452 lines
11 KiB
C
Raw Normal View History

/* SPDX-License-Identifier: GPL-2.0 */
/*
* kselftest.h: low-level kselftest framework to include from
* selftest programs. When possible, please use
* kselftest_harness.h instead.
*
* Copyright (c) 2014 Shuah Khan <shuahkh@osg.samsung.com>
* Copyright (c) 2014 Samsung Electronics Co., Ltd.
*
* Using this API consists of first counting how many tests your code
* has to run, and then starting up the reporting:
*
* ksft_print_header();
* ksft_set_plan(total_number_of_tests);
*
* For each test, report any progress, debugging, etc with:
*
* ksft_print_msg(fmt, ...);
* ksft_perror(msg);
*
* and finally report the pass/fail/skip/xfail state of the test with one of:
*
* ksft_test_result(condition, fmt, ...);
* ksft_test_result_report(result, fmt, ...);
* ksft_test_result_pass(fmt, ...);
* ksft_test_result_fail(fmt, ...);
* ksft_test_result_skip(fmt, ...);
* ksft_test_result_xfail(fmt, ...);
* ksft_test_result_error(fmt, ...);
* ksft_test_result_code(exit_code, test_name, fmt, ...);
*
* When all tests are finished, clean up and exit the program with one of:
*
* ksft_finished();
* ksft_exit(condition);
* ksft_exit_pass();
* ksft_exit_fail();
*
* If the program wants to report details on why the entire program has
* failed, it can instead exit with a message (this is usually done when
* the program is aborting before finishing all tests):
*
* ksft_exit_fail_msg(fmt, ...);
* ksft_exit_fail_perror(msg);
*
*/
#ifndef __KSELFTEST_H
#define __KSELFTEST_H
#ifndef NOLIBC
#include <errno.h>
#include <stdlib.h>
#include <unistd.h>
#include <stdarg.h>
#include <string.h>
#include <stdio.h>
selftests/timers/posix_timers: Reimplement check_timer_distribution() check_timer_distribution() runs ten threads in a busy loop and tries to test that the kernel distributes a process posix CPU timer signal to every thread over time. There is not guarantee that this is true even after commit bcb7ee79029d ("posix-timers: Prefer delivery of signals to the current thread") because that commit only avoids waking up the sleeping process leader thread, but that has nothing to do with the actual signal delivery. As the signal is process wide the first thread which observes sigpending and wins the race to lock sighand will deliver the signal. Testing shows that this hangs on a regular base because some threads never win the race. The comment "This primarily tests that the kernel does not favour any one." is wrong. The kernel does favour a thread which hits the timer interrupt when CLOCK_PROCESS_CPUTIME_ID expires. Rewrite the test so it only checks that the group leader sleeping in join() never receives SIGALRM and the thread which burns CPU cycles receives all signals. In older kernels which do not have commit bcb7ee79029d ("posix-timers: Prefer delivery of signals to the current thread") the test-case fails immediately, the very 1st tick wakes the leader up. Otherwise it quickly succeeds after 100 ticks. CI testing wants to use newer selftest versions on stable kernels. In this case the test is guaranteed to fail. So check in the failure case whether the kernel version is less than v6.3 and skip the test result in that case. [ tglx: Massaged change log, renamed the version check helper ] Fixes: e797203fb3ba ("selftests/timers/posix_timers: Test delivery of signals across threads") Signed-off-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20240409133802.GD29396@redhat.com
2024-04-09 15:38:03 +02:00
#include <sys/utsname.h>
#endif
#ifndef ARRAY_SIZE
#define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]))
#endif
/*
* gcc cpuid.h provides __cpuid_count() since v4.4.
* Clang/LLVM cpuid.h provides __cpuid_count() since v3.4.0.
*
* Provide local define for tests needing __cpuid_count() because
* selftests need to work in older environments that do not yet
* have __cpuid_count().
*/
#ifndef __cpuid_count
#define __cpuid_count(level, count, a, b, c, d) \
__asm__ __volatile__ ("cpuid\n\t" \
: "=a" (a), "=b" (b), "=c" (c), "=d" (d) \
: "0" (level), "2" (count))
#endif
/* define kselftest exit codes */
#define KSFT_PASS 0
#define KSFT_FAIL 1
#define KSFT_XFAIL 2
#define KSFT_XPASS 3
#define KSFT_SKIP 4
selftests: kselftest: Mark functions that unconditionally call exit() as __noreturn After commit 6d029c25b71f ("selftests/timers/posix_timers: Reimplement check_timer_distribution()"), clang warns: tools/testing/selftests/timers/../kselftest.h:398:6: warning: variable 'major' is used uninitialized whenever '||' condition is true [-Wsometimes-uninitialized] 398 | if (uname(&info) || sscanf(info.release, "%u.%u.", &major, &minor) != 2) | ^~~~~~~~~~~~ tools/testing/selftests/timers/../kselftest.h:401:9: note: uninitialized use occurs here 401 | return major > min_major || (major == min_major && minor >= min_minor); | ^~~~~ tools/testing/selftests/timers/../kselftest.h:398:6: note: remove the '||' if its condition is always false 398 | if (uname(&info) || sscanf(info.release, "%u.%u.", &major, &minor) != 2) | ^~~~~~~~~~~~~~~ tools/testing/selftests/timers/../kselftest.h:395:20: note: initialize the variable 'major' to silence this warning 395 | unsigned int major, minor; | ^ | = 0 This is a false positive because if uname() fails, ksft_exit_fail_msg() will be called, which unconditionally calls exit(), a noreturn function. However, clang does not know that ksft_exit_fail_msg() will call exit() at the point in the pipeline that the warning is emitted because inlining has not occurred, so it assumes control flow will resume normally after ksft_exit_fail_msg() is called. Make it clear to clang that all of the functions that call exit() unconditionally in kselftest.h are noreturn transitively by marking them explicitly with '__attribute__((__noreturn__))', which clears up the warning above and any future warnings that may appear for the same reason. Fixes: 6d029c25b71f ("selftests/timers/posix_timers: Reimplement check_timer_distribution()") Reported-by: John Stultz <jstultz@google.com> Signed-off-by: Nathan Chancellor <nathan@kernel.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Shuah Khan <skhan@linuxfoundation.org> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20240411-mark-kselftest-exit-funcs-noreturn-v1-1-b027c948f586@kernel.org Closes: https://lore.kernel.org/all/20240410232637.4135564-2-jstultz@google.com/
2024-04-11 11:45:40 -07:00
#ifndef __noreturn
#define __noreturn __attribute__((__noreturn__))
#endif
#define __printf(a, b) __attribute__((format(printf, a, b)))
/* counters */
struct ksft_count {
unsigned int ksft_pass;
unsigned int ksft_fail;
unsigned int ksft_xfail;
unsigned int ksft_xpass;
unsigned int ksft_xskip;
unsigned int ksft_error;
};
static struct ksft_count ksft_cnt;
static unsigned int ksft_plan;
static inline unsigned int ksft_test_num(void)
{
return ksft_cnt.ksft_pass + ksft_cnt.ksft_fail +
ksft_cnt.ksft_xfail + ksft_cnt.ksft_xpass +
ksft_cnt.ksft_xskip + ksft_cnt.ksft_error;
}
static inline void ksft_inc_pass_cnt(void) { ksft_cnt.ksft_pass++; }
static inline void ksft_inc_fail_cnt(void) { ksft_cnt.ksft_fail++; }
static inline void ksft_inc_xfail_cnt(void) { ksft_cnt.ksft_xfail++; }
static inline void ksft_inc_xpass_cnt(void) { ksft_cnt.ksft_xpass++; }
static inline void ksft_inc_xskip_cnt(void) { ksft_cnt.ksft_xskip++; }
static inline void ksft_inc_error_cnt(void) { ksft_cnt.ksft_error++; }
static inline int ksft_get_pass_cnt(void) { return ksft_cnt.ksft_pass; }
static inline int ksft_get_fail_cnt(void) { return ksft_cnt.ksft_fail; }
static inline int ksft_get_xfail_cnt(void) { return ksft_cnt.ksft_xfail; }
static inline int ksft_get_xpass_cnt(void) { return ksft_cnt.ksft_xpass; }
static inline int ksft_get_xskip_cnt(void) { return ksft_cnt.ksft_xskip; }
static inline int ksft_get_error_cnt(void) { return ksft_cnt.ksft_error; }
static inline void ksft_print_header(void)
{
selftests: line buffer test program's stdout Patch series "selftests/mm fixes for arm64", v3. Given my on-going work on large anon folios and contpte mappings, I decided it would be a good idea to start running mm selftests to help guard against regressions. However, it soon became clear that I couldn't get the suite to run cleanly on arm64 with a vanilla v6.5-rc1 kernel (perhaps I'm just doing it wrong??), so got stuck in a rabbit hole trying to debug and fix all the issues. Some were down to misconfigurations, but I also found a number of issues with the tests and even a couple of issues with the kernel. This patch (of 8): The selftests runner pipes the test program's stdout to tap_prefix. The presence of the pipe means that the test program sets its stdout to be fully buffered (as aposed to line buffered when directly connected to the terminal). The block buffering means that there is often content in the buffer at fork() time, which causes the output to end up duplicated. This was causing problems for mm:cow where test results were duplicated 20-30x. Solve this by using `stdbuf`, when available to force the test program to use line buffered mode. This means previously printf'ed results are flushed out of the program before any fork(). Additionally, explicitly set line buffer mode in ksft_print_header(), which means that all test programs that use the ksft framework will benefit even if stdbuf is not present on the system. [ryan.roberts@arm.com: add setvbuf() to set buffering mode] Link: https://lkml.kernel.org/r/20230726070655.2713530-1-ryan.roberts@arm.com Link: https://lkml.kernel.org/r/20230724082522.1202616-1-ryan.roberts@arm.com Link: https://lkml.kernel.org/r/20230724082522.1202616-2-ryan.roberts@arm.com Signed-off-by: Ryan Roberts <ryan.roberts@arm.com> Reviewed-by: Mark Brown <broonie@kernel.org> Cc: David Hildenbrand <david@redhat.com> Cc: Florent Revest <revest@chromium.org> Cc: Jérôme Glisse <jglisse@redhat.com> Cc: John Hubbard <jhubbard@nvidia.com> Cc: Peter Xu <peterx@redhat.com> Cc: Shuah Khan <shuah@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
2023-07-24 09:25:15 +01:00
/*
* Force line buffering; If stdout is not connected to a terminal, it
* will otherwise default to fully buffered, which can cause output
* duplication if there is content in the buffer when fork()ing. If
* there is a crash, line buffering also means the most recent output
* line will be visible.
*/
setvbuf(stdout, NULL, _IOLBF, 0);
if (!(getenv("KSFT_TAP_LEVEL")))
printf("TAP version 13\n");
}
static inline void ksft_set_plan(unsigned int plan)
{
ksft_plan = plan;
printf("1..%u\n", ksft_plan);
}
static inline void ksft_print_cnts(void)
{
if (ksft_plan != ksft_test_num())
printf("# Planned tests != run tests (%u != %u)\n",
ksft_plan, ksft_test_num());
printf("# Totals: pass:%u fail:%u xfail:%u xpass:%u skip:%u error:%u\n",
ksft_cnt.ksft_pass, ksft_cnt.ksft_fail,
ksft_cnt.ksft_xfail, ksft_cnt.ksft_xpass,
ksft_cnt.ksft_xskip, ksft_cnt.ksft_error);
}
static inline __printf(1, 2) void ksft_print_msg(const char *msg, ...)
{
int saved_errno = errno;
va_list args;
va_start(args, msg);
printf("# ");
errno = saved_errno;
vprintf(msg, args);
va_end(args);
}
static inline void ksft_perror(const char *msg)
{
#ifndef NOLIBC
ksft_print_msg("%s: %s (%d)\n", msg, strerror(errno), errno);
#else
/*
* nolibc doesn't provide strerror() and it seems
* inappropriate to add one, just print the errno.
*/
ksft_print_msg("%s: %d)\n", msg, errno);
#endif
}
static inline __printf(1, 2) void ksft_test_result_pass(const char *msg, ...)
{
int saved_errno = errno;
va_list args;
ksft_cnt.ksft_pass++;
va_start(args, msg);
printf("ok %u ", ksft_test_num());
errno = saved_errno;
vprintf(msg, args);
va_end(args);
}
static inline __printf(1, 2) void ksft_test_result_fail(const char *msg, ...)
{
int saved_errno = errno;
va_list args;
ksft_cnt.ksft_fail++;
va_start(args, msg);
printf("not ok %u ", ksft_test_num());
errno = saved_errno;
vprintf(msg, args);
va_end(args);
}
/**
* ksft_test_result() - Report test success based on truth of condition
*
* @condition: if true, report test success, otherwise failure.
*/
#define ksft_test_result(condition, fmt, ...) do { \
if (!!(condition)) \
ksft_test_result_pass(fmt, ##__VA_ARGS__);\
else \
ksft_test_result_fail(fmt, ##__VA_ARGS__);\
} while (0)
static inline __printf(1, 2) void ksft_test_result_xfail(const char *msg, ...)
{
int saved_errno = errno;
va_list args;
ksft_cnt.ksft_xfail++;
va_start(args, msg);
printf("ok %u # XFAIL ", ksft_test_num());
errno = saved_errno;
vprintf(msg, args);
va_end(args);
}
static inline __printf(1, 2) void ksft_test_result_skip(const char *msg, ...)
{
int saved_errno = errno;
va_list args;
ksft_cnt.ksft_xskip++;
va_start(args, msg);
printf("ok %u # SKIP ", ksft_test_num());
errno = saved_errno;
vprintf(msg, args);
va_end(args);
}
/* TODO: how does "error" differ from "fail" or "skip"? */
static inline __printf(1, 2) void ksft_test_result_error(const char *msg, ...)
{
int saved_errno = errno;
va_list args;
ksft_cnt.ksft_error++;
va_start(args, msg);
printf("not ok %u # error ", ksft_test_num());
errno = saved_errno;
vprintf(msg, args);
va_end(args);
}
static inline __printf(3, 4)
void ksft_test_result_code(int exit_code, const char *test_name,
const char *msg, ...)
{
const char *tap_code = "ok";
const char *directive = "";
int saved_errno = errno;
va_list args;
switch (exit_code) {
case KSFT_PASS:
ksft_cnt.ksft_pass++;
break;
case KSFT_XFAIL:
directive = " # XFAIL ";
ksft_cnt.ksft_xfail++;
break;
case KSFT_XPASS:
directive = " # XPASS ";
ksft_cnt.ksft_xpass++;
break;
case KSFT_SKIP:
directive = " # SKIP ";
ksft_cnt.ksft_xskip++;
break;
case KSFT_FAIL:
default:
tap_code = "not ok";
ksft_cnt.ksft_fail++;
break;
}
/* Docs seem to call for double space if directive is absent */
if (!directive[0] && msg)
directive = " # ";
printf("%s %u %s%s", tap_code, ksft_test_num(), test_name, directive);
errno = saved_errno;
if (msg) {
va_start(args, msg);
vprintf(msg, args);
va_end(args);
}
printf("\n");
}
/**
* ksft_test_result() - Report test success based on truth of condition
*
* @condition: if true, report test success, otherwise failure.
*/
#define ksft_test_result_report(result, fmt, ...) do { \
switch (result) { \
case KSFT_PASS: \
ksft_test_result_pass(fmt, ##__VA_ARGS__); \
break; \
case KSFT_FAIL: \
ksft_test_result_fail(fmt, ##__VA_ARGS__); \
break; \
case KSFT_XFAIL: \
ksft_test_result_xfail(fmt, ##__VA_ARGS__); \
break; \
case KSFT_SKIP: \
ksft_test_result_skip(fmt, ##__VA_ARGS__); \
break; \
} } while (0)
static inline __noreturn void ksft_exit_pass(void)
{
ksft_print_cnts();
exit(KSFT_PASS);
}
static inline __noreturn void ksft_exit_fail(void)
{
ksft_print_cnts();
exit(KSFT_FAIL);
}
/**
* ksft_exit() - Exit selftest based on truth of condition
*
* @condition: if true, exit self test with success, otherwise fail.
*/
#define ksft_exit(condition) do { \
if (!!(condition)) \
ksft_exit_pass(); \
else \
ksft_exit_fail(); \
} while (0)
/**
* ksft_finished() - Exit selftest with success if all tests passed
*/
#define ksft_finished() \
ksft_exit(ksft_plan == \
ksft_cnt.ksft_pass + \
ksft_cnt.ksft_xfail + \
ksft_cnt.ksft_xskip)
static inline __noreturn __printf(1, 2) void ksft_exit_fail_msg(const char *msg, ...)
{
int saved_errno = errno;
va_list args;
va_start(args, msg);
printf("Bail out! ");
errno = saved_errno;
vprintf(msg, args);
va_end(args);
ksft_print_cnts();
exit(KSFT_FAIL);
}
static inline __noreturn void ksft_exit_fail_perror(const char *msg)
{
#ifndef NOLIBC
ksft_exit_fail_msg("%s: %s (%d)\n", msg, strerror(errno), errno);
#else
/*
* nolibc doesn't provide strerror() and it seems
* inappropriate to add one, just print the errno.
*/
ksft_exit_fail_msg("%s: %d)\n", msg, errno);
#endif
}
static inline __noreturn void ksft_exit_xfail(void)
{
ksft_print_cnts();
exit(KSFT_XFAIL);
}
static inline __noreturn void ksft_exit_xpass(void)
{
ksft_print_cnts();
exit(KSFT_XPASS);
}
static inline __noreturn __printf(1, 2) void ksft_exit_skip(const char *msg, ...)
{
int saved_errno = errno;
va_list args;
va_start(args, msg);
/*
* FIXME: several tests misuse ksft_exit_skip so produce
* something sensible if some tests have already been run
* or a plan has been printed. Those tests should use
* ksft_test_result_skip or ksft_exit_fail_msg instead.
*/
if (ksft_plan || ksft_test_num()) {
ksft_cnt.ksft_xskip++;
printf("ok %d # SKIP ", 1 + ksft_test_num());
} else {
printf("1..0 # SKIP ");
}
if (msg) {
errno = saved_errno;
vprintf(msg, args);
va_end(args);
}
if (ksft_test_num())
ksft_print_cnts();
exit(KSFT_SKIP);
}
selftests/timers/posix_timers: Reimplement check_timer_distribution() check_timer_distribution() runs ten threads in a busy loop and tries to test that the kernel distributes a process posix CPU timer signal to every thread over time. There is not guarantee that this is true even after commit bcb7ee79029d ("posix-timers: Prefer delivery of signals to the current thread") because that commit only avoids waking up the sleeping process leader thread, but that has nothing to do with the actual signal delivery. As the signal is process wide the first thread which observes sigpending and wins the race to lock sighand will deliver the signal. Testing shows that this hangs on a regular base because some threads never win the race. The comment "This primarily tests that the kernel does not favour any one." is wrong. The kernel does favour a thread which hits the timer interrupt when CLOCK_PROCESS_CPUTIME_ID expires. Rewrite the test so it only checks that the group leader sleeping in join() never receives SIGALRM and the thread which burns CPU cycles receives all signals. In older kernels which do not have commit bcb7ee79029d ("posix-timers: Prefer delivery of signals to the current thread") the test-case fails immediately, the very 1st tick wakes the leader up. Otherwise it quickly succeeds after 100 ticks. CI testing wants to use newer selftest versions on stable kernels. In this case the test is guaranteed to fail. So check in the failure case whether the kernel version is less than v6.3 and skip the test result in that case. [ tglx: Massaged change log, renamed the version check helper ] Fixes: e797203fb3ba ("selftests/timers/posix_timers: Test delivery of signals across threads") Signed-off-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20240409133802.GD29396@redhat.com
2024-04-09 15:38:03 +02:00
static inline int ksft_min_kernel_version(unsigned int min_major,
unsigned int min_minor)
{
#ifdef NOLIBC
ksft_print_msg("NOLIBC: Can't check kernel version: Function not implemented\n");
return 0;
#else
selftests/timers/posix_timers: Reimplement check_timer_distribution() check_timer_distribution() runs ten threads in a busy loop and tries to test that the kernel distributes a process posix CPU timer signal to every thread over time. There is not guarantee that this is true even after commit bcb7ee79029d ("posix-timers: Prefer delivery of signals to the current thread") because that commit only avoids waking up the sleeping process leader thread, but that has nothing to do with the actual signal delivery. As the signal is process wide the first thread which observes sigpending and wins the race to lock sighand will deliver the signal. Testing shows that this hangs on a regular base because some threads never win the race. The comment "This primarily tests that the kernel does not favour any one." is wrong. The kernel does favour a thread which hits the timer interrupt when CLOCK_PROCESS_CPUTIME_ID expires. Rewrite the test so it only checks that the group leader sleeping in join() never receives SIGALRM and the thread which burns CPU cycles receives all signals. In older kernels which do not have commit bcb7ee79029d ("posix-timers: Prefer delivery of signals to the current thread") the test-case fails immediately, the very 1st tick wakes the leader up. Otherwise it quickly succeeds after 100 ticks. CI testing wants to use newer selftest versions on stable kernels. In this case the test is guaranteed to fail. So check in the failure case whether the kernel version is less than v6.3 and skip the test result in that case. [ tglx: Massaged change log, renamed the version check helper ] Fixes: e797203fb3ba ("selftests/timers/posix_timers: Test delivery of signals across threads") Signed-off-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20240409133802.GD29396@redhat.com
2024-04-09 15:38:03 +02:00
unsigned int major, minor;
struct utsname info;
if (uname(&info) || sscanf(info.release, "%u.%u.", &major, &minor) != 2)
ksft_exit_fail_msg("Can't parse kernel version\n");
return major > min_major || (major == min_major && minor >= min_minor);
#endif
selftests/timers/posix_timers: Reimplement check_timer_distribution() check_timer_distribution() runs ten threads in a busy loop and tries to test that the kernel distributes a process posix CPU timer signal to every thread over time. There is not guarantee that this is true even after commit bcb7ee79029d ("posix-timers: Prefer delivery of signals to the current thread") because that commit only avoids waking up the sleeping process leader thread, but that has nothing to do with the actual signal delivery. As the signal is process wide the first thread which observes sigpending and wins the race to lock sighand will deliver the signal. Testing shows that this hangs on a regular base because some threads never win the race. The comment "This primarily tests that the kernel does not favour any one." is wrong. The kernel does favour a thread which hits the timer interrupt when CLOCK_PROCESS_CPUTIME_ID expires. Rewrite the test so it only checks that the group leader sleeping in join() never receives SIGALRM and the thread which burns CPU cycles receives all signals. In older kernels which do not have commit bcb7ee79029d ("posix-timers: Prefer delivery of signals to the current thread") the test-case fails immediately, the very 1st tick wakes the leader up. Otherwise it quickly succeeds after 100 ticks. CI testing wants to use newer selftest versions on stable kernels. In this case the test is guaranteed to fail. So check in the failure case whether the kernel version is less than v6.3 and skip the test result in that case. [ tglx: Massaged change log, renamed the version check helper ] Fixes: e797203fb3ba ("selftests/timers/posix_timers: Test delivery of signals across threads") Signed-off-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20240409133802.GD29396@redhat.com
2024-04-09 15:38:03 +02:00
}
#endif /* __KSELFTEST_H */