summaryrefslogtreecommitdiffstats
path: root/tools/testing/selftests/pidfd
diff options
context:
space:
mode:
authorDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-07 18:49:45 +0000
committerDaniel Baumann <daniel.baumann@progress-linux.org>2024-04-07 18:49:45 +0000
commit2c3c1048746a4622d8c89a29670120dc8fab93c4 (patch)
tree848558de17fb3008cdf4d861b01ac7781903ce39 /tools/testing/selftests/pidfd
parentInitial commit. (diff)
downloadlinux-upstream.tar.xz
linux-upstream.zip
Adding upstream version 6.1.76.upstream/6.1.76upstream
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'tools/testing/selftests/pidfd')
-rw-r--r--tools/testing/selftests/pidfd/.gitignore8
-rw-r--r--tools/testing/selftests/pidfd/Makefile8
-rw-r--r--tools/testing/selftests/pidfd/config7
-rw-r--r--tools/testing/selftests/pidfd/pidfd.h118
-rw-r--r--tools/testing/selftests/pidfd/pidfd_fdinfo_test.c310
-rw-r--r--tools/testing/selftests/pidfd/pidfd_getfd_test.c246
-rw-r--r--tools/testing/selftests/pidfd/pidfd_open_test.c163
-rw-r--r--tools/testing/selftests/pidfd/pidfd_poll_test.c116
-rw-r--r--tools/testing/selftests/pidfd/pidfd_setns_test.c559
-rw-r--r--tools/testing/selftests/pidfd/pidfd_test.c575
-rw-r--r--tools/testing/selftests/pidfd/pidfd_wait.c233
11 files changed, 2343 insertions, 0 deletions
diff --git a/tools/testing/selftests/pidfd/.gitignore b/tools/testing/selftests/pidfd/.gitignore
new file mode 100644
index 000000000..973198a3e
--- /dev/null
+++ b/tools/testing/selftests/pidfd/.gitignore
@@ -0,0 +1,8 @@
+# SPDX-License-Identifier: GPL-2.0-only
+pidfd_open_test
+pidfd_poll_test
+pidfd_test
+pidfd_wait
+pidfd_fdinfo_test
+pidfd_getfd_test
+pidfd_setns_test
diff --git a/tools/testing/selftests/pidfd/Makefile b/tools/testing/selftests/pidfd/Makefile
new file mode 100644
index 000000000..d731e3e76
--- /dev/null
+++ b/tools/testing/selftests/pidfd/Makefile
@@ -0,0 +1,8 @@
+# SPDX-License-Identifier: GPL-2.0-only
+CFLAGS += -g $(KHDR_INCLUDES) -pthread -Wall
+
+TEST_GEN_PROGS := pidfd_test pidfd_fdinfo_test pidfd_open_test \
+ pidfd_poll_test pidfd_wait pidfd_getfd_test pidfd_setns_test
+
+include ../lib.mk
+
diff --git a/tools/testing/selftests/pidfd/config b/tools/testing/selftests/pidfd/config
new file mode 100644
index 000000000..f6f2965e1
--- /dev/null
+++ b/tools/testing/selftests/pidfd/config
@@ -0,0 +1,7 @@
+CONFIG_UTS_NS=y
+CONFIG_IPC_NS=y
+CONFIG_USER_NS=y
+CONFIG_PID_NS=y
+CONFIG_NET_NS=y
+CONFIG_CGROUPS=y
+CONFIG_CHECKPOINT_RESTORE=y
diff --git a/tools/testing/selftests/pidfd/pidfd.h b/tools/testing/selftests/pidfd/pidfd.h
new file mode 100644
index 000000000..6922d6417
--- /dev/null
+++ b/tools/testing/selftests/pidfd/pidfd.h
@@ -0,0 +1,118 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __PIDFD_H
+#define __PIDFD_H
+
+#define _GNU_SOURCE
+#include <errno.h>
+#include <fcntl.h>
+#include <sched.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syscall.h>
+#include <sys/mount.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+
+#include "../kselftest.h"
+
+#ifndef P_PIDFD
+#define P_PIDFD 3
+#endif
+
+#ifndef CLONE_NEWTIME
+#define CLONE_NEWTIME 0x00000080
+#endif
+
+#ifndef CLONE_PIDFD
+#define CLONE_PIDFD 0x00001000
+#endif
+
+#ifndef __NR_pidfd_open
+#define __NR_pidfd_open -1
+#endif
+
+#ifndef __NR_pidfd_send_signal
+#define __NR_pidfd_send_signal -1
+#endif
+
+#ifndef __NR_clone3
+#define __NR_clone3 -1
+#endif
+
+#ifndef __NR_pidfd_getfd
+#define __NR_pidfd_getfd -1
+#endif
+
+#ifndef PIDFD_NONBLOCK
+#define PIDFD_NONBLOCK O_NONBLOCK
+#endif
+
+/*
+ * The kernel reserves 300 pids via RESERVED_PIDS in kernel/pid.c
+ * That means, when it wraps around any pid < 300 will be skipped.
+ * So we need to use a pid > 300 in order to test recycling.
+ */
+#define PID_RECYCLE 1000
+
+/*
+ * Define a few custom error codes for the child process to clearly indicate
+ * what is happening. This way we can tell the difference between a system
+ * error, a test error, etc.
+ */
+#define PIDFD_PASS 0
+#define PIDFD_FAIL 1
+#define PIDFD_ERROR 2
+#define PIDFD_SKIP 3
+#define PIDFD_XFAIL 4
+
+static inline int wait_for_pid(pid_t pid)
+{
+ int status, ret;
+
+again:
+ ret = waitpid(pid, &status, 0);
+ if (ret == -1) {
+ if (errno == EINTR)
+ goto again;
+
+ ksft_print_msg("waitpid returned -1, errno=%d\n", errno);
+ return -1;
+ }
+
+ if (!WIFEXITED(status)) {
+ ksft_print_msg(
+ "waitpid !WIFEXITED, WIFSIGNALED=%d, WTERMSIG=%d\n",
+ WIFSIGNALED(status), WTERMSIG(status));
+ return -1;
+ }
+
+ ret = WEXITSTATUS(status);
+ ksft_print_msg("waitpid WEXITSTATUS=%d\n", ret);
+ return ret;
+}
+
+static inline int sys_pidfd_open(pid_t pid, unsigned int flags)
+{
+ return syscall(__NR_pidfd_open, pid, flags);
+}
+
+static inline int sys_pidfd_send_signal(int pidfd, int sig, siginfo_t *info,
+ unsigned int flags)
+{
+ return syscall(__NR_pidfd_send_signal, pidfd, sig, info, flags);
+}
+
+static inline int sys_pidfd_getfd(int pidfd, int fd, int flags)
+{
+ return syscall(__NR_pidfd_getfd, pidfd, fd, flags);
+}
+
+static inline int sys_memfd_create(const char *name, unsigned int flags)
+{
+ return syscall(__NR_memfd_create, name, flags);
+}
+
+#endif /* __PIDFD_H */
diff --git a/tools/testing/selftests/pidfd/pidfd_fdinfo_test.c b/tools/testing/selftests/pidfd/pidfd_fdinfo_test.c
new file mode 100644
index 000000000..3bc46d615
--- /dev/null
+++ b/tools/testing/selftests/pidfd/pidfd_fdinfo_test.c
@@ -0,0 +1,310 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#define _GNU_SOURCE
+#include <assert.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <linux/types.h>
+#include <sched.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syscall.h>
+#include <sys/wait.h>
+#include <sys/mman.h>
+
+#include "pidfd.h"
+#include "../kselftest.h"
+
+struct error {
+ int code;
+ char msg[512];
+};
+
+static int error_set(struct error *err, int code, const char *fmt, ...)
+{
+ va_list args;
+ int r;
+
+ if (code == PIDFD_PASS || !err || err->code != PIDFD_PASS)
+ return code;
+
+ err->code = code;
+ va_start(args, fmt);
+ r = vsnprintf(err->msg, sizeof(err->msg), fmt, args);
+ assert((size_t)r < sizeof(err->msg));
+ va_end(args);
+
+ return code;
+}
+
+static void error_report(struct error *err, const char *test_name)
+{
+ switch (err->code) {
+ case PIDFD_ERROR:
+ ksft_exit_fail_msg("%s test: Fatal: %s\n", test_name, err->msg);
+ break;
+
+ case PIDFD_FAIL:
+ /* will be: not ok %d # error %s test: %s */
+ ksft_test_result_error("%s test: %s\n", test_name, err->msg);
+ break;
+
+ case PIDFD_SKIP:
+ /* will be: not ok %d # SKIP %s test: %s */
+ ksft_test_result_skip("%s test: %s\n", test_name, err->msg);
+ break;
+
+ case PIDFD_XFAIL:
+ ksft_test_result_pass("%s test: Expected failure: %s\n",
+ test_name, err->msg);
+ break;
+
+ case PIDFD_PASS:
+ ksft_test_result_pass("%s test: Passed\n", test_name);
+ break;
+
+ default:
+ ksft_exit_fail_msg("%s test: Unknown code: %d %s\n",
+ test_name, err->code, err->msg);
+ break;
+ }
+}
+
+static inline int error_check(struct error *err, const char *test_name)
+{
+ /* In case of error we bail out and terminate the test program */
+ if (err->code == PIDFD_ERROR)
+ error_report(err, test_name);
+
+ return err->code;
+}
+
+#define CHILD_STACK_SIZE 8192
+
+struct child {
+ char *stack;
+ pid_t pid;
+ int fd;
+};
+
+static struct child clone_newns(int (*fn)(void *), void *args,
+ struct error *err)
+{
+ static int flags = CLONE_PIDFD | CLONE_NEWPID | CLONE_NEWNS | SIGCHLD;
+ struct child ret;
+
+ if (!(flags & CLONE_NEWUSER) && geteuid() != 0)
+ flags |= CLONE_NEWUSER;
+
+ ret.stack = mmap(NULL, CHILD_STACK_SIZE, PROT_READ | PROT_WRITE,
+ MAP_PRIVATE | MAP_ANONYMOUS | MAP_STACK, -1, 0);
+ if (ret.stack == MAP_FAILED) {
+ error_set(err, -1, "mmap of stack failed (errno %d)", errno);
+ return ret;
+ }
+
+#ifdef __ia64__
+ ret.pid = __clone2(fn, ret.stack, CHILD_STACK_SIZE, flags, args, &ret.fd);
+#else
+ ret.pid = clone(fn, ret.stack + CHILD_STACK_SIZE, flags, args, &ret.fd);
+#endif
+
+ if (ret.pid < 0) {
+ error_set(err, PIDFD_ERROR, "clone failed (ret %d, errno %d)",
+ ret.fd, errno);
+ return ret;
+ }
+
+ ksft_print_msg("New child: %d, fd: %d\n", ret.pid, ret.fd);
+
+ return ret;
+}
+
+static inline void child_close(struct child *child)
+{
+ close(child->fd);
+}
+
+static inline int child_join(struct child *child, struct error *err)
+{
+ int r;
+
+ r = wait_for_pid(child->pid);
+ if (r < 0)
+ error_set(err, PIDFD_ERROR, "waitpid failed (ret %d, errno %d)",
+ r, errno);
+ else if (r > 0)
+ error_set(err, r, "child %d reported: %d", child->pid, r);
+
+ if (munmap(child->stack, CHILD_STACK_SIZE)) {
+ error_set(err, -1, "munmap of child stack failed (errno %d)", errno);
+ r = -1;
+ }
+
+ return r;
+}
+
+static inline int child_join_close(struct child *child, struct error *err)
+{
+ child_close(child);
+ return child_join(child, err);
+}
+
+static inline void trim_newline(char *str)
+{
+ char *pos = strrchr(str, '\n');
+
+ if (pos)
+ *pos = '\0';
+}
+
+static int verify_fdinfo(int pidfd, struct error *err, const char *prefix,
+ size_t prefix_len, const char *expect, ...)
+{
+ char buffer[512] = {0, };
+ char path[512] = {0, };
+ va_list args;
+ FILE *f;
+ char *line = NULL;
+ size_t n = 0;
+ int found = 0;
+ int r;
+
+ va_start(args, expect);
+ r = vsnprintf(buffer, sizeof(buffer), expect, args);
+ assert((size_t)r < sizeof(buffer));
+ va_end(args);
+
+ snprintf(path, sizeof(path), "/proc/self/fdinfo/%d", pidfd);
+ f = fopen(path, "re");
+ if (!f)
+ return error_set(err, PIDFD_ERROR, "fdinfo open failed for %d",
+ pidfd);
+
+ while (getline(&line, &n, f) != -1) {
+ char *val;
+
+ if (strncmp(line, prefix, prefix_len))
+ continue;
+
+ found = 1;
+
+ val = line + prefix_len;
+ r = strcmp(val, buffer);
+ if (r != 0) {
+ trim_newline(line);
+ trim_newline(buffer);
+ error_set(err, PIDFD_FAIL, "%s '%s' != '%s'",
+ prefix, val, buffer);
+ }
+ break;
+ }
+
+ free(line);
+ fclose(f);
+
+ if (found == 0)
+ return error_set(err, PIDFD_FAIL, "%s not found for fd %d",
+ prefix, pidfd);
+
+ return PIDFD_PASS;
+}
+
+static int child_fdinfo_nspid_test(void *args)
+{
+ struct error err;
+ int pidfd;
+ int r;
+
+ /* if we got no fd for the sibling, we are done */
+ if (!args)
+ return PIDFD_PASS;
+
+ /* verify that we can not resolve the pidfd for a process
+ * in a sibling pid namespace, i.e. a pid namespace it is
+ * not in our or a descended namespace
+ */
+ r = mount(NULL, "/", NULL, MS_REC | MS_PRIVATE, 0);
+ if (r < 0) {
+ ksft_print_msg("Failed to remount / private\n");
+ return PIDFD_ERROR;
+ }
+
+ (void)umount2("/proc", MNT_DETACH);
+ r = mount("proc", "/proc", "proc", 0, NULL);
+ if (r < 0) {
+ ksft_print_msg("Failed to remount /proc\n");
+ return PIDFD_ERROR;
+ }
+
+ pidfd = *(int *)args;
+ r = verify_fdinfo(pidfd, &err, "NSpid:", 6, "\t0\n");
+
+ if (r != PIDFD_PASS)
+ ksft_print_msg("NSpid fdinfo check failed: %s\n", err.msg);
+
+ return r;
+}
+
+static void test_pidfd_fdinfo_nspid(void)
+{
+ struct child a, b;
+ struct error err = {0, };
+ const char *test_name = "pidfd check for NSpid in fdinfo";
+
+ /* Create a new child in a new pid and mount namespace */
+ a = clone_newns(child_fdinfo_nspid_test, NULL, &err);
+ error_check(&err, test_name);
+
+ /* Pass the pidfd representing the first child to the
+ * second child, which will be in a sibling pid namespace,
+ * which means that the fdinfo NSpid entry for the pidfd
+ * should only contain '0'.
+ */
+ b = clone_newns(child_fdinfo_nspid_test, &a.fd, &err);
+ error_check(&err, test_name);
+
+ /* The children will have pid 1 in the new pid namespace,
+ * so the line must be 'NSPid:\t<pid>\t1'.
+ */
+ verify_fdinfo(a.fd, &err, "NSpid:", 6, "\t%d\t%d\n", a.pid, 1);
+ verify_fdinfo(b.fd, &err, "NSpid:", 6, "\t%d\t%d\n", b.pid, 1);
+
+ /* wait for the process, check the exit status and set
+ * 'err' accordingly, if it is not already set.
+ */
+ child_join_close(&a, &err);
+ child_join_close(&b, &err);
+
+ error_report(&err, test_name);
+}
+
+static void test_pidfd_dead_fdinfo(void)
+{
+ struct child a;
+ struct error err = {0, };
+ const char *test_name = "pidfd check fdinfo for dead process";
+
+ /* Create a new child in a new pid and mount namespace */
+ a = clone_newns(child_fdinfo_nspid_test, NULL, &err);
+ error_check(&err, test_name);
+ child_join(&a, &err);
+
+ verify_fdinfo(a.fd, &err, "Pid:", 4, "\t-1\n");
+ verify_fdinfo(a.fd, &err, "NSpid:", 6, "\t-1\n");
+ child_close(&a);
+ error_report(&err, test_name);
+}
+
+int main(int argc, char **argv)
+{
+ ksft_print_header();
+ ksft_set_plan(2);
+
+ test_pidfd_fdinfo_nspid();
+ test_pidfd_dead_fdinfo();
+
+ return ksft_exit_pass();
+}
diff --git a/tools/testing/selftests/pidfd/pidfd_getfd_test.c b/tools/testing/selftests/pidfd/pidfd_getfd_test.c
new file mode 100644
index 000000000..0930e2411
--- /dev/null
+++ b/tools/testing/selftests/pidfd/pidfd_getfd_test.c
@@ -0,0 +1,246 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#define _GNU_SOURCE
+#include <errno.h>
+#include <fcntl.h>
+#include <limits.h>
+#include <linux/types.h>
+#include <sched.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syscall.h>
+#include <sys/prctl.h>
+#include <sys/wait.h>
+#include <unistd.h>
+#include <sys/socket.h>
+#include <linux/kcmp.h>
+
+#include "pidfd.h"
+#include "../kselftest_harness.h"
+
+/*
+ * UNKNOWN_FD is an fd number that should never exist in the child, as it is
+ * used to check the negative case.
+ */
+#define UNKNOWN_FD 111
+#define UID_NOBODY 65535
+
+static int sys_kcmp(pid_t pid1, pid_t pid2, int type, unsigned long idx1,
+ unsigned long idx2)
+{
+ return syscall(__NR_kcmp, pid1, pid2, type, idx1, idx2);
+}
+
+static int __child(int sk, int memfd)
+{
+ int ret;
+ char buf;
+
+ /*
+ * Ensure we don't leave around a bunch of orphaned children if our
+ * tests fail.
+ */
+ ret = prctl(PR_SET_PDEATHSIG, SIGKILL);
+ if (ret) {
+ fprintf(stderr, "%s: Child could not set DEATHSIG\n",
+ strerror(errno));
+ return -1;
+ }
+
+ ret = send(sk, &memfd, sizeof(memfd), 0);
+ if (ret != sizeof(memfd)) {
+ fprintf(stderr, "%s: Child failed to send fd number\n",
+ strerror(errno));
+ return -1;
+ }
+
+ /*
+ * The fixture setup is completed at this point. The tests will run.
+ *
+ * This blocking recv enables the parent to message the child.
+ * Either we will read 'P' off of the sk, indicating that we need
+ * to disable ptrace, or we will read a 0, indicating that the other
+ * side has closed the sk. This occurs during fixture teardown time,
+ * indicating that the child should exit.
+ */
+ while ((ret = recv(sk, &buf, sizeof(buf), 0)) > 0) {
+ if (buf == 'P') {
+ ret = prctl(PR_SET_DUMPABLE, 0);
+ if (ret < 0) {
+ fprintf(stderr,
+ "%s: Child failed to disable ptrace\n",
+ strerror(errno));
+ return -1;
+ }
+ } else {
+ fprintf(stderr, "Child received unknown command %c\n",
+ buf);
+ return -1;
+ }
+ ret = send(sk, &buf, sizeof(buf), 0);
+ if (ret != 1) {
+ fprintf(stderr, "%s: Child failed to ack\n",
+ strerror(errno));
+ return -1;
+ }
+ }
+ if (ret < 0) {
+ fprintf(stderr, "%s: Child failed to read from socket\n",
+ strerror(errno));
+ return -1;
+ }
+
+ return 0;
+}
+
+static int child(int sk)
+{
+ int memfd, ret;
+
+ memfd = sys_memfd_create("test", 0);
+ if (memfd < 0) {
+ fprintf(stderr, "%s: Child could not create memfd\n",
+ strerror(errno));
+ ret = -1;
+ } else {
+ ret = __child(sk, memfd);
+ close(memfd);
+ }
+
+ close(sk);
+ return ret;
+}
+
+FIXTURE(child)
+{
+ /*
+ * remote_fd is the number of the FD which we are trying to retrieve
+ * from the child.
+ */
+ int remote_fd;
+ /* pid points to the child which we are fetching FDs from */
+ pid_t pid;
+ /* pidfd is the pidfd of the child */
+ int pidfd;
+ /*
+ * sk is our side of the socketpair used to communicate with the child.
+ * When it is closed, the child will exit.
+ */
+ int sk;
+};
+
+FIXTURE_SETUP(child)
+{
+ int ret, sk_pair[2];
+
+ ASSERT_EQ(0, socketpair(PF_LOCAL, SOCK_SEQPACKET, 0, sk_pair)) {
+ TH_LOG("%s: failed to create socketpair", strerror(errno));
+ }
+ self->sk = sk_pair[0];
+
+ self->pid = fork();
+ ASSERT_GE(self->pid, 0);
+
+ if (self->pid == 0) {
+ close(sk_pair[0]);
+ if (child(sk_pair[1]))
+ _exit(EXIT_FAILURE);
+ _exit(EXIT_SUCCESS);
+ }
+
+ close(sk_pair[1]);
+
+ self->pidfd = sys_pidfd_open(self->pid, 0);
+ ASSERT_GE(self->pidfd, 0);
+
+ /*
+ * Wait for the child to complete setup. It'll send the remote memfd's
+ * number when ready.
+ */
+ ret = recv(sk_pair[0], &self->remote_fd, sizeof(self->remote_fd), 0);
+ ASSERT_EQ(sizeof(self->remote_fd), ret);
+}
+
+FIXTURE_TEARDOWN(child)
+{
+ EXPECT_EQ(0, close(self->pidfd));
+ EXPECT_EQ(0, close(self->sk));
+
+ EXPECT_EQ(0, wait_for_pid(self->pid));
+}
+
+TEST_F(child, disable_ptrace)
+{
+ int uid, fd;
+ char c;
+
+ /*
+ * Turn into nobody if we're root, to avoid CAP_SYS_PTRACE
+ *
+ * The tests should run in their own process, so even this test fails,
+ * it shouldn't result in subsequent tests failing.
+ */
+ uid = getuid();
+ if (uid == 0)
+ ASSERT_EQ(0, seteuid(UID_NOBODY));
+
+ ASSERT_EQ(1, send(self->sk, "P", 1, 0));
+ ASSERT_EQ(1, recv(self->sk, &c, 1, 0));
+
+ fd = sys_pidfd_getfd(self->pidfd, self->remote_fd, 0);
+ EXPECT_EQ(-1, fd);
+ EXPECT_EQ(EPERM, errno);
+
+ if (uid == 0)
+ ASSERT_EQ(0, seteuid(0));
+}
+
+TEST_F(child, fetch_fd)
+{
+ int fd, ret;
+
+ fd = sys_pidfd_getfd(self->pidfd, self->remote_fd, 0);
+ ASSERT_GE(fd, 0);
+
+ ret = sys_kcmp(getpid(), self->pid, KCMP_FILE, fd, self->remote_fd);
+ if (ret < 0 && errno == ENOSYS)
+ SKIP(return, "kcmp() syscall not supported");
+ EXPECT_EQ(ret, 0);
+
+ ret = fcntl(fd, F_GETFD);
+ ASSERT_GE(ret, 0);
+ EXPECT_GE(ret & FD_CLOEXEC, 0);
+
+ close(fd);
+}
+
+TEST_F(child, test_unknown_fd)
+{
+ int fd;
+
+ fd = sys_pidfd_getfd(self->pidfd, UNKNOWN_FD, 0);
+ EXPECT_EQ(-1, fd) {
+ TH_LOG("getfd succeeded while fetching unknown fd");
+ };
+ EXPECT_EQ(EBADF, errno) {
+ TH_LOG("%s: getfd did not get EBADF", strerror(errno));
+ }
+}
+
+TEST(flags_set)
+{
+ ASSERT_EQ(-1, sys_pidfd_getfd(0, 0, 1));
+ EXPECT_EQ(errno, EINVAL);
+}
+
+#if __NR_pidfd_getfd == -1
+int main(void)
+{
+ fprintf(stderr, "__NR_pidfd_getfd undefined. The pidfd_getfd syscall is unavailable. Test aborting\n");
+ return KSFT_SKIP;
+}
+#else
+TEST_HARNESS_MAIN
+#endif
diff --git a/tools/testing/selftests/pidfd/pidfd_open_test.c b/tools/testing/selftests/pidfd/pidfd_open_test.c
new file mode 100644
index 000000000..8a59438cc
--- /dev/null
+++ b/tools/testing/selftests/pidfd/pidfd_open_test.c
@@ -0,0 +1,163 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#define _GNU_SOURCE
+#include <errno.h>
+#include <fcntl.h>
+#include <inttypes.h>
+#include <limits.h>
+#include <linux/types.h>
+#include <sched.h>
+#include <signal.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syscall.h>
+#include <sys/mount.h>
+#include <sys/prctl.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+#include "pidfd.h"
+#include "../kselftest.h"
+
+static int safe_int(const char *numstr, int *converted)
+{
+ char *err = NULL;
+ long sli;
+
+ errno = 0;
+ sli = strtol(numstr, &err, 0);
+ if (errno == ERANGE && (sli == LONG_MAX || sli == LONG_MIN))
+ return -ERANGE;
+
+ if (errno != 0 && sli == 0)
+ return -EINVAL;
+
+ if (err == numstr || *err != '\0')
+ return -EINVAL;
+
+ if (sli > INT_MAX || sli < INT_MIN)
+ return -ERANGE;
+
+ *converted = (int)sli;
+ return 0;
+}
+
+static int char_left_gc(const char *buffer, size_t len)
+{
+ size_t i;
+
+ for (i = 0; i < len; i++) {
+ if (buffer[i] == ' ' ||
+ buffer[i] == '\t')
+ continue;
+
+ return i;
+ }
+
+ return 0;
+}
+
+static int char_right_gc(const char *buffer, size_t len)
+{
+ int i;
+
+ for (i = len - 1; i >= 0; i--) {
+ if (buffer[i] == ' ' ||
+ buffer[i] == '\t' ||
+ buffer[i] == '\n' ||
+ buffer[i] == '\0')
+ continue;
+
+ return i + 1;
+ }
+
+ return 0;
+}
+
+static char *trim_whitespace_in_place(char *buffer)
+{
+ buffer += char_left_gc(buffer, strlen(buffer));
+ buffer[char_right_gc(buffer, strlen(buffer))] = '\0';
+ return buffer;
+}
+
+static pid_t get_pid_from_fdinfo_file(int pidfd, const char *key, size_t keylen)
+{
+ int ret;
+ char path[512];
+ FILE *f;
+ size_t n = 0;
+ pid_t result = -1;
+ char *line = NULL;
+
+ snprintf(path, sizeof(path), "/proc/self/fdinfo/%d", pidfd);
+
+ f = fopen(path, "re");
+ if (!f)
+ return -1;
+
+ while (getline(&line, &n, f) != -1) {
+ char *numstr;
+
+ if (strncmp(line, key, keylen))
+ continue;
+
+ numstr = trim_whitespace_in_place(line + 4);
+ ret = safe_int(numstr, &result);
+ if (ret < 0)
+ goto out;
+
+ break;
+ }
+
+out:
+ free(line);
+ fclose(f);
+ return result;
+}
+
+int main(int argc, char **argv)
+{
+ int pidfd = -1, ret = 1;
+ pid_t pid;
+
+ ksft_set_plan(3);
+
+ pidfd = sys_pidfd_open(-1, 0);
+ if (pidfd >= 0) {
+ ksft_print_msg(
+ "%s - succeeded to open pidfd for invalid pid -1\n",
+ strerror(errno));
+ goto on_error;
+ }
+ ksft_test_result_pass("do not allow invalid pid test: passed\n");
+
+ pidfd = sys_pidfd_open(getpid(), 1);
+ if (pidfd >= 0) {
+ ksft_print_msg(
+ "%s - succeeded to open pidfd with invalid flag value specified\n",
+ strerror(errno));
+ goto on_error;
+ }
+ ksft_test_result_pass("do not allow invalid flag test: passed\n");
+
+ pidfd = sys_pidfd_open(getpid(), 0);
+ if (pidfd < 0) {
+ ksft_print_msg("%s - failed to open pidfd\n", strerror(errno));
+ goto on_error;
+ }
+ ksft_test_result_pass("open a new pidfd test: passed\n");
+
+ pid = get_pid_from_fdinfo_file(pidfd, "Pid:", sizeof("Pid:") - 1);
+ ksft_print_msg("pidfd %d refers to process with pid %d\n", pidfd, pid);
+
+ ret = 0;
+
+on_error:
+ if (pidfd >= 0)
+ close(pidfd);
+
+ return !ret ? ksft_exit_pass() : ksft_exit_fail();
+}
diff --git a/tools/testing/selftests/pidfd/pidfd_poll_test.c b/tools/testing/selftests/pidfd/pidfd_poll_test.c
new file mode 100644
index 000000000..610811275
--- /dev/null
+++ b/tools/testing/selftests/pidfd/pidfd_poll_test.c
@@ -0,0 +1,116 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#define _GNU_SOURCE
+#include <errno.h>
+#include <linux/types.h>
+#include <poll.h>
+#include <signal.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syscall.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+#include "pidfd.h"
+#include "../kselftest.h"
+
+static bool timeout;
+
+static void handle_alarm(int sig)
+{
+ timeout = true;
+}
+
+int main(int argc, char **argv)
+{
+ struct pollfd fds;
+ int iter, nevents;
+ int nr_iterations = 10000;
+
+ fds.events = POLLIN;
+
+ if (argc > 2)
+ ksft_exit_fail_msg("Unexpected command line argument\n");
+
+ if (argc == 2) {
+ nr_iterations = atoi(argv[1]);
+ if (nr_iterations <= 0)
+ ksft_exit_fail_msg("invalid input parameter %s\n",
+ argv[1]);
+ }
+
+ ksft_print_msg("running pidfd poll test for %d iterations\n",
+ nr_iterations);
+
+ for (iter = 0; iter < nr_iterations; iter++) {
+ int pidfd;
+ int child_pid = fork();
+
+ if (child_pid < 0) {
+ if (errno == EAGAIN) {
+ iter--;
+ continue;
+ }
+ ksft_exit_fail_msg(
+ "%s - failed to fork a child process\n",
+ strerror(errno));
+ }
+
+ if (child_pid == 0) {
+ /* Child process just sleeps for a min and exits */
+ sleep(60);
+ exit(EXIT_SUCCESS);
+ }
+
+ /* Parent kills the child and waits for its death */
+ pidfd = sys_pidfd_open(child_pid, 0);
+ if (pidfd < 0)
+ ksft_exit_fail_msg("%s - pidfd_open failed\n",
+ strerror(errno));
+
+ /* Setup 3 sec alarm - plenty of time */
+ if (signal(SIGALRM, handle_alarm) == SIG_ERR)
+ ksft_exit_fail_msg("%s - signal failed\n",
+ strerror(errno));
+ alarm(3);
+
+ /* Send SIGKILL to the child */
+ if (sys_pidfd_send_signal(pidfd, SIGKILL, NULL, 0))
+ ksft_exit_fail_msg("%s - pidfd_send_signal failed\n",
+ strerror(errno));
+
+ /* Wait for the death notification */
+ fds.fd = pidfd;
+ nevents = poll(&fds, 1, -1);
+
+ /* Check for error conditions */
+ if (nevents < 0)
+ ksft_exit_fail_msg("%s - poll failed\n",
+ strerror(errno));
+
+ if (nevents != 1)
+ ksft_exit_fail_msg("unexpected poll result: %d\n",
+ nevents);
+
+ if (!(fds.revents & POLLIN))
+ ksft_exit_fail_msg(
+ "unexpected event type received: 0x%x\n",
+ fds.revents);
+
+ if (timeout)
+ ksft_exit_fail_msg(
+ "death notification wait timeout\n");
+
+ close(pidfd);
+ /* Wait for child to prevent zombies */
+ if (waitpid(child_pid, NULL, 0) < 0)
+ ksft_exit_fail_msg("%s - waitpid failed\n",
+ strerror(errno));
+
+ }
+
+ ksft_test_result_pass("pidfd poll test: pass\n");
+ return ksft_exit_pass();
+}
diff --git a/tools/testing/selftests/pidfd/pidfd_setns_test.c b/tools/testing/selftests/pidfd/pidfd_setns_test.c
new file mode 100644
index 000000000..6e2f2cd40
--- /dev/null
+++ b/tools/testing/selftests/pidfd/pidfd_setns_test.c
@@ -0,0 +1,559 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#define _GNU_SOURCE
+#include <errno.h>
+#include <fcntl.h>
+#include <limits.h>
+#include <linux/types.h>
+#include <sched.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syscall.h>
+#include <sys/prctl.h>
+#include <sys/wait.h>
+#include <unistd.h>
+#include <sys/socket.h>
+#include <sys/stat.h>
+
+#include "pidfd.h"
+#include "../clone3/clone3_selftests.h"
+#include "../kselftest_harness.h"
+
+enum {
+ PIDFD_NS_USER,
+ PIDFD_NS_MNT,
+ PIDFD_NS_PID,
+ PIDFD_NS_UTS,
+ PIDFD_NS_IPC,
+ PIDFD_NS_NET,
+ PIDFD_NS_CGROUP,
+ PIDFD_NS_PIDCLD,
+ PIDFD_NS_TIME,
+ PIDFD_NS_MAX
+};
+
+const struct ns_info {
+ const char *name;
+ int flag;
+} ns_info[] = {
+ [PIDFD_NS_USER] = { "user", CLONE_NEWUSER, },
+ [PIDFD_NS_MNT] = { "mnt", CLONE_NEWNS, },
+ [PIDFD_NS_PID] = { "pid", CLONE_NEWPID, },
+ [PIDFD_NS_UTS] = { "uts", CLONE_NEWUTS, },
+ [PIDFD_NS_IPC] = { "ipc", CLONE_NEWIPC, },
+ [PIDFD_NS_NET] = { "net", CLONE_NEWNET, },
+ [PIDFD_NS_CGROUP] = { "cgroup", CLONE_NEWCGROUP, },
+ [PIDFD_NS_PIDCLD] = { "pid_for_children", 0, },
+ [PIDFD_NS_TIME] = { "time", CLONE_NEWTIME, },
+};
+
+FIXTURE(current_nsset)
+{
+ pid_t pid;
+ int pidfd;
+ int nsfds[PIDFD_NS_MAX];
+
+ pid_t child_pid_exited;
+ int child_pidfd_exited;
+
+ pid_t child_pid1;
+ int child_pidfd1;
+ int child_nsfds1[PIDFD_NS_MAX];
+
+ pid_t child_pid2;
+ int child_pidfd2;
+ int child_nsfds2[PIDFD_NS_MAX];
+};
+
+static int sys_waitid(int which, pid_t pid, int options)
+{
+ return syscall(__NR_waitid, which, pid, NULL, options, NULL);
+}
+
+pid_t create_child(int *pidfd, unsigned flags)
+{
+ struct __clone_args args = {
+ .flags = CLONE_PIDFD | flags,
+ .exit_signal = SIGCHLD,
+ .pidfd = ptr_to_u64(pidfd),
+ };
+
+ return sys_clone3(&args, sizeof(struct clone_args));
+}
+
+static bool switch_timens(void)
+{
+ int fd, ret;
+
+ if (unshare(CLONE_NEWTIME))
+ return false;
+
+ fd = open("/proc/self/ns/time_for_children", O_RDONLY | O_CLOEXEC);
+ if (fd < 0)
+ return false;
+
+ ret = setns(fd, CLONE_NEWTIME);
+ close(fd);
+ return ret == 0;
+}
+
+static ssize_t read_nointr(int fd, void *buf, size_t count)
+{
+ ssize_t ret;
+
+ do {
+ ret = read(fd, buf, count);
+ } while (ret < 0 && errno == EINTR);
+
+ return ret;
+}
+
+static ssize_t write_nointr(int fd, const void *buf, size_t count)
+{
+ ssize_t ret;
+
+ do {
+ ret = write(fd, buf, count);
+ } while (ret < 0 && errno == EINTR);
+
+ return ret;
+}
+
+FIXTURE_SETUP(current_nsset)
+{
+ int i, proc_fd, ret;
+ int ipc_sockets[2];
+ char c;
+
+ for (i = 0; i < PIDFD_NS_MAX; i++) {
+ self->nsfds[i] = -EBADF;
+ self->child_nsfds1[i] = -EBADF;
+ self->child_nsfds2[i] = -EBADF;
+ }
+
+ proc_fd = open("/proc/self/ns", O_DIRECTORY | O_CLOEXEC);
+ ASSERT_GE(proc_fd, 0) {
+ TH_LOG("%m - Failed to open /proc/self/ns");
+ }
+
+ self->pid = getpid();
+ for (i = 0; i < PIDFD_NS_MAX; i++) {
+ const struct ns_info *info = &ns_info[i];
+ self->nsfds[i] = openat(proc_fd, info->name, O_RDONLY | O_CLOEXEC);
+ if (self->nsfds[i] < 0) {
+ EXPECT_EQ(errno, ENOENT) {
+ TH_LOG("%m - Failed to open %s namespace for process %d",
+ info->name, self->pid);
+ }
+ }
+ }
+
+ self->pidfd = sys_pidfd_open(self->pid, 0);
+ EXPECT_GT(self->pidfd, 0) {
+ TH_LOG("%m - Failed to open pidfd for process %d", self->pid);
+ }
+
+ /* Create task that exits right away. */
+ self->child_pid_exited = create_child(&self->child_pidfd_exited,
+ CLONE_NEWUSER | CLONE_NEWNET);
+ EXPECT_GT(self->child_pid_exited, 0);
+
+ if (self->child_pid_exited == 0)
+ _exit(EXIT_SUCCESS);
+
+ ASSERT_EQ(sys_waitid(P_PID, self->child_pid_exited, WEXITED | WNOWAIT), 0);
+
+ self->pidfd = sys_pidfd_open(self->pid, 0);
+ EXPECT_GE(self->pidfd, 0) {
+ TH_LOG("%m - Failed to open pidfd for process %d", self->pid);
+ }
+
+ ret = socketpair(AF_LOCAL, SOCK_STREAM | SOCK_CLOEXEC, 0, ipc_sockets);
+ EXPECT_EQ(ret, 0);
+
+ /* Create tasks that will be stopped. */
+ self->child_pid1 = create_child(&self->child_pidfd1,
+ CLONE_NEWUSER | CLONE_NEWNS |
+ CLONE_NEWCGROUP | CLONE_NEWIPC |
+ CLONE_NEWUTS | CLONE_NEWPID |
+ CLONE_NEWNET);
+ EXPECT_GE(self->child_pid1, 0);
+
+ if (self->child_pid1 == 0) {
+ close(ipc_sockets[0]);
+
+ if (!switch_timens())
+ _exit(EXIT_FAILURE);
+
+ if (write_nointr(ipc_sockets[1], "1", 1) < 0)
+ _exit(EXIT_FAILURE);
+
+ close(ipc_sockets[1]);
+
+ pause();
+ _exit(EXIT_SUCCESS);
+ }
+
+ close(ipc_sockets[1]);
+ ASSERT_EQ(read_nointr(ipc_sockets[0], &c, 1), 1);
+ close(ipc_sockets[0]);
+
+ ret = socketpair(AF_LOCAL, SOCK_STREAM | SOCK_CLOEXEC, 0, ipc_sockets);
+ EXPECT_EQ(ret, 0);
+
+ self->child_pid2 = create_child(&self->child_pidfd2,
+ CLONE_NEWUSER | CLONE_NEWNS |
+ CLONE_NEWCGROUP | CLONE_NEWIPC |
+ CLONE_NEWUTS | CLONE_NEWPID |
+ CLONE_NEWNET);
+ EXPECT_GE(self->child_pid2, 0);
+
+ if (self->child_pid2 == 0) {
+ close(ipc_sockets[0]);
+
+ if (!switch_timens())
+ _exit(EXIT_FAILURE);
+
+ if (write_nointr(ipc_sockets[1], "1", 1) < 0)
+ _exit(EXIT_FAILURE);
+
+ close(ipc_sockets[1]);
+
+ pause();
+ _exit(EXIT_SUCCESS);
+ }
+
+ close(ipc_sockets[1]);
+ ASSERT_EQ(read_nointr(ipc_sockets[0], &c, 1), 1);
+ close(ipc_sockets[0]);
+
+ for (i = 0; i < PIDFD_NS_MAX; i++) {
+ char p[100];
+
+ const struct ns_info *info = &ns_info[i];
+
+ self->nsfds[i] = openat(proc_fd, info->name, O_RDONLY | O_CLOEXEC);
+ if (self->nsfds[i] < 0) {
+ EXPECT_EQ(errno, ENOENT) {
+ TH_LOG("%m - Failed to open %s namespace for process %d",
+ info->name, self->pid);
+ }
+ }
+
+ ret = snprintf(p, sizeof(p), "/proc/%d/ns/%s",
+ self->child_pid1, info->name);
+ EXPECT_GT(ret, 0);
+ EXPECT_LT(ret, sizeof(p));
+
+ self->child_nsfds1[i] = open(p, O_RDONLY | O_CLOEXEC);
+ if (self->child_nsfds1[i] < 0) {
+ EXPECT_EQ(errno, ENOENT) {
+ TH_LOG("%m - Failed to open %s namespace for process %d",
+ info->name, self->child_pid1);
+ }
+ }
+
+ ret = snprintf(p, sizeof(p), "/proc/%d/ns/%s",
+ self->child_pid2, info->name);
+ EXPECT_GT(ret, 0);
+ EXPECT_LT(ret, sizeof(p));
+
+ self->child_nsfds2[i] = open(p, O_RDONLY | O_CLOEXEC);
+ if (self->child_nsfds2[i] < 0) {
+ EXPECT_EQ(errno, ENOENT) {
+ TH_LOG("%m - Failed to open %s namespace for process %d",
+ info->name, self->child_pid1);
+ }
+ }
+ }
+
+ close(proc_fd);
+}
+
+FIXTURE_TEARDOWN(current_nsset)
+{
+ int i;
+
+ ASSERT_EQ(sys_pidfd_send_signal(self->child_pidfd1,
+ SIGKILL, NULL, 0), 0);
+ ASSERT_EQ(sys_pidfd_send_signal(self->child_pidfd2,
+ SIGKILL, NULL, 0), 0);
+
+ for (i = 0; i < PIDFD_NS_MAX; i++) {
+ if (self->nsfds[i] >= 0)
+ close(self->nsfds[i]);
+ if (self->child_nsfds1[i] >= 0)
+ close(self->child_nsfds1[i]);
+ if (self->child_nsfds2[i] >= 0)
+ close(self->child_nsfds2[i]);
+ }
+
+ if (self->child_pidfd1 >= 0)
+ EXPECT_EQ(0, close(self->child_pidfd1));
+ if (self->child_pidfd2 >= 0)
+ EXPECT_EQ(0, close(self->child_pidfd2));
+ ASSERT_EQ(sys_waitid(P_PID, self->child_pid_exited, WEXITED), 0);
+ ASSERT_EQ(sys_waitid(P_PID, self->child_pid1, WEXITED), 0);
+ ASSERT_EQ(sys_waitid(P_PID, self->child_pid2, WEXITED), 0);
+}
+
+static int preserve_ns(const int pid, const char *ns)
+{
+ int ret;
+ char path[50];
+
+ ret = snprintf(path, sizeof(path), "/proc/%d/ns/%s", pid, ns);
+ if (ret < 0 || (size_t)ret >= sizeof(path))
+ return -EIO;
+
+ return open(path, O_RDONLY | O_CLOEXEC);
+}
+
+static int in_same_namespace(int ns_fd1, pid_t pid2, const char *ns)
+{
+ int ns_fd2 = -EBADF;
+ int ret = -1;
+ struct stat ns_st1, ns_st2;
+
+ ret = fstat(ns_fd1, &ns_st1);
+ if (ret < 0)
+ return -1;
+
+ ns_fd2 = preserve_ns(pid2, ns);
+ if (ns_fd2 < 0)
+ return -1;
+
+ ret = fstat(ns_fd2, &ns_st2);
+ close(ns_fd2);
+ if (ret < 0)
+ return -1;
+
+ /* processes are in the same namespace */
+ if ((ns_st1.st_dev == ns_st2.st_dev) &&
+ (ns_st1.st_ino == ns_st2.st_ino))
+ return 1;
+
+ /* processes are in different namespaces */
+ return 0;
+}
+
+/* Test that we can't pass garbage to the kernel. */
+TEST_F(current_nsset, invalid_flags)
+{
+ ASSERT_NE(setns(self->pidfd, 0), 0);
+ EXPECT_EQ(errno, EINVAL);
+
+ ASSERT_NE(setns(self->pidfd, -1), 0);
+ EXPECT_EQ(errno, EINVAL);
+
+ ASSERT_NE(setns(self->pidfd, CLONE_VM), 0);
+ EXPECT_EQ(errno, EINVAL);
+
+ ASSERT_NE(setns(self->pidfd, CLONE_NEWUSER | CLONE_VM), 0);
+ EXPECT_EQ(errno, EINVAL);
+}
+
+/* Test that we can't attach to a task that has already exited. */
+TEST_F(current_nsset, pidfd_exited_child)
+{
+ int i;
+ pid_t pid;
+
+ ASSERT_NE(setns(self->child_pidfd_exited, CLONE_NEWUSER | CLONE_NEWNET),
+ 0);
+ EXPECT_EQ(errno, ESRCH);
+
+ pid = getpid();
+ for (i = 0; i < PIDFD_NS_MAX; i++) {
+ const struct ns_info *info = &ns_info[i];
+ /* Verify that we haven't changed any namespaces. */
+ if (self->nsfds[i] >= 0)
+ ASSERT_EQ(in_same_namespace(self->nsfds[i], pid, info->name), 1);
+ }
+}
+
+TEST_F(current_nsset, pidfd_incremental_setns)
+{
+ int i;
+ pid_t pid;
+
+ pid = getpid();
+ for (i = 0; i < PIDFD_NS_MAX; i++) {
+ const struct ns_info *info = &ns_info[i];
+ int nsfd;
+
+ if (self->child_nsfds1[i] < 0)
+ continue;
+
+ if (info->flag) {
+ ASSERT_EQ(setns(self->child_pidfd1, info->flag), 0) {
+ TH_LOG("%m - Failed to setns to %s namespace of %d via pidfd %d",
+ info->name, self->child_pid1,
+ self->child_pidfd1);
+ }
+ }
+
+ /* Verify that we have changed to the correct namespaces. */
+ if (info->flag == CLONE_NEWPID)
+ nsfd = self->nsfds[i];
+ else
+ nsfd = self->child_nsfds1[i];
+ ASSERT_EQ(in_same_namespace(nsfd, pid, info->name), 1) {
+ TH_LOG("setns failed to place us correctly into %s namespace of %d via pidfd %d",
+ info->name, self->child_pid1,
+ self->child_pidfd1);
+ }
+ TH_LOG("Managed to correctly setns to %s namespace of %d via pidfd %d",
+ info->name, self->child_pid1, self->child_pidfd1);
+ }
+}
+
+TEST_F(current_nsset, nsfd_incremental_setns)
+{
+ int i;
+ pid_t pid;
+
+ pid = getpid();
+ for (i = 0; i < PIDFD_NS_MAX; i++) {
+ const struct ns_info *info = &ns_info[i];
+ int nsfd;
+
+ if (self->child_nsfds1[i] < 0)
+ continue;
+
+ if (info->flag) {
+ ASSERT_EQ(setns(self->child_nsfds1[i], info->flag), 0) {
+ TH_LOG("%m - Failed to setns to %s namespace of %d via nsfd %d",
+ info->name, self->child_pid1,
+ self->child_nsfds1[i]);
+ }
+ }
+
+ /* Verify that we have changed to the correct namespaces. */
+ if (info->flag == CLONE_NEWPID)
+ nsfd = self->nsfds[i];
+ else
+ nsfd = self->child_nsfds1[i];
+ ASSERT_EQ(in_same_namespace(nsfd, pid, info->name), 1) {
+ TH_LOG("setns failed to place us correctly into %s namespace of %d via nsfd %d",
+ info->name, self->child_pid1,
+ self->child_nsfds1[i]);
+ }
+ TH_LOG("Managed to correctly setns to %s namespace of %d via nsfd %d",
+ info->name, self->child_pid1, self->child_nsfds1[i]);
+ }
+}
+
+TEST_F(current_nsset, pidfd_one_shot_setns)
+{
+ unsigned flags = 0;
+ int i;
+ pid_t pid;
+
+ for (i = 0; i < PIDFD_NS_MAX; i++) {
+ const struct ns_info *info = &ns_info[i];
+
+ if (self->child_nsfds1[i] < 0)
+ continue;
+
+ flags |= info->flag;
+ TH_LOG("Adding %s namespace of %d to list of namespaces to attach to",
+ info->name, self->child_pid1);
+ }
+
+ ASSERT_EQ(setns(self->child_pidfd1, flags), 0) {
+ TH_LOG("%m - Failed to setns to namespaces of %d",
+ self->child_pid1);
+ }
+
+ pid = getpid();
+ for (i = 0; i < PIDFD_NS_MAX; i++) {
+ const struct ns_info *info = &ns_info[i];
+ int nsfd;
+
+ if (self->child_nsfds1[i] < 0)
+ continue;
+
+ /* Verify that we have changed to the correct namespaces. */
+ if (info->flag == CLONE_NEWPID)
+ nsfd = self->nsfds[i];
+ else
+ nsfd = self->child_nsfds1[i];
+ ASSERT_EQ(in_same_namespace(nsfd, pid, info->name), 1) {
+ TH_LOG("setns failed to place us correctly into %s namespace of %d",
+ info->name, self->child_pid1);
+ }
+ TH_LOG("Managed to correctly setns to %s namespace of %d",
+ info->name, self->child_pid1);
+ }
+}
+
+TEST_F(current_nsset, no_foul_play)
+{
+ unsigned flags = 0;
+ int i;
+
+ for (i = 0; i < PIDFD_NS_MAX; i++) {
+ const struct ns_info *info = &ns_info[i];
+
+ if (self->child_nsfds1[i] < 0)
+ continue;
+
+ flags |= info->flag;
+ if (info->flag) /* No use logging pid_for_children. */
+ TH_LOG("Adding %s namespace of %d to list of namespaces to attach to",
+ info->name, self->child_pid1);
+ }
+
+ ASSERT_EQ(setns(self->child_pidfd1, flags), 0) {
+ TH_LOG("%m - Failed to setns to namespaces of %d vid pidfd %d",
+ self->child_pid1, self->child_pidfd1);
+ }
+
+ /*
+ * Can't setns to a user namespace outside of our hierarchy since we
+ * don't have caps in there and didn't create it. That means that under
+ * no circumstances should we be able to setns to any of the other
+ * ones since they aren't owned by our user namespace.
+ */
+ for (i = 0; i < PIDFD_NS_MAX; i++) {
+ const struct ns_info *info = &ns_info[i];
+
+ if (self->child_nsfds2[i] < 0 || !info->flag)
+ continue;
+
+ ASSERT_NE(setns(self->child_pidfd2, info->flag), 0) {
+ TH_LOG("Managed to setns to %s namespace of %d via pidfd %d",
+ info->name, self->child_pid2,
+ self->child_pidfd2);
+ }
+ TH_LOG("%m - Correctly failed to setns to %s namespace of %d via pidfd %d",
+ info->name, self->child_pid2,
+ self->child_pidfd2);
+
+ ASSERT_NE(setns(self->child_nsfds2[i], info->flag), 0) {
+ TH_LOG("Managed to setns to %s namespace of %d via nsfd %d",
+ info->name, self->child_pid2,
+ self->child_nsfds2[i]);
+ }
+ TH_LOG("%m - Correctly failed to setns to %s namespace of %d via nsfd %d",
+ info->name, self->child_pid2,
+ self->child_nsfds2[i]);
+ }
+}
+
+TEST(setns_einval)
+{
+ int fd;
+
+ fd = sys_memfd_create("rostock", 0);
+ EXPECT_GT(fd, 0);
+
+ ASSERT_NE(setns(fd, 0), 0);
+ EXPECT_EQ(errno, EINVAL);
+ close(fd);
+}
+
+TEST_HARNESS_MAIN
diff --git a/tools/testing/selftests/pidfd/pidfd_test.c b/tools/testing/selftests/pidfd/pidfd_test.c
new file mode 100644
index 000000000..cf4f3174c
--- /dev/null
+++ b/tools/testing/selftests/pidfd/pidfd_test.c
@@ -0,0 +1,575 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#define _GNU_SOURCE
+#include <errno.h>
+#include <fcntl.h>
+#include <linux/types.h>
+#include <pthread.h>
+#include <sched.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdbool.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syscall.h>
+#include <sys/epoll.h>
+#include <sys/mman.h>
+#include <sys/mount.h>
+#include <sys/wait.h>
+#include <time.h>
+#include <unistd.h>
+
+#include "pidfd.h"
+#include "../kselftest.h"
+
+#define str(s) _str(s)
+#define _str(s) #s
+#define CHILD_THREAD_MIN_WAIT 3 /* seconds */
+
+#define MAX_EVENTS 5
+
+static bool have_pidfd_send_signal;
+
+static pid_t pidfd_clone(int flags, int *pidfd, int (*fn)(void *))
+{
+ size_t stack_size = 1024;
+ char *stack[1024] = { 0 };
+
+#ifdef __ia64__
+ return __clone2(fn, stack, stack_size, flags | SIGCHLD, NULL, pidfd);
+#else
+ return clone(fn, stack + stack_size, flags | SIGCHLD, NULL, pidfd);
+#endif
+}
+
+static int signal_received;
+
+static void set_signal_received_on_sigusr1(int sig)
+{
+ if (sig == SIGUSR1)
+ signal_received = 1;
+}
+
+/*
+ * Straightforward test to see whether pidfd_send_signal() works is to send
+ * a signal to ourself.
+ */
+static int test_pidfd_send_signal_simple_success(void)
+{
+ int pidfd, ret;
+ const char *test_name = "pidfd_send_signal send SIGUSR1";
+
+ if (!have_pidfd_send_signal) {
+ ksft_test_result_skip(
+ "%s test: pidfd_send_signal() syscall not supported\n",
+ test_name);
+ return 0;
+ }
+
+ pidfd = open("/proc/self", O_DIRECTORY | O_CLOEXEC);
+ if (pidfd < 0)
+ ksft_exit_fail_msg(
+ "%s test: Failed to open process file descriptor\n",
+ test_name);
+
+ signal(SIGUSR1, set_signal_received_on_sigusr1);
+
+ ret = sys_pidfd_send_signal(pidfd, SIGUSR1, NULL, 0);
+ close(pidfd);
+ if (ret < 0)
+ ksft_exit_fail_msg("%s test: Failed to send signal\n",
+ test_name);
+
+ if (signal_received != 1)
+ ksft_exit_fail_msg("%s test: Failed to receive signal\n",
+ test_name);
+
+ signal_received = 0;
+ ksft_test_result_pass("%s test: Sent signal\n", test_name);
+ return 0;
+}
+
+static int test_pidfd_send_signal_exited_fail(void)
+{
+ int pidfd, ret, saved_errno;
+ char buf[256];
+ pid_t pid;
+ const char *test_name = "pidfd_send_signal signal exited process";
+
+ if (!have_pidfd_send_signal) {
+ ksft_test_result_skip(
+ "%s test: pidfd_send_signal() syscall not supported\n",
+ test_name);
+ return 0;
+ }
+
+ pid = fork();
+ if (pid < 0)
+ ksft_exit_fail_msg("%s test: Failed to create new process\n",
+ test_name);
+
+ if (pid == 0)
+ _exit(EXIT_SUCCESS);
+
+ snprintf(buf, sizeof(buf), "/proc/%d", pid);
+
+ pidfd = open(buf, O_DIRECTORY | O_CLOEXEC);
+
+ (void)wait_for_pid(pid);
+
+ if (pidfd < 0)
+ ksft_exit_fail_msg(
+ "%s test: Failed to open process file descriptor\n",
+ test_name);
+
+ ret = sys_pidfd_send_signal(pidfd, 0, NULL, 0);
+ saved_errno = errno;
+ close(pidfd);
+ if (ret == 0)
+ ksft_exit_fail_msg(
+ "%s test: Managed to send signal to process even though it should have failed\n",
+ test_name);
+
+ if (saved_errno != ESRCH)
+ ksft_exit_fail_msg(
+ "%s test: Expected to receive ESRCH as errno value but received %d instead\n",
+ test_name, saved_errno);
+
+ ksft_test_result_pass("%s test: Failed to send signal as expected\n",
+ test_name);
+ return 0;
+}
+
+/*
+ * Maximum number of cycles we allow. This is equivalent to PID_MAX_DEFAULT.
+ * If users set a higher limit or we have cycled PIDFD_MAX_DEFAULT number of
+ * times then we skip the test to not go into an infinite loop or block for a
+ * long time.
+ */
+#define PIDFD_MAX_DEFAULT 0x8000
+
+static int test_pidfd_send_signal_recycled_pid_fail(void)
+{
+ int i, ret;
+ pid_t pid1;
+ const char *test_name = "pidfd_send_signal signal recycled pid";
+
+ if (!have_pidfd_send_signal) {
+ ksft_test_result_skip(
+ "%s test: pidfd_send_signal() syscall not supported\n",
+ test_name);
+ return 0;
+ }
+
+ ret = unshare(CLONE_NEWPID);
+ if (ret < 0) {
+ if (errno == EPERM) {
+ ksft_test_result_skip("%s test: Unsharing pid namespace not permitted\n",
+ test_name);
+ return 0;
+ }
+ ksft_exit_fail_msg("%s test: Failed to unshare pid namespace\n",
+ test_name);
+ }
+
+ ret = unshare(CLONE_NEWNS);
+ if (ret < 0) {
+ if (errno == EPERM) {
+ ksft_test_result_skip("%s test: Unsharing mount namespace not permitted\n",
+ test_name);
+ return 0;
+ }
+ ksft_exit_fail_msg("%s test: Failed to unshare mount namespace\n",
+ test_name);
+ }
+
+ ret = mount(NULL, "/", NULL, MS_REC | MS_PRIVATE, 0);
+ if (ret < 0)
+ ksft_exit_fail_msg("%s test: Failed to remount / private\n",
+ test_name);
+
+ /* pid 1 in new pid namespace */
+ pid1 = fork();
+ if (pid1 < 0)
+ ksft_exit_fail_msg("%s test: Failed to create new process\n",
+ test_name);
+
+ if (pid1 == 0) {
+ char buf[256];
+ pid_t pid2;
+ int pidfd = -1;
+
+ (void)umount2("/proc", MNT_DETACH);
+ ret = mount("proc", "/proc", "proc", 0, NULL);
+ if (ret < 0)
+ _exit(PIDFD_ERROR);
+
+ /* grab pid PID_RECYCLE */
+ for (i = 0; i <= PIDFD_MAX_DEFAULT; i++) {
+ pid2 = fork();
+ if (pid2 < 0)
+ _exit(PIDFD_ERROR);
+
+ if (pid2 == 0)
+ _exit(PIDFD_PASS);
+
+ if (pid2 == PID_RECYCLE) {
+ snprintf(buf, sizeof(buf), "/proc/%d", pid2);
+ ksft_print_msg("pid to recycle is %d\n", pid2);
+ pidfd = open(buf, O_DIRECTORY | O_CLOEXEC);
+ }
+
+ if (wait_for_pid(pid2))
+ _exit(PIDFD_ERROR);
+
+ if (pid2 >= PID_RECYCLE)
+ break;
+ }
+
+ /*
+ * We want to be as predictable as we can so if we haven't been
+ * able to grab pid PID_RECYCLE skip the test.
+ */
+ if (pid2 != PID_RECYCLE) {
+ /* skip test */
+ close(pidfd);
+ _exit(PIDFD_SKIP);
+ }
+
+ if (pidfd < 0)
+ _exit(PIDFD_ERROR);
+
+ for (i = 0; i <= PIDFD_MAX_DEFAULT; i++) {
+ char c;
+ int pipe_fds[2];
+ pid_t recycled_pid;
+ int child_ret = PIDFD_PASS;
+
+ ret = pipe2(pipe_fds, O_CLOEXEC);
+ if (ret < 0)
+ _exit(PIDFD_ERROR);
+
+ recycled_pid = fork();
+ if (recycled_pid < 0)
+ _exit(PIDFD_ERROR);
+
+ if (recycled_pid == 0) {
+ close(pipe_fds[1]);
+ (void)read(pipe_fds[0], &c, 1);
+ close(pipe_fds[0]);
+
+ _exit(PIDFD_PASS);
+ }
+
+ /*
+ * Stop the child so we can inspect whether we have
+ * recycled pid PID_RECYCLE.
+ */
+ close(pipe_fds[0]);
+ ret = kill(recycled_pid, SIGSTOP);
+ close(pipe_fds[1]);
+ if (ret) {
+ (void)wait_for_pid(recycled_pid);
+ _exit(PIDFD_ERROR);
+ }
+
+ /*
+ * We have recycled the pid. Try to signal it. This
+ * needs to fail since this is a different process than
+ * the one the pidfd refers to.
+ */
+ if (recycled_pid == PID_RECYCLE) {
+ ret = sys_pidfd_send_signal(pidfd, SIGCONT,
+ NULL, 0);
+ if (ret && errno == ESRCH)
+ child_ret = PIDFD_XFAIL;
+ else
+ child_ret = PIDFD_FAIL;
+ }
+
+ /* let the process move on */
+ ret = kill(recycled_pid, SIGCONT);
+ if (ret)
+ (void)kill(recycled_pid, SIGKILL);
+
+ if (wait_for_pid(recycled_pid))
+ _exit(PIDFD_ERROR);
+
+ switch (child_ret) {
+ case PIDFD_FAIL:
+ /* fallthrough */
+ case PIDFD_XFAIL:
+ _exit(child_ret);
+ case PIDFD_PASS:
+ break;
+ default:
+ /* not reached */
+ _exit(PIDFD_ERROR);
+ }
+
+ /*
+ * If the user set a custom pid_max limit we could be
+ * in the millions.
+ * Skip the test in this case.
+ */
+ if (recycled_pid > PIDFD_MAX_DEFAULT)
+ _exit(PIDFD_SKIP);
+ }
+
+ /* failed to recycle pid */
+ _exit(PIDFD_SKIP);
+ }
+
+ ret = wait_for_pid(pid1);
+ switch (ret) {
+ case PIDFD_FAIL:
+ ksft_exit_fail_msg(
+ "%s test: Managed to signal recycled pid %d\n",
+ test_name, PID_RECYCLE);
+ case PIDFD_PASS:
+ ksft_exit_fail_msg("%s test: Failed to recycle pid %d\n",
+ test_name, PID_RECYCLE);
+ case PIDFD_SKIP:
+ ksft_test_result_skip("%s test: Skipping test\n", test_name);
+ ret = 0;
+ break;
+ case PIDFD_XFAIL:
+ ksft_test_result_pass(
+ "%s test: Failed to signal recycled pid as expected\n",
+ test_name);
+ ret = 0;
+ break;
+ default /* PIDFD_ERROR */:
+ ksft_exit_fail_msg("%s test: Error while running tests\n",
+ test_name);
+ }
+
+ return ret;
+}
+
+static int test_pidfd_send_signal_syscall_support(void)
+{
+ int pidfd, ret;
+ const char *test_name = "pidfd_send_signal check for support";
+
+ pidfd = open("/proc/self", O_DIRECTORY | O_CLOEXEC);
+ if (pidfd < 0)
+ ksft_exit_fail_msg(
+ "%s test: Failed to open process file descriptor\n",
+ test_name);
+
+ ret = sys_pidfd_send_signal(pidfd, 0, NULL, 0);
+ if (ret < 0) {
+ if (errno == ENOSYS) {
+ ksft_test_result_skip(
+ "%s test: pidfd_send_signal() syscall not supported\n",
+ test_name);
+ return 0;
+ }
+ ksft_exit_fail_msg("%s test: Failed to send signal\n",
+ test_name);
+ }
+
+ have_pidfd_send_signal = true;
+ close(pidfd);
+ ksft_test_result_pass(
+ "%s test: pidfd_send_signal() syscall is supported. Tests can be executed\n",
+ test_name);
+ return 0;
+}
+
+static void *test_pidfd_poll_exec_thread(void *priv)
+{
+ ksft_print_msg("Child Thread: starting. pid %d tid %ld ; and sleeping\n",
+ getpid(), syscall(SYS_gettid));
+ ksft_print_msg("Child Thread: doing exec of sleep\n");
+
+ execl("/bin/sleep", "sleep", str(CHILD_THREAD_MIN_WAIT), (char *)NULL);
+
+ ksft_print_msg("Child Thread: DONE. pid %d tid %ld\n",
+ getpid(), syscall(SYS_gettid));
+ return NULL;
+}
+
+static void poll_pidfd(const char *test_name, int pidfd)
+{
+ int c;
+ int epoll_fd = epoll_create1(EPOLL_CLOEXEC);
+ struct epoll_event event, events[MAX_EVENTS];
+
+ if (epoll_fd == -1)
+ ksft_exit_fail_msg("%s test: Failed to create epoll file descriptor "
+ "(errno %d)\n",
+ test_name, errno);
+
+ event.events = EPOLLIN;
+ event.data.fd = pidfd;
+
+ if (epoll_ctl(epoll_fd, EPOLL_CTL_ADD, pidfd, &event)) {
+ ksft_exit_fail_msg("%s test: Failed to add epoll file descriptor "
+ "(errno %d)\n",
+ test_name, errno);
+ }
+
+ c = epoll_wait(epoll_fd, events, MAX_EVENTS, 5000);
+ if (c != 1 || !(events[0].events & EPOLLIN))
+ ksft_exit_fail_msg("%s test: Unexpected epoll_wait result (c=%d, events=%x) "
+ "(errno %d)\n",
+ test_name, c, events[0].events, errno);
+
+ close(epoll_fd);
+ return;
+
+}
+
+static int child_poll_exec_test(void *args)
+{
+ pthread_t t1;
+
+ ksft_print_msg("Child (pidfd): starting. pid %d tid %ld\n", getpid(),
+ syscall(SYS_gettid));
+ pthread_create(&t1, NULL, test_pidfd_poll_exec_thread, NULL);
+ /*
+ * Exec in the non-leader thread will destroy the leader immediately.
+ * If the wait in the parent returns too soon, the test fails.
+ */
+ while (1)
+ sleep(1);
+
+ return 0;
+}
+
+static void test_pidfd_poll_exec(int use_waitpid)
+{
+ int pid, pidfd = 0;
+ int status, ret;
+ time_t prog_start = time(NULL);
+ const char *test_name = "pidfd_poll check for premature notification on child thread exec";
+
+ ksft_print_msg("Parent: pid: %d\n", getpid());
+ pid = pidfd_clone(CLONE_PIDFD, &pidfd, child_poll_exec_test);
+ if (pid < 0)
+ ksft_exit_fail_msg("%s test: pidfd_clone failed (ret %d, errno %d)\n",
+ test_name, pid, errno);
+
+ ksft_print_msg("Parent: Waiting for Child (%d) to complete.\n", pid);
+
+ if (use_waitpid) {
+ ret = waitpid(pid, &status, 0);
+ if (ret == -1)
+ ksft_print_msg("Parent: error\n");
+
+ if (ret == pid)
+ ksft_print_msg("Parent: Child process waited for.\n");
+ } else {
+ poll_pidfd(test_name, pidfd);
+ }
+
+ time_t prog_time = time(NULL) - prog_start;
+
+ ksft_print_msg("Time waited for child: %lu\n", prog_time);
+
+ close(pidfd);
+
+ if (prog_time < CHILD_THREAD_MIN_WAIT || prog_time > CHILD_THREAD_MIN_WAIT + 2)
+ ksft_exit_fail_msg("%s test: Failed\n", test_name);
+ else
+ ksft_test_result_pass("%s test: Passed\n", test_name);
+}
+
+static void *test_pidfd_poll_leader_exit_thread(void *priv)
+{
+ ksft_print_msg("Child Thread: starting. pid %d tid %ld ; and sleeping\n",
+ getpid(), syscall(SYS_gettid));
+ sleep(CHILD_THREAD_MIN_WAIT);
+ ksft_print_msg("Child Thread: DONE. pid %d tid %ld\n", getpid(), syscall(SYS_gettid));
+ return NULL;
+}
+
+static time_t *child_exit_secs;
+static int child_poll_leader_exit_test(void *args)
+{
+ pthread_t t1, t2;
+
+ ksft_print_msg("Child: starting. pid %d tid %ld\n", getpid(), syscall(SYS_gettid));
+ pthread_create(&t1, NULL, test_pidfd_poll_leader_exit_thread, NULL);
+ pthread_create(&t2, NULL, test_pidfd_poll_leader_exit_thread, NULL);
+
+ /*
+ * glibc exit calls exit_group syscall, so explicity call exit only
+ * so that only the group leader exits, leaving the threads alone.
+ */
+ *child_exit_secs = time(NULL);
+ syscall(SYS_exit, 0);
+ /* Never reached, but appeases compiler thinking we should return. */
+ exit(0);
+}
+
+static void test_pidfd_poll_leader_exit(int use_waitpid)
+{
+ int pid, pidfd = 0;
+ int status, ret = 0;
+ const char *test_name = "pidfd_poll check for premature notification on non-empty"
+ "group leader exit";
+
+ child_exit_secs = mmap(NULL, sizeof *child_exit_secs, PROT_READ | PROT_WRITE,
+ MAP_SHARED | MAP_ANONYMOUS, -1, 0);
+
+ if (child_exit_secs == MAP_FAILED)
+ ksft_exit_fail_msg("%s test: mmap failed (errno %d)\n",
+ test_name, errno);
+
+ ksft_print_msg("Parent: pid: %d\n", getpid());
+ pid = pidfd_clone(CLONE_PIDFD, &pidfd, child_poll_leader_exit_test);
+ if (pid < 0)
+ ksft_exit_fail_msg("%s test: pidfd_clone failed (ret %d, errno %d)\n",
+ test_name, pid, errno);
+
+ ksft_print_msg("Parent: Waiting for Child (%d) to complete.\n", pid);
+
+ if (use_waitpid) {
+ ret = waitpid(pid, &status, 0);
+ if (ret == -1)
+ ksft_print_msg("Parent: error\n");
+ } else {
+ /*
+ * This sleep tests for the case where if the child exits, and is in
+ * EXIT_ZOMBIE, but the thread group leader is non-empty, then the poll
+ * doesn't prematurely return even though there are active threads
+ */
+ sleep(1);
+ poll_pidfd(test_name, pidfd);
+ }
+
+ if (ret == pid)
+ ksft_print_msg("Parent: Child process waited for.\n");
+
+ time_t since_child_exit = time(NULL) - *child_exit_secs;
+
+ ksft_print_msg("Time since child exit: %lu\n", since_child_exit);
+
+ close(pidfd);
+
+ if (since_child_exit < CHILD_THREAD_MIN_WAIT ||
+ since_child_exit > CHILD_THREAD_MIN_WAIT + 2)
+ ksft_exit_fail_msg("%s test: Failed\n", test_name);
+ else
+ ksft_test_result_pass("%s test: Passed\n", test_name);
+}
+
+int main(int argc, char **argv)
+{
+ ksft_print_header();
+ ksft_set_plan(8);
+
+ test_pidfd_poll_exec(0);
+ test_pidfd_poll_exec(1);
+ test_pidfd_poll_leader_exit(0);
+ test_pidfd_poll_leader_exit(1);
+ test_pidfd_send_signal_syscall_support();
+ test_pidfd_send_signal_simple_success();
+ test_pidfd_send_signal_exited_fail();
+ test_pidfd_send_signal_recycled_pid_fail();
+
+ return ksft_exit_pass();
+}
diff --git a/tools/testing/selftests/pidfd/pidfd_wait.c b/tools/testing/selftests/pidfd/pidfd_wait.c
new file mode 100644
index 000000000..0dcb8365d
--- /dev/null
+++ b/tools/testing/selftests/pidfd/pidfd_wait.c
@@ -0,0 +1,233 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#define _GNU_SOURCE
+#include <errno.h>
+#include <linux/sched.h>
+#include <linux/types.h>
+#include <signal.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sched.h>
+#include <string.h>
+#include <sys/resource.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+#include "pidfd.h"
+#include "../kselftest_harness.h"
+
+#define ptr_to_u64(ptr) ((__u64)((uintptr_t)(ptr)))
+
+/* Attempt to de-conflict with the selftests tree. */
+#ifndef SKIP
+#define SKIP(s, ...) XFAIL(s, ##__VA_ARGS__)
+#endif
+
+static pid_t sys_clone3(struct clone_args *args)
+{
+ return syscall(__NR_clone3, args, sizeof(struct clone_args));
+}
+
+static int sys_waitid(int which, pid_t pid, siginfo_t *info, int options,
+ struct rusage *ru)
+{
+ return syscall(__NR_waitid, which, pid, info, options, ru);
+}
+
+TEST(wait_simple)
+{
+ int pidfd = -1;
+ pid_t parent_tid = -1;
+ struct clone_args args = {
+ .parent_tid = ptr_to_u64(&parent_tid),
+ .pidfd = ptr_to_u64(&pidfd),
+ .flags = CLONE_PIDFD | CLONE_PARENT_SETTID,
+ .exit_signal = SIGCHLD,
+ };
+ pid_t pid;
+ siginfo_t info = {
+ .si_signo = 0,
+ };
+
+ pidfd = open("/proc/self", O_DIRECTORY | O_RDONLY | O_CLOEXEC);
+ ASSERT_GE(pidfd, 0);
+
+ pid = sys_waitid(P_PIDFD, pidfd, &info, WEXITED, NULL);
+ ASSERT_NE(pid, 0);
+ EXPECT_EQ(close(pidfd), 0);
+ pidfd = -1;
+
+ pidfd = open("/dev/null", O_RDONLY | O_CLOEXEC);
+ ASSERT_GE(pidfd, 0);
+
+ pid = sys_waitid(P_PIDFD, pidfd, &info, WEXITED, NULL);
+ ASSERT_NE(pid, 0);
+ EXPECT_EQ(close(pidfd), 0);
+ pidfd = -1;
+
+ pid = sys_clone3(&args);
+ ASSERT_GE(pid, 0);
+
+ if (pid == 0)
+ exit(EXIT_SUCCESS);
+
+ pid = sys_waitid(P_PIDFD, pidfd, &info, WEXITED, NULL);
+ ASSERT_GE(pid, 0);
+ ASSERT_EQ(WIFEXITED(info.si_status), true);
+ ASSERT_EQ(WEXITSTATUS(info.si_status), 0);
+ EXPECT_EQ(close(pidfd), 0);
+
+ ASSERT_EQ(info.si_signo, SIGCHLD);
+ ASSERT_EQ(info.si_code, CLD_EXITED);
+ ASSERT_EQ(info.si_pid, parent_tid);
+}
+
+TEST(wait_states)
+{
+ int pidfd = -1;
+ pid_t parent_tid = -1;
+ struct clone_args args = {
+ .parent_tid = ptr_to_u64(&parent_tid),
+ .pidfd = ptr_to_u64(&pidfd),
+ .flags = CLONE_PIDFD | CLONE_PARENT_SETTID,
+ .exit_signal = SIGCHLD,
+ };
+ int pfd[2];
+ pid_t pid;
+ siginfo_t info = {
+ .si_signo = 0,
+ };
+
+ ASSERT_EQ(pipe(pfd), 0);
+ pid = sys_clone3(&args);
+ ASSERT_GE(pid, 0);
+
+ if (pid == 0) {
+ char buf[2];
+
+ close(pfd[1]);
+ kill(getpid(), SIGSTOP);
+ ASSERT_EQ(read(pfd[0], buf, 1), 1);
+ close(pfd[0]);
+ kill(getpid(), SIGSTOP);
+ exit(EXIT_SUCCESS);
+ }
+
+ close(pfd[0]);
+ ASSERT_EQ(sys_waitid(P_PIDFD, pidfd, &info, WSTOPPED, NULL), 0);
+ ASSERT_EQ(info.si_signo, SIGCHLD);
+ ASSERT_EQ(info.si_code, CLD_STOPPED);
+ ASSERT_EQ(info.si_pid, parent_tid);
+
+ ASSERT_EQ(sys_pidfd_send_signal(pidfd, SIGCONT, NULL, 0), 0);
+
+ ASSERT_EQ(sys_waitid(P_PIDFD, pidfd, &info, WCONTINUED, NULL), 0);
+ ASSERT_EQ(write(pfd[1], "C", 1), 1);
+ close(pfd[1]);
+ ASSERT_EQ(info.si_signo, SIGCHLD);
+ ASSERT_EQ(info.si_code, CLD_CONTINUED);
+ ASSERT_EQ(info.si_pid, parent_tid);
+
+ ASSERT_EQ(sys_waitid(P_PIDFD, pidfd, &info, WUNTRACED, NULL), 0);
+ ASSERT_EQ(info.si_signo, SIGCHLD);
+ ASSERT_EQ(info.si_code, CLD_STOPPED);
+ ASSERT_EQ(info.si_pid, parent_tid);
+
+ ASSERT_EQ(sys_pidfd_send_signal(pidfd, SIGKILL, NULL, 0), 0);
+
+ ASSERT_EQ(sys_waitid(P_PIDFD, pidfd, &info, WEXITED, NULL), 0);
+ ASSERT_EQ(info.si_signo, SIGCHLD);
+ ASSERT_EQ(info.si_code, CLD_KILLED);
+ ASSERT_EQ(info.si_pid, parent_tid);
+
+ EXPECT_EQ(close(pidfd), 0);
+}
+
+TEST(wait_nonblock)
+{
+ int pidfd;
+ unsigned int flags = 0;
+ pid_t parent_tid = -1;
+ struct clone_args args = {
+ .parent_tid = ptr_to_u64(&parent_tid),
+ .flags = CLONE_PARENT_SETTID,
+ .exit_signal = SIGCHLD,
+ };
+ int ret;
+ pid_t pid;
+ siginfo_t info = {
+ .si_signo = 0,
+ };
+
+ /*
+ * Callers need to see ECHILD with non-blocking pidfds when no child
+ * processes exists.
+ */
+ pidfd = sys_pidfd_open(getpid(), PIDFD_NONBLOCK);
+ EXPECT_GE(pidfd, 0) {
+ /* pidfd_open() doesn't support PIDFD_NONBLOCK. */
+ ASSERT_EQ(errno, EINVAL);
+ SKIP(return, "Skipping PIDFD_NONBLOCK test");
+ }
+
+ ret = sys_waitid(P_PIDFD, pidfd, &info, WEXITED, NULL);
+ ASSERT_LT(ret, 0);
+ ASSERT_EQ(errno, ECHILD);
+ EXPECT_EQ(close(pidfd), 0);
+
+ pid = sys_clone3(&args);
+ ASSERT_GE(pid, 0);
+
+ if (pid == 0) {
+ kill(getpid(), SIGSTOP);
+ exit(EXIT_SUCCESS);
+ }
+
+ pidfd = sys_pidfd_open(pid, PIDFD_NONBLOCK);
+ EXPECT_GE(pidfd, 0) {
+ /* pidfd_open() doesn't support PIDFD_NONBLOCK. */
+ ASSERT_EQ(errno, EINVAL);
+ SKIP(return, "Skipping PIDFD_NONBLOCK test");
+ }
+
+ flags = fcntl(pidfd, F_GETFL, 0);
+ ASSERT_GT(flags, 0);
+ ASSERT_GT((flags & O_NONBLOCK), 0);
+
+ /*
+ * Callers need to see EAGAIN/EWOULDBLOCK with non-blocking pidfd when
+ * child processes exist but none have exited.
+ */
+ ret = sys_waitid(P_PIDFD, pidfd, &info, WEXITED, NULL);
+ ASSERT_LT(ret, 0);
+ ASSERT_EQ(errno, EAGAIN);
+
+ /*
+ * Callers need to continue seeing 0 with non-blocking pidfd and
+ * WNOHANG raised explicitly when child processes exist but none have
+ * exited.
+ */
+ ret = sys_waitid(P_PIDFD, pidfd, &info, WEXITED | WNOHANG, NULL);
+ ASSERT_EQ(ret, 0);
+
+ ASSERT_EQ(fcntl(pidfd, F_SETFL, (flags & ~O_NONBLOCK)), 0);
+
+ ASSERT_EQ(sys_waitid(P_PIDFD, pidfd, &info, WSTOPPED, NULL), 0);
+ ASSERT_EQ(info.si_signo, SIGCHLD);
+ ASSERT_EQ(info.si_code, CLD_STOPPED);
+ ASSERT_EQ(info.si_pid, parent_tid);
+
+ ASSERT_EQ(sys_pidfd_send_signal(pidfd, SIGCONT, NULL, 0), 0);
+
+ ASSERT_EQ(sys_waitid(P_PIDFD, pidfd, &info, WEXITED, NULL), 0);
+ ASSERT_EQ(info.si_signo, SIGCHLD);
+ ASSERT_EQ(info.si_code, CLD_EXITED);
+ ASSERT_EQ(info.si_pid, parent_tid);
+
+ EXPECT_EQ(close(pidfd), 0);
+}
+
+TEST_HARNESS_MAIN