for all the tests.
It looks like you are still missing a test for the libc semtimedop function with
TST_LIBC_TIMESPEC, but perhaps this is in a different place?
There is one more complication regarding semtimedop, and that is how some
architectures never had this as a separate syscall but instead require calling
sys_ipc(SEMTIMEDOP, ...). All architectures that have sys_ipc() let you use
both SEMOP and SEMTIMEDOP with __kernel_old_timespec. Some
architectures define more than one of __NR_ipc, __NR_semtimedop and
__NR_semop.
As of linux-5.1, all 64-bit architectures at least have __NR_semtimedop, and
all 32-bit architectures at least have __NR_semtimedop64, plus whatever
other variants they supported before then.
Arnd
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-11 7:20 ` Arnd Bergmann
@ 2020-05-11 7:33 ` Viresh Kumar
2020-05-11 7:58 ` Arnd Bergmann
0 siblings, 1 reply; 37+ messages in thread
From: Viresh Kumar @ 2020-05-11 7:33 UTC (permalink / raw)
To: ltp
On 11-05-20, 09:20, Arnd Bergmann wrote:
> It looks like you are still missing a test for the libc semtimedop function with
> TST_LIBC_TIMESPEC, but perhaps this is in a different place?
The libc test is missing for few of the syscalls, as it was already missing for
them because the headers may not have the definition.
And so I didn't add them separately.
--
viresh
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-11 7:33 ` Viresh Kumar
@ 2020-05-11 7:58 ` Arnd Bergmann
2020-05-11 11:23 ` Viresh Kumar
0 siblings, 1 reply; 37+ messages in thread
From: Arnd Bergmann @ 2020-05-11 7:58 UTC (permalink / raw)
To: ltp
On Mon, May 11, 2020 at 9:33 AM Viresh Kumar <viresh.kumar@linaro.org> wrote:
>
> On 11-05-20, 09:20, Arnd Bergmann wrote:
> > It looks like you are still missing a test for the libc semtimedop function with
> > TST_LIBC_TIMESPEC, but perhaps this is in a different place?
>
> The libc test is missing for few of the syscalls, as it was already missing for
> them because the headers may not have the definition.
>
> And so I didn't add them separately.
Ok, I see. This one just looked curious because the you add tests for the kernel
syscall (semtimedop) in a file that checks an older libc syscall (semop).
The libc semop is implemented in terms of the kernel's semtimedop for
architectures that don't have their own semop by passing a NULL timeout
argument. We should still check that five variants of it (sys_semop,
sys_semtimedop, sys_semtimedop64, ipc(SEMTIMEDOP) and ipc(SEMOP))
correctly implement the semop semantics, but I would also note that
testing the differences of the time types is only meaningful when you
actually pass a non-NULL timeout.
Arnd
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 16/17] syscalls/utimensat: Migrate to new test framework
2020-05-08 4:24 ` [LTP] [PATCH V2 16/17] syscalls/utimensat: Migrate to new test framework Viresh Kumar
2020-05-09 3:50 ` Li Wang
@ 2020-05-11 9:33 ` Cyril Hrubis
1 sibling, 0 replies; 37+ messages in thread
From: Cyril Hrubis @ 2020-05-11 9:33 UTC (permalink / raw)
To: ltp
Hi!
> + if (tc->rattr != NULL)
> + run_command("chattr", tc->rattr, tc->pathname);
Can we please use the FS_IOC_GETFLAGS and FS_IOC_SETFLAGS instead?
Have a look at fallocate06.c function toggle_cow()
--
Cyril Hrubis
chrubis@suse.cz
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-09 4:21 ` Li Wang
@ 2020-05-11 10:43 ` Viresh Kumar
2020-05-11 12:29 ` Li Wang
0 siblings, 1 reply; 37+ messages in thread
From: Viresh Kumar @ 2020-05-11 10:43 UTC (permalink / raw)
To: ltp
On 09-05-20, 12:21, Li Wang wrote:
> RHEL8.2(x84_64) failed as:
>
> # ./semop05
> tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
> semop05.c:147: INFO: Testing variant: semop: vDSO or syscall
> semop05.c:132: PASS: expected failure: EIDRM (43)
> semop05.c:132: PASS: expected failure: EIDRM (43)
> semop05.c:132: PASS: expected failure: EINTR (4)
> semop05.c:132: PASS: expected failure: EINTR (4)
> tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
> semop05.c:147: INFO: Testing variant: semtimedop: syscall with kernel spec64
> semop05.c:134: FAIL: unexpected failure: EAGAIN/EWOULDBLOCK (11)
> Test timeouted, sending SIGKILL!
> tst_test.c:1286: INFO: If you are running on slow machine, try exporting
> LTP_TIMEOUT_MUL > 1
> tst_test.c:1287: BROK: Test killed! (timeout?)
>
>
> RHEL8.2(s390x) failed as:
>
> tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
> semop05.c:147: INFO: Testing variant: semop: vDSO or syscall
> semop05.c:132: PASS: expected failure: EIDRM (43)
> semop05.c:132: PASS: expected failure: EIDRM (43)
> semop05.c:132: PASS: expected failure: EINTR (4)
> semop05.c:132: PASS: expected failure: EINTR (4)
> tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
> semop05.c:147: INFO: Testing variant: semtimedop: syscall with kernel spec64
> semop.h:13: CONF: syscall(392) __NR_semtimedop not supported
> Test timeouted, sending SIGKILL!
> tst_test.c:1286: INFO: If you are running on slow machine, try exporting
> LTP_TIMEOUT_MUL > 1
> tst_test.c:1287: BROK: Test killed! (timeout?)
Some part of the timeout was left uninitialized and so was causing
some issues. Can you please try with this delta:
diff --git a/testcases/kernel/syscalls/ipc/semop/semop05.c b/testcases/kernel/syscalls/ipc/semop/semop05.c
index f892cd519ecf..0e18f77ba1f5 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop05.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop05.c
@@ -120,6 +120,7 @@ static void do_child(int i)
struct tst_ts timeout;
timeout.type = tv->type;
+ tst_ts_set_sec(&timeout, 1);
tst_ts_set_nsec(&timeout, 10000);
TEST(call_semop(tv, sem_id, &s_buf, 1, &timeout));
--
viresh
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 06/17] syscalls/futex: Add support for time64 tests
2020-05-09 9:02 ` Li Wang
@ 2020-05-11 11:08 ` Viresh Kumar
0 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-11 11:08 UTC (permalink / raw)
To: ltp
On 09-05-20, 17:02, Li Wang wrote:
> This might not safe to use it directly if only check __NR_futex_time64 being
> defined. i.e Test it on the platform without achieving the system
> syscall(but
> define __NR_futex_time64), the test children exit with TCONF will cause
> kind of issues.
I will add this to the patch then, I am wondering if I need to add
this to all tests now ? Cyril ?
diff --git a/testcases/kernel/syscalls/futex/futex_cmp_requeue01.c b/testcases/kernel/syscalls/futex/futex_cmp_requeue01.c
index efde3410f564..537641766357 100644
--- a/testcases/kernel/syscalls/futex/futex_cmp_requeue01.c
+++ b/testcases/kernel/syscalls/futex/futex_cmp_requeue01.c
@@ -194,7 +194,11 @@ static void verify_futex_cmp_requeue(unsigned int n)
static void setup(void)
{
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ futex_supported_by_kernel(tv->fntype);
+
max_sleep_ms = tst_multiply_timeout(5000);
sd = SAFE_MMAP(NULL, sizeof(*sd), PROT_READ | PROT_WRITE,
diff --git a/testcases/kernel/syscalls/futex/futex_cmp_requeue02.c b/testcases/kernel/syscalls/futex/futex_cmp_requeue02.c
index e1af5ddab509..dd8fafb3ee38 100644
--- a/testcases/kernel/syscalls/futex/futex_cmp_requeue02.c
+++ b/testcases/kernel/syscalls/futex/futex_cmp_requeue02.c
@@ -69,7 +69,11 @@ static void verify_futex_cmp_requeue(unsigned int n)
static void setup(void)
{
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ futex_supported_by_kernel(tv->fntype);
+
futexes = SAFE_MMAP(NULL, sizeof(futex_t) * 2, PROT_READ | PROT_WRITE,
MAP_ANONYMOUS | MAP_SHARED, -1, 0);
diff --git a/testcases/kernel/syscalls/futex/futex_wait01.c b/testcases/kernel/syscalls/futex/futex_wait01.c
index 41b21a71f2c6..e884e696ef55 100644
--- a/testcases/kernel/syscalls/futex/futex_wait01.c
+++ b/testcases/kernel/syscalls/futex/futex_wait01.c
@@ -74,7 +74,10 @@ static void run(unsigned int n)
static void setup(void)
{
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ futex_supported_by_kernel(tv->fntype);
}
static struct tst_test test = {
diff --git a/testcases/kernel/syscalls/futex/futex_wait02.c b/testcases/kernel/syscalls/futex/futex_wait02.c
index 7b19115073dd..34b78357f6e6 100644
--- a/testcases/kernel/syscalls/futex/futex_wait02.c
+++ b/testcases/kernel/syscalls/futex/futex_wait02.c
@@ -74,7 +74,10 @@ static void run(void)
static void setup(void)
{
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ futex_supported_by_kernel(tv->fntype);
futex = SAFE_MMAP(NULL, sizeof(*futex), PROT_READ | PROT_WRITE,
MAP_ANONYMOUS | MAP_SHARED, -1, 0);
diff --git a/testcases/kernel/syscalls/futex/futex_wait03.c b/testcases/kernel/syscalls/futex/futex_wait03.c
index 83601479adc0..e436b238a165 100644
--- a/testcases/kernel/syscalls/futex/futex_wait03.c
+++ b/testcases/kernel/syscalls/futex/futex_wait03.c
@@ -73,7 +73,10 @@ static void run(void)
static void setup(void)
{
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ futex_supported_by_kernel(tv->fntype);
}
static struct tst_test test = {
diff --git a/testcases/kernel/syscalls/futex/futex_wait04.c b/testcases/kernel/syscalls/futex/futex_wait04.c
index ba6988a20e98..8de3c90b08d3 100644
--- a/testcases/kernel/syscalls/futex/futex_wait04.c
+++ b/testcases/kernel/syscalls/futex/futex_wait04.c
@@ -53,7 +53,10 @@ static void run(void)
static void setup(void)
{
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ futex_supported_by_kernel(tv->fntype);
}
static struct tst_test test = {
diff --git a/testcases/kernel/syscalls/futex/futex_wait_bitset01.c b/testcases/kernel/syscalls/futex/futex_wait_bitset01.c
index 9f2f4025d4a4..a871af992572 100644
--- a/testcases/kernel/syscalls/futex/futex_wait_bitset01.c
+++ b/testcases/kernel/syscalls/futex/futex_wait_bitset01.c
@@ -102,7 +102,10 @@ static void run(unsigned int n)
static void setup(void)
{
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ futex_supported_by_kernel(tv->fntype);
}
static struct tst_test test = {
diff --git a/testcases/kernel/syscalls/futex/futex_wake01.c b/testcases/kernel/syscalls/futex/futex_wake01.c
index 21cf158daac7..2dbe8cc2f3aa 100644
--- a/testcases/kernel/syscalls/futex/futex_wake01.c
+++ b/testcases/kernel/syscalls/futex/futex_wake01.c
@@ -58,7 +58,10 @@ static void run(unsigned int n)
static void setup(void)
{
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ futex_supported_by_kernel(tv->fntype);
}
static struct tst_test test = {
diff --git a/testcases/kernel/syscalls/futex/futex_wake02.c b/testcases/kernel/syscalls/futex/futex_wake02.c
index cd535163a6e9..3491839872ac 100644
--- a/testcases/kernel/syscalls/futex/futex_wake02.c
+++ b/testcases/kernel/syscalls/futex/futex_wake02.c
@@ -147,7 +147,10 @@ static void run(void)
static void setup(void)
{
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ futex_supported_by_kernel(tv->fntype);
}
static struct tst_test test = {
diff --git a/testcases/kernel/syscalls/futex/futex_wake03.c b/testcases/kernel/syscalls/futex/futex_wake03.c
index b059cf6fa272..29e586c40132 100644
--- a/testcases/kernel/syscalls/futex/futex_wake03.c
+++ b/testcases/kernel/syscalls/futex/futex_wake03.c
@@ -103,7 +103,10 @@ static void run(void)
static void setup(void)
{
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ futex_supported_by_kernel(tv->fntype);
futex = SAFE_MMAP(NULL, sizeof(*futex), PROT_READ | PROT_WRITE,
MAP_ANONYMOUS | MAP_SHARED, -1, 0);
diff --git a/testcases/kernel/syscalls/futex/futex_wake04.c b/testcases/kernel/syscalls/futex/futex_wake04.c
index 2f58efc1eb2a..edbb9c644c57 100644
--- a/testcases/kernel/syscalls/futex/futex_wake04.c
+++ b/testcases/kernel/syscalls/futex/futex_wake04.c
@@ -62,6 +62,7 @@ static void setup(void)
struct test_variants *tv = &variants[tst_variant];
tst_res(TINFO, "Testing variant: %s", tv->desc);
+ futex_supported_by_kernel(tv->fntype);
to.type = tv->tstype;
tst_ts_set_sec(&to, 30);
diff --git a/testcases/kernel/syscalls/futex/futextest.h b/testcases/kernel/syscalls/futex/futextest.h
index a9dcbbedd249..816f2f1f4689 100644
--- a/testcases/kernel/syscalls/futex/futextest.h
+++ b/testcases/kernel/syscalls/futex/futextest.h
@@ -88,6 +88,17 @@ enum futex_fn_type {
FUTEX_FN_FUTEX64,
};
+static inline void futex_supported_by_kernel(enum futex_fn_type fntype)
+{
+ if (fntype != FUTEX_FN_FUTEX64)
+ return;
+
+ /* Check if the syscall is backported on an older kernel */
+ TEST(sys_futex_time64(NULL, 0, 0, NULL, NULL, 0));
+ if (TST_RET == -1 && TST_ERR == ENOSYS)
+ tst_brk(TCONF, "Test not supported on kernel");
+}
+
/**
* futex_syscall() - futex syscall wrapper
* @fntype: Futex function type
--
viresh
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-11 7:58 ` Arnd Bergmann
@ 2020-05-11 11:23 ` Viresh Kumar
0 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-11 11:23 UTC (permalink / raw)
To: ltp
On 11-05-20, 09:58, Arnd Bergmann wrote:
> We should still check that five variants of it (sys_semop,
> sys_semtimedop, sys_semtimedop64, ipc(SEMTIMEDOP) and ipc(SEMOP))
> correctly implement the semop semantics
How do you suggest I add tests for ipc() types? Should they be always
tested? Or they should be tested only if the normal sys_* variant
isn't implemented by the kernel/arch ?
--
viresh
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-11 10:43 ` Viresh Kumar
@ 2020-05-11 12:29 ` Li Wang
2020-05-12 4:19 ` Viresh Kumar
0 siblings, 1 reply; 37+ messages in thread
From: Li Wang @ 2020-05-11 12:29 UTC (permalink / raw)
To: ltp
On Mon, May 11, 2020 at 6:43 PM Viresh Kumar <viresh.kumar@linaro.org>
wrote:
> On 09-05-20, 12:21, Li Wang wrote:
> > RHEL8.2(x84_64) failed as:
> >
> > # ./semop05
> > tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
> > semop05.c:147: INFO: Testing variant: semop: vDSO or syscall
> > semop05.c:132: PASS: expected failure: EIDRM (43)
> > semop05.c:132: PASS: expected failure: EIDRM (43)
> > semop05.c:132: PASS: expected failure: EINTR (4)
> > semop05.c:132: PASS: expected failure: EINTR (4)
> > tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
> > semop05.c:147: INFO: Testing variant: semtimedop: syscall with kernel
> spec64
> > semop05.c:134: FAIL: unexpected failure: EAGAIN/EWOULDBLOCK (11)
> > Test timeouted, sending SIGKILL!
> > tst_test.c:1286: INFO: If you are running on slow machine, try exporting
> > LTP_TIMEOUT_MUL > 1
> > tst_test.c:1287: BROK: Test killed! (timeout?)
> >
> >
> > RHEL8.2(s390x) failed as:
> >
> > tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
> > semop05.c:147: INFO: Testing variant: semop: vDSO or syscall
> > semop05.c:132: PASS: expected failure: EIDRM (43)
> > semop05.c:132: PASS: expected failure: EIDRM (43)
> > semop05.c:132: PASS: expected failure: EINTR (4)
> > semop05.c:132: PASS: expected failure: EINTR (4)
> > tst_test.c:1246: INFO: Timeout per run is 0h 05m 00s
> > semop05.c:147: INFO: Testing variant: semtimedop: syscall with kernel
> spec64
> > semop.h:13: CONF: syscall(392) __NR_semtimedop not supported
> > Test timeouted, sending SIGKILL!
> > tst_test.c:1286: INFO: If you are running on slow machine, try exporting
> > LTP_TIMEOUT_MUL > 1
> > tst_test.c:1287: BROK: Test killed! (timeout?)
>
> Some part of the timeout was left uninitialized and so was causing
> some issues. Can you please try with this delta:
>
This delta works(on x86_64) for me. But the s390x issue looks like the same
as the futex TCONF in another email.
--
Regards,
Li Wang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linux.it/pipermail/ltp/attachments/20200511/67b4f9ab/attachment-0001.htm>
^ permalink raw reply [flat|nested] 37+ messages in thread
* [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version
2020-05-11 12:29 ` Li Wang
@ 2020-05-12 4:19 ` Viresh Kumar
0 siblings, 0 replies; 37+ messages in thread
From: Viresh Kumar @ 2020-05-12 4:19 UTC (permalink / raw)
To: ltp
On 11-05-20, 20:29, Li Wang wrote:
> This delta works(on x86_64) for me. But the s390x issue looks like the same
> as the futex TCONF in another email.
Ah, this should fix it then:
diff --git a/testcases/kernel/syscalls/ipc/semop/semop.h b/testcases/kernel/syscalls/ipc/semop/semop.h
index 7180f6e35185..584d12c68e0d 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop.h
+++ b/testcases/kernel/syscalls/ipc/semop/semop.h
@@ -44,4 +44,12 @@ static inline int call_semop(struct test_variants *tv, int semid,
return tv->semtimedop(semid, sops, nsops, tst_ts_get(timeout));
}
+static inline void semop_supported_by_kernel(struct test_variants *tv)
+{
+ /* Check if the syscall is implemented on the platform */
+ TEST(call_semop(tv, 0, NULL, 0, NULL));
+ if (TST_RET == -1 && TST_ERR == ENOSYS)
+ tst_brk(TCONF, "Test not supported on kernel/platform");
+}
+
#endif /* SEMOP_VAR__ */
diff --git a/testcases/kernel/syscalls/ipc/semop/semop01.c b/testcases/kernel/syscalls/ipc/semop/semop01.c
index cc8725bdbed3..7947183c837b 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop01.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop01.c
@@ -99,9 +99,11 @@ static void run(void)
static void setup(void)
{
+ struct test_variants *tv = &variants[tst_variant];
int i;
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ semop_supported_by_kernel(tv);
get_arr.array = malloc(sizeof(unsigned short int) * PSEMS);
if (get_arr.array == NULL)
diff --git a/testcases/kernel/syscalls/ipc/semop/semop02.c b/testcases/kernel/syscalls/ipc/semop/semop02.c
index 5af1909480b5..4ffb9dd72c1c 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop02.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop02.c
@@ -47,13 +47,15 @@ static struct test_case_t {
static void setup(void)
{
+ struct test_variants *tv = &variants[tst_variant];
char nobody_uid[] = "nobody";
struct passwd *ltpuser;
key_t semkey2;
struct seminfo ipc_buf;
union semun arr;
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ semop_supported_by_kernel(tv);
ltpuser = SAFE_GETPWNAM(nobody_uid);
SAFE_SETUID(ltpuser->pw_uid);
diff --git a/testcases/kernel/syscalls/ipc/semop/semop03.c b/testcases/kernel/syscalls/ipc/semop/semop03.c
index e877f989edaa..40a6b095c57d 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop03.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop03.c
@@ -88,7 +88,10 @@ static void run(unsigned int i)
static void setup(void)
{
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ semop_supported_by_kernel(tv);
/* get an IPC resource key */
semkey = GETIPCKEY();
diff --git a/testcases/kernel/syscalls/ipc/semop/semop04.c b/testcases/kernel/syscalls/ipc/semop/semop04.c
index 226a64603943..5e4bfeb625e2 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop04.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop04.c
@@ -94,7 +94,10 @@ static void run(unsigned int i)
static void setup(void)
{
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ semop_supported_by_kernel(tv);
val = 1;
diff --git a/testcases/kernel/syscalls/ipc/semop/semop05.c b/testcases/kernel/syscalls/ipc/semop/semop05.c
index 0e18f77ba1f5..770646bc564c 100644
--- a/testcases/kernel/syscalls/ipc/semop/semop05.c
+++ b/testcases/kernel/syscalls/ipc/semop/semop05.c
@@ -145,7 +145,10 @@ static void sighandler(int sig)
static void setup(void)
{
- tst_res(TINFO, "Testing variant: %s", variants[tst_variant].desc);
+ struct test_variants *tv = &variants[tst_variant];
+
+ tst_res(TINFO, "Testing variant: %s", tv->desc);
+ semop_supported_by_kernel(tv);
SAFE_SIGNAL(SIGHUP, sighandler);
--
viresh
^ permalink raw reply related [flat|nested] 37+ messages in thread
end of thread, other threads:[~2020-05-12 4:19 UTC | newest]
Thread overview: 37+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-05-08 4:23 [LTP] [PATCH V2 00/17] Syscalls: Add support for time64 variants Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 01/17] syscalls/timer_gettime: Add support for time64 tests Viresh Kumar
2020-05-08 8:30 ` Arnd Bergmann
2020-05-08 4:23 ` [LTP] [PATCH V2 02/17] syscalls/timer_settime: " Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 03/17] syscalls/timerfd: " Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 04/17] syscalls/sched_rr_get_interval: " Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 05/17] syscalls/futex: Merge futex_wait_bitset tests Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 06/17] syscalls/futex: Add support for time64 tests Viresh Kumar
2020-05-09 9:02 ` Li Wang
2020-05-11 11:08 ` Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 07/17] syscalls/io_pgetevents: " Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 08/17] syscalls/sigwaitinfo: Migrate to new test framework Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 09/17] syscalls/rt_sigtimedwait: Add support for time64 tests Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 10/17] syscalls/mq_timed{send|receive}: " Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 11/17] syscalls/recvmmsg: " Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 12/17] syscalls/ppoll: " Viresh Kumar
2020-05-08 4:23 ` [LTP] [PATCH V2 13/17] syscalls/select6: " Viresh Kumar
2020-05-08 7:21 ` Arnd Bergmann
2020-05-08 4:23 ` [LTP] [PATCH V2 14/17] syscalls/semop: Migrate to new test framework Viresh Kumar
2020-05-08 4:24 ` [LTP] [PATCH V2 15/17] syscalls/semtimedop: Add support for semtimedop and its time64 version Viresh Kumar
2020-05-08 7:18 ` Arnd Bergmann
2020-05-08 8:56 ` Viresh Kumar
2020-05-08 9:24 ` Arnd Bergmann
2020-05-11 5:44 ` Viresh Kumar
2020-05-11 6:49 ` Viresh Kumar
2020-05-11 7:20 ` Arnd Bergmann
2020-05-11 7:33 ` Viresh Kumar
2020-05-11 7:58 ` Arnd Bergmann
2020-05-11 11:23 ` Viresh Kumar
2020-05-09 4:21 ` Li Wang
2020-05-11 10:43 ` Viresh Kumar
2020-05-11 12:29 ` Li Wang
2020-05-12 4:19 ` Viresh Kumar
2020-05-08 4:24 ` [LTP] [PATCH V2 16/17] syscalls/utimensat: Migrate to new test framework Viresh Kumar
2020-05-09 3:50 ` Li Wang
2020-05-11 9:33 ` Cyril Hrubis
2020-05-08 4:24 ` [LTP] [PATCH V2 17/17] syscalls/utimensat: Add support for time64 tests Viresh Kumar
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.