Merge tag 'threads-v5.5' of git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux

Pull thread management updates from Christian Brauner:

 - A pidfd's fdinfo file currently contains the field "Pid:\t<pid>"
   where <pid> is the pid of the process in the pid namespace of the
   procfs instance the fdinfo file for the pidfd was opened in.

   The fdinfo file has now gained a new "NSpid:\t<ns-pid1>[\t<ns-pid2>[...]]"
   field which lists the pids of the process in all child pid namespaces
   provided the pid namespace of the procfs instance it is looked up
   under has an ancestoral relationship with the pid namespace of the
   process. If it does not 0 will be shown and no further pid namespaces
   will be listed. Tests included. (Christian Kellner)

 - If the process the pidfd references has already exited, print -1 for
   the Pid and NSpid fields in the pidfd's fdinfo file. Tests included.
   (me)

 - Add CLONE_CLEAR_SIGHAND. This lets callers clear all signal handler
   that are not SIG_DFL or SIG_IGN at process creation time. This
   originated as a feature request from glibc to improve performance and
   elimate races in their posix_spawn() implementation. Tests included.
   (me)

 - Add support for choosing a specific pid for a process with clone3().
   This is the feature which was part of the thread update for v5.4 but
   after a discussion at LPC in Lisbon we decided to delay it for one
   more cycle in order to make the interface more generic. This has now
   done. It is now possible to choose a specific pid in a whole pid
   namespaces (sub)hierarchy instead of just one pid namespace. In order
   to choose a specific pid the caller must have CAP_SYS_ADMIN in all
   owning user namespaces of the target pid namespaces. Tests included.
   (Adrian Reber)

 - Test improvements and extensions. (Andrei Vagin, me)

* tag 'threads-v5.5' of git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux:
  selftests/clone3: skip if clone3() is ENOSYS
  selftests/clone3: check that all pids are released on error paths
  selftests/clone3: report a correct number of fails
  selftests/clone3: flush stdout and stderr before clone3() and _exit()
  selftests: add tests for clone3() with *set_tid
  fork: extend clone3() to support setting a PID
  selftests: add tests for clone3()
  tests: test CLONE_CLEAR_SIGHAND
  clone3: add CLONE_CLEAR_SIGHAND
  pid: use pid_has_task() in pidfd_open()
  exit: use pid_has_task() in do_wait()
  pid: use pid_has_task() in __change_pid()
  test: verify fdinfo for pidfd of reaped process
  pidfd: check pid has attached task in fdinfo
  pidfd: add tests for NSpid info in fdinfo
  pidfd: add NSpid entries to fdinfo

1 files changed, 202 insertions, 0 deletions

diff --git a/tools/testing/selftests/clone3/clone3.c b/tools/testing/selftests/clone3/clone3.c
new file mode 100644
index 000000000000..f14c269a5a18
--- /dev/null
+++ b/tools/testing/selftests/clone3/clone3.c
@@ -0,0 +1,202 @@
+// SPDX-License-Identifier: GPL-2.0
+
+/* Based on Christian Brauner's clone3() example */
+
+#define _GNU_SOURCE
+#include <errno.h>
+#include <inttypes.h>
+#include <linux/types.h>
+#include <linux/sched.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/syscall.h>
+#include <sys/types.h>
+#include <sys/un.h>
+#include <sys/wait.h>
+#include <unistd.h>
+#include <sched.h>
+
+#include "../kselftest.h"
+#include "clone3_selftests.h"
+
+/*
+ * Different sizes of struct clone_args
+ */
+#ifndef CLONE3_ARGS_SIZE_V0
+#define CLONE3_ARGS_SIZE_V0 64
+#endif
+
+enum test_mode {
+	CLONE3_ARGS_NO_TEST,
+	CLONE3_ARGS_ALL_0,
+	CLONE3_ARGS_INVAL_EXIT_SIGNAL_BIG,
+	CLONE3_ARGS_INVAL_EXIT_SIGNAL_NEG,
+	CLONE3_ARGS_INVAL_EXIT_SIGNAL_CSIG,
+	CLONE3_ARGS_INVAL_EXIT_SIGNAL_NSIG,
+};
+
+static int call_clone3(uint64_t flags, size_t size, enum test_mode test_mode)
+{
+	struct clone_args args = {
+		.flags = flags,
+		.exit_signal = SIGCHLD,
+	};
+
+	struct clone_args_extended {
+		struct clone_args args;
+		__aligned_u64 excess_space[2];
+	} args_ext;
+
+	pid_t pid = -1;
+	int status;
+
+	memset(&args_ext, 0, sizeof(args_ext));
+	if (size > sizeof(struct clone_args))
+		args_ext.excess_space[1] = 1;
+
+	if (size == 0)
+		size = sizeof(struct clone_args);
+
+	switch (test_mode) {
+	case CLONE3_ARGS_ALL_0:
+		args.flags = 0;
+		args.exit_signal = 0;
+		break;
+	case CLONE3_ARGS_INVAL_EXIT_SIGNAL_BIG:
+		args.exit_signal = 0xbadc0ded00000000ULL;
+		break;
+	case CLONE3_ARGS_INVAL_EXIT_SIGNAL_NEG:
+		args.exit_signal = 0x0000000080000000ULL;
+		break;
+	case CLONE3_ARGS_INVAL_EXIT_SIGNAL_CSIG:
+		args.exit_signal = 0x0000000000000100ULL;
+		break;
+	case CLONE3_ARGS_INVAL_EXIT_SIGNAL_NSIG:
+		args.exit_signal = 0x00000000000000f0ULL;
+		break;
+	}
+
+	memcpy(&args_ext.args, &args, sizeof(struct clone_args));
+
+	pid = sys_clone3((struct clone_args *)&args_ext, size);
+	if (pid < 0) {
+		ksft_print_msg("%s - Failed to create new process\n",
+				strerror(errno));
+		return -errno;
+	}
+
+	if (pid == 0) {
+		ksft_print_msg("I am the child, my PID is %d\n", getpid());
+		_exit(EXIT_SUCCESS);
+	}
+
+	ksft_print_msg("I am the parent (%d). My child's pid is %d\n",
+			getpid(), pid);
+
+	if (waitpid(-1, &status, __WALL) < 0) {
+		ksft_print_msg("Child returned %s\n", strerror(errno));
+		return -errno;
+	}
+	if (WEXITSTATUS(status))
+		return WEXITSTATUS(status);
+
+	return 0;
+}
+
+static void test_clone3(uint64_t flags, size_t size, int expected,
+		       enum test_mode test_mode)
+{
+	int ret;
+
+	ksft_print_msg(
+		"[%d] Trying clone3() with flags %#" PRIx64 " (size %zu)\n",
+		getpid(), flags, size);
+	ret = call_clone3(flags, size, test_mode);
+	ksft_print_msg("[%d] clone3() with flags says: %d expected %d\n",
+			getpid(), ret, expected);
+	if (ret != expected)
+		ksft_test_result_fail(
+			"[%d] Result (%d) is different than expected (%d)\n",
+			getpid(), ret, expected);
+	else
+		ksft_test_result_pass(
+			"[%d] Result (%d) matches expectation (%d)\n",
+			getpid(), ret, expected);
+}
+
+int main(int argc, char *argv[])
+{
+	pid_t pid;
+
+	uid_t uid = getuid();
+
+	test_clone3_supported();
+	ksft_print_header();
+	ksft_set_plan(17);
+
+	/* Just a simple clone3() should return 0.*/
+	test_clone3(0, 0, 0, CLONE3_ARGS_NO_TEST);
+
+	/* Do a clone3() in a new PID NS.*/
+	if (uid == 0)
+		test_clone3(CLONE_NEWPID, 0, 0, CLONE3_ARGS_NO_TEST);
+	else
+		ksft_test_result_skip("Skipping clone3() with CLONE_NEWPID\n");
+
+	/* Do a clone3() with CLONE3_ARGS_SIZE_V0. */
+	test_clone3(0, CLONE3_ARGS_SIZE_V0, 0, CLONE3_ARGS_NO_TEST);
+
+	/* Do a clone3() with CLONE3_ARGS_SIZE_V0 - 8 */
+	test_clone3(0, CLONE3_ARGS_SIZE_V0 - 8, -EINVAL, CLONE3_ARGS_NO_TEST);
+
+	/* Do a clone3() with sizeof(struct clone_args) + 8 */
+	test_clone3(0, sizeof(struct clone_args) + 8, 0, CLONE3_ARGS_NO_TEST);
+
+	/* Do a clone3() with exit_signal having highest 32 bits non-zero */
+	test_clone3(0, 0, -EINVAL, CLONE3_ARGS_INVAL_EXIT_SIGNAL_BIG);
+
+	/* Do a clone3() with negative 32-bit exit_signal */
+	test_clone3(0, 0, -EINVAL, CLONE3_ARGS_INVAL_EXIT_SIGNAL_NEG);
+
+	/* Do a clone3() with exit_signal not fitting into CSIGNAL mask */
+	test_clone3(0, 0, -EINVAL, CLONE3_ARGS_INVAL_EXIT_SIGNAL_CSIG);
+
+	/* Do a clone3() with NSIG < exit_signal < CSIG */
+	test_clone3(0, 0, -EINVAL, CLONE3_ARGS_INVAL_EXIT_SIGNAL_NSIG);
+
+	test_clone3(0, sizeof(struct clone_args) + 8, 0, CLONE3_ARGS_ALL_0);
+
+	test_clone3(0, sizeof(struct clone_args) + 16, -E2BIG,
+			CLONE3_ARGS_ALL_0);
+
+	test_clone3(0, sizeof(struct clone_args) * 2, -E2BIG,
+			CLONE3_ARGS_ALL_0);
+
+	/* Do a clone3() with > page size */
+	test_clone3(0, getpagesize() + 8, -E2BIG, CLONE3_ARGS_NO_TEST);
+
+	/* Do a clone3() with CLONE3_ARGS_SIZE_V0 in a new PID NS. */
+	if (uid == 0)
+		test_clone3(CLONE_NEWPID, CLONE3_ARGS_SIZE_V0, 0,
+				CLONE3_ARGS_NO_TEST);
+	else
+		ksft_test_result_skip("Skipping clone3() with CLONE_NEWPID\n");
+
+	/* Do a clone3() with CLONE3_ARGS_SIZE_V0 - 8 in a new PID NS */
+	test_clone3(CLONE_NEWPID, CLONE3_ARGS_SIZE_V0 - 8, -EINVAL,
+			CLONE3_ARGS_NO_TEST);
+
+	/* Do a clone3() with sizeof(struct clone_args) + 8 in a new PID NS */
+	if (uid == 0)
+		test_clone3(CLONE_NEWPID, sizeof(struct clone_args) + 8, 0,
+				CLONE3_ARGS_NO_TEST);
+	else
+		ksft_test_result_skip("Skipping clone3() with CLONE_NEWPID\n");
+
+	/* Do a clone3() with > page size in a new PID NS */
+	test_clone3(CLONE_NEWPID, getpagesize() + 8, -E2BIG,
+			CLONE3_ARGS_NO_TEST);
+
+	return !ksft_get_fail_cnt() ? ksft_exit_pass() : ksft_exit_fail();
+}