summaryrefslogtreecommitdiff
path: root/tools/testing/selftests/filesystems
diff options
context:
space:
mode:
Diffstat (limited to 'tools/testing/selftests/filesystems')
-rw-r--r--tools/testing/selftests/filesystems/.gitignore1
-rw-r--r--tools/testing/selftests/filesystems/Makefile2
-rw-r--r--tools/testing/selftests/filesystems/binderfs/Makefile2
-rw-r--r--tools/testing/selftests/filesystems/binderfs/binderfs_test.c2
-rw-r--r--tools/testing/selftests/filesystems/eventfd/.gitignore2
-rw-r--r--tools/testing/selftests/filesystems/eventfd/Makefile7
-rw-r--r--tools/testing/selftests/filesystems/eventfd/eventfd_test.c312
-rw-r--r--tools/testing/selftests/filesystems/file_stressor.c194
-rw-r--r--tools/testing/selftests/filesystems/nsfs/.gitignore4
-rw-r--r--tools/testing/selftests/filesystems/nsfs/Makefile6
-rw-r--r--tools/testing/selftests/filesystems/nsfs/config3
-rw-r--r--tools/testing/selftests/filesystems/nsfs/iterate_mntns.c149
-rw-r--r--tools/testing/selftests/filesystems/nsfs/owner.c92
-rw-r--r--tools/testing/selftests/filesystems/nsfs/pidns.c79
-rw-r--r--tools/testing/selftests/filesystems/overlayfs/.gitignore1
-rw-r--r--tools/testing/selftests/filesystems/overlayfs/Makefile2
-rw-r--r--tools/testing/selftests/filesystems/overlayfs/dev_in_maps.c28
-rw-r--r--tools/testing/selftests/filesystems/overlayfs/set_layers_via_fds.c217
-rw-r--r--tools/testing/selftests/filesystems/overlayfs/wrappers.h47
-rw-r--r--tools/testing/selftests/filesystems/statmount/.gitignore1
-rw-r--r--tools/testing/selftests/filesystems/statmount/Makefile2
-rw-r--r--tools/testing/selftests/filesystems/statmount/listmount_test.c66
-rw-r--r--tools/testing/selftests/filesystems/statmount/statmount.h46
-rw-r--r--tools/testing/selftests/filesystems/statmount/statmount_test.c179
-rw-r--r--tools/testing/selftests/filesystems/statmount/statmount_test_ns.c367
25 files changed, 1736 insertions, 75 deletions
diff --git a/tools/testing/selftests/filesystems/.gitignore b/tools/testing/selftests/filesystems/.gitignore
index f0c0ff20d6cf..828b66a10c63 100644
--- a/tools/testing/selftests/filesystems/.gitignore
+++ b/tools/testing/selftests/filesystems/.gitignore
@@ -1,3 +1,4 @@
# SPDX-License-Identifier: GPL-2.0-only
dnotify_test
devpts_pts
+file_stressor
diff --git a/tools/testing/selftests/filesystems/Makefile b/tools/testing/selftests/filesystems/Makefile
index c647fd6a0446..66305fc34c60 100644
--- a/tools/testing/selftests/filesystems/Makefile
+++ b/tools/testing/selftests/filesystems/Makefile
@@ -1,7 +1,7 @@
# SPDX-License-Identifier: GPL-2.0
CFLAGS += $(KHDR_INCLUDES)
-TEST_GEN_PROGS := devpts_pts
+TEST_GEN_PROGS := devpts_pts file_stressor
TEST_GEN_PROGS_EXTENDED := dnotify_test
include ../lib.mk
diff --git a/tools/testing/selftests/filesystems/binderfs/Makefile b/tools/testing/selftests/filesystems/binderfs/Makefile
index c2f7cef919c0..eb4c3b411934 100644
--- a/tools/testing/selftests/filesystems/binderfs/Makefile
+++ b/tools/testing/selftests/filesystems/binderfs/Makefile
@@ -3,6 +3,4 @@
CFLAGS += $(KHDR_INCLUDES) -pthread
TEST_GEN_PROGS := binderfs_test
-binderfs_test: binderfs_test.c ../../kselftest.h ../../kselftest_harness.h
-
include ../../lib.mk
diff --git a/tools/testing/selftests/filesystems/binderfs/binderfs_test.c b/tools/testing/selftests/filesystems/binderfs/binderfs_test.c
index 5f362c0fd890..81db85a5cc16 100644
--- a/tools/testing/selftests/filesystems/binderfs/binderfs_test.c
+++ b/tools/testing/selftests/filesystems/binderfs/binderfs_test.c
@@ -57,7 +57,6 @@ static int __do_binderfs_test(struct __test_metadata *_metadata)
{
int fd, ret, saved_errno, result = 1;
size_t len;
- ssize_t wret;
struct binderfs_device device = { 0 };
struct binder_version version = { 0 };
char binderfs_mntpt[] = P_tmpdir "/binderfs_XXXXXX",
@@ -65,6 +64,7 @@ static int __do_binderfs_test(struct __test_metadata *_metadata)
static const char * const binder_features[] = {
"oneway_spam_detection",
"extended_error",
+ "freeze_notification",
};
change_mountns(_metadata);
diff --git a/tools/testing/selftests/filesystems/eventfd/.gitignore b/tools/testing/selftests/filesystems/eventfd/.gitignore
new file mode 100644
index 000000000000..483faf59fe4a
--- /dev/null
+++ b/tools/testing/selftests/filesystems/eventfd/.gitignore
@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0-only
+eventfd_test
diff --git a/tools/testing/selftests/filesystems/eventfd/Makefile b/tools/testing/selftests/filesystems/eventfd/Makefile
new file mode 100644
index 000000000000..0a8e3910df15
--- /dev/null
+++ b/tools/testing/selftests/filesystems/eventfd/Makefile
@@ -0,0 +1,7 @@
+# SPDX-License-Identifier: GPL-2.0
+
+CFLAGS += $(KHDR_INCLUDES)
+LDLIBS += -lpthread
+TEST_GEN_PROGS := eventfd_test
+
+include ../../lib.mk
diff --git a/tools/testing/selftests/filesystems/eventfd/eventfd_test.c b/tools/testing/selftests/filesystems/eventfd/eventfd_test.c
new file mode 100644
index 000000000000..85acb4e3ef00
--- /dev/null
+++ b/tools/testing/selftests/filesystems/eventfd/eventfd_test.c
@@ -0,0 +1,312 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#define _GNU_SOURCE
+#include <errno.h>
+#include <fcntl.h>
+#include <asm/unistd.h>
+#include <linux/time_types.h>
+#include <unistd.h>
+#include <assert.h>
+#include <signal.h>
+#include <pthread.h>
+#include <sys/epoll.h>
+#include <sys/eventfd.h>
+#include "../../kselftest_harness.h"
+
+#define EVENTFD_TEST_ITERATIONS 100000UL
+
+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 == 0 || !err || err->code != 0)
+ 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 inline int sys_eventfd2(unsigned int count, int flags)
+{
+ return syscall(__NR_eventfd2, count, flags);
+}
+
+TEST(eventfd_check_flag_rdwr)
+{
+ int fd, flags;
+
+ fd = sys_eventfd2(0, 0);
+ ASSERT_GE(fd, 0);
+
+ flags = fcntl(fd, F_GETFL);
+ // since the kernel automatically added O_RDWR.
+ EXPECT_EQ(flags, O_RDWR);
+
+ close(fd);
+}
+
+TEST(eventfd_check_flag_cloexec)
+{
+ int fd, flags;
+
+ fd = sys_eventfd2(0, EFD_CLOEXEC);
+ ASSERT_GE(fd, 0);
+
+ flags = fcntl(fd, F_GETFD);
+ ASSERT_GT(flags, -1);
+ EXPECT_EQ(flags, FD_CLOEXEC);
+
+ close(fd);
+}
+
+TEST(eventfd_check_flag_nonblock)
+{
+ int fd, flags;
+
+ fd = sys_eventfd2(0, EFD_NONBLOCK);
+ ASSERT_GE(fd, 0);
+
+ flags = fcntl(fd, F_GETFL);
+ ASSERT_GT(flags, -1);
+ EXPECT_EQ(flags & EFD_NONBLOCK, EFD_NONBLOCK);
+ EXPECT_EQ(flags & O_RDWR, O_RDWR);
+
+ close(fd);
+}
+
+TEST(eventfd_chek_flag_cloexec_and_nonblock)
+{
+ int fd, flags;
+
+ fd = sys_eventfd2(0, EFD_CLOEXEC|EFD_NONBLOCK);
+ ASSERT_GE(fd, 0);
+
+ flags = fcntl(fd, F_GETFL);
+ ASSERT_GT(flags, -1);
+ EXPECT_EQ(flags & EFD_NONBLOCK, EFD_NONBLOCK);
+ EXPECT_EQ(flags & O_RDWR, O_RDWR);
+
+ flags = fcntl(fd, F_GETFD);
+ ASSERT_GT(flags, -1);
+ EXPECT_EQ(flags, FD_CLOEXEC);
+
+ close(fd);
+}
+
+static inline void trim_newline(char *str)
+{
+ char *pos = strrchr(str, '\n');
+
+ if (pos)
+ *pos = '\0';
+}
+
+static int verify_fdinfo(int fd, 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", fd);
+ f = fopen(path, "re");
+ if (!f)
+ return error_set(err, -1, "fdinfo open failed for %d", fd);
+
+ 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, -1, "%s '%s' != '%s'",
+ prefix, val, buffer);
+ }
+ break;
+ }
+
+ free(line);
+ fclose(f);
+
+ if (found == 0)
+ return error_set(err, -1, "%s not found for fd %d",
+ prefix, fd);
+
+ return 0;
+}
+
+TEST(eventfd_check_flag_semaphore)
+{
+ struct error err = {0};
+ int fd, ret;
+
+ fd = sys_eventfd2(0, EFD_SEMAPHORE);
+ ASSERT_GE(fd, 0);
+
+ ret = fcntl(fd, F_GETFL);
+ ASSERT_GT(ret, -1);
+ EXPECT_EQ(ret & O_RDWR, O_RDWR);
+
+ // The semaphore could only be obtained from fdinfo.
+ ret = verify_fdinfo(fd, &err, "eventfd-semaphore: ", 19, "1\n");
+ if (ret != 0)
+ ksft_print_msg("eventfd-semaphore check failed, msg: %s\n",
+ err.msg);
+ EXPECT_EQ(ret, 0);
+
+ close(fd);
+}
+
+/*
+ * A write(2) fails with the error EINVAL if the size of the supplied buffer
+ * is less than 8 bytes, or if an attempt is made to write the value
+ * 0xffffffffffffffff.
+ */
+TEST(eventfd_check_write)
+{
+ uint64_t value = 1;
+ ssize_t size;
+ int fd;
+
+ fd = sys_eventfd2(0, 0);
+ ASSERT_GE(fd, 0);
+
+ size = write(fd, &value, sizeof(int));
+ EXPECT_EQ(size, -1);
+ EXPECT_EQ(errno, EINVAL);
+
+ size = write(fd, &value, sizeof(value));
+ EXPECT_EQ(size, sizeof(value));
+
+ value = (uint64_t)-1;
+ size = write(fd, &value, sizeof(value));
+ EXPECT_EQ(size, -1);
+ EXPECT_EQ(errno, EINVAL);
+
+ close(fd);
+}
+
+/*
+ * A read(2) fails with the error EINVAL if the size of the supplied buffer is
+ * less than 8 bytes.
+ */
+TEST(eventfd_check_read)
+{
+ uint64_t value;
+ ssize_t size;
+ int fd;
+
+ fd = sys_eventfd2(1, 0);
+ ASSERT_GE(fd, 0);
+
+ size = read(fd, &value, sizeof(int));
+ EXPECT_EQ(size, -1);
+ EXPECT_EQ(errno, EINVAL);
+
+ size = read(fd, &value, sizeof(value));
+ EXPECT_EQ(size, sizeof(value));
+ EXPECT_EQ(value, 1);
+
+ close(fd);
+}
+
+
+/*
+ * If EFD_SEMAPHORE was not specified and the eventfd counter has a nonzero
+ * value, then a read(2) returns 8 bytes containing that value, and the
+ * counter's value is reset to zero.
+ * If the eventfd counter is zero at the time of the call to read(2), then the
+ * call fails with the error EAGAIN if the file descriptor has been made nonblocking.
+ */
+TEST(eventfd_check_read_with_nonsemaphore)
+{
+ uint64_t value;
+ ssize_t size;
+ int fd;
+ int i;
+
+ fd = sys_eventfd2(0, EFD_NONBLOCK);
+ ASSERT_GE(fd, 0);
+
+ value = 1;
+ for (i = 0; i < EVENTFD_TEST_ITERATIONS; i++) {
+ size = write(fd, &value, sizeof(value));
+ EXPECT_EQ(size, sizeof(value));
+ }
+
+ size = read(fd, &value, sizeof(value));
+ EXPECT_EQ(size, sizeof(uint64_t));
+ EXPECT_EQ(value, EVENTFD_TEST_ITERATIONS);
+
+ size = read(fd, &value, sizeof(value));
+ EXPECT_EQ(size, -1);
+ EXPECT_EQ(errno, EAGAIN);
+
+ close(fd);
+}
+
+/*
+ * If EFD_SEMAPHORE was specified and the eventfd counter has a nonzero value,
+ * then a read(2) returns 8 bytes containing the value 1, and the counter's
+ * value is decremented by 1.
+ * If the eventfd counter is zero at the time of the call to read(2), then the
+ * call fails with the error EAGAIN if the file descriptor has been made nonblocking.
+ */
+TEST(eventfd_check_read_with_semaphore)
+{
+ uint64_t value;
+ ssize_t size;
+ int fd;
+ int i;
+
+ fd = sys_eventfd2(0, EFD_SEMAPHORE|EFD_NONBLOCK);
+ ASSERT_GE(fd, 0);
+
+ value = 1;
+ for (i = 0; i < EVENTFD_TEST_ITERATIONS; i++) {
+ size = write(fd, &value, sizeof(value));
+ EXPECT_EQ(size, sizeof(value));
+ }
+
+ for (i = 0; i < EVENTFD_TEST_ITERATIONS; i++) {
+ size = read(fd, &value, sizeof(value));
+ EXPECT_EQ(size, sizeof(value));
+ EXPECT_EQ(value, 1);
+ }
+
+ size = read(fd, &value, sizeof(value));
+ EXPECT_EQ(size, -1);
+ EXPECT_EQ(errno, EAGAIN);
+
+ close(fd);
+}
+
+TEST_HARNESS_MAIN
diff --git a/tools/testing/selftests/filesystems/file_stressor.c b/tools/testing/selftests/filesystems/file_stressor.c
new file mode 100644
index 000000000000..1136f93a9977
--- /dev/null
+++ b/tools/testing/selftests/filesystems/file_stressor.c
@@ -0,0 +1,194 @@
+// SPDX-License-Identifier: GPL-2.0
+#define _GNU_SOURCE
+#define __SANE_USERSPACE_TYPES__
+
+#include <fcntl.h>
+#include <limits.h>
+#include <pthread.h>
+#include <sched.h>
+#include <stdio.h>
+#include <string.h>
+#include <sys/stat.h>
+#include <sys/mount.h>
+#include <unistd.h>
+
+#include "../kselftest_harness.h"
+
+#include <linux/types.h>
+#include <linux/mount.h>
+#include <sys/syscall.h>
+
+static inline int sys_fsopen(const char *fsname, unsigned int flags)
+{
+ return syscall(__NR_fsopen, fsname, flags);
+}
+
+static inline int sys_fsconfig(int fd, unsigned int cmd, const char *key,
+ const char *value, int aux)
+{
+ return syscall(__NR_fsconfig, fd, cmd, key, value, aux);
+}
+
+static inline int sys_fsmount(int fd, unsigned int flags,
+ unsigned int attr_flags)
+{
+ return syscall(__NR_fsmount, fd, flags, attr_flags);
+}
+
+#ifndef MOVE_MOUNT_F_EMPTY_PATH
+#define MOVE_MOUNT_F_EMPTY_PATH 0x00000004 /* Empty from path permitted */
+#endif
+
+static inline int sys_move_mount(int from_dfd, const char *from_pathname,
+ int to_dfd, const char *to_pathname,
+ unsigned int flags)
+{
+ return syscall(__NR_move_mount, from_dfd, from_pathname, to_dfd,
+ to_pathname, flags);
+}
+
+FIXTURE(file_stressor) {
+ int fd_tmpfs;
+ int nr_procs;
+ int max_fds;
+ pid_t *pids_openers;
+ pid_t *pids_getdents;
+ int *fd_proc_pid;
+};
+
+FIXTURE_SETUP(file_stressor)
+{
+ int fd_context;
+
+ ASSERT_EQ(unshare(CLONE_NEWNS), 0);
+ ASSERT_EQ(mount(NULL, "/", NULL, MS_SLAVE | MS_REC, NULL), 0);
+ ASSERT_EQ(mkdir("/slab_typesafe_by_rcu", 0755), 0);
+
+ fd_context = sys_fsopen("tmpfs", 0);
+ ASSERT_GE(fd_context, 0);
+
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_CMD_CREATE, NULL, NULL, 0), 0);
+ self->fd_tmpfs = sys_fsmount(fd_context, 0, 0);
+ ASSERT_GE(self->fd_tmpfs, 0);
+ ASSERT_EQ(close(fd_context), 0);
+
+ ASSERT_EQ(sys_move_mount(self->fd_tmpfs, "", -EBADF, "/slab_typesafe_by_rcu", MOVE_MOUNT_F_EMPTY_PATH), 0);
+
+ self->nr_procs = sysconf(_SC_NPROCESSORS_ONLN);
+ self->pids_openers = malloc(sizeof(pid_t) * self->nr_procs);
+ ASSERT_NE(self->pids_openers, NULL);
+ self->pids_getdents = malloc(sizeof(pid_t) * self->nr_procs);
+ ASSERT_NE(self->pids_getdents, NULL);
+ self->fd_proc_pid = malloc(sizeof(int) * self->nr_procs);
+ ASSERT_NE(self->fd_proc_pid, NULL);
+ self->max_fds = 500;
+}
+
+FIXTURE_TEARDOWN(file_stressor)
+{
+ for (int i = 0; i < self->nr_procs; i++) {
+ int wstatus;
+ pid_t pid;
+
+ pid = waitpid(self->pids_openers[i], &wstatus, 0);
+ ASSERT_EQ(pid, self->pids_openers[i]);
+ ASSERT_TRUE(!WIFEXITED(wstatus) || !WIFSIGNALED(wstatus));
+
+ pid = waitpid(self->pids_getdents[i], &wstatus, 0);
+ ASSERT_EQ(pid, self->pids_getdents[i]);
+ ASSERT_TRUE(!WIFEXITED(wstatus) || !WIFSIGNALED(wstatus));
+ }
+ free(self->pids_openers);
+ free(self->pids_getdents);
+ ASSERT_EQ(close(self->fd_tmpfs), 0);
+
+ umount2("/slab_typesafe_by_rcu", 0);
+ ASSERT_EQ(rmdir("/slab_typesafe_by_rcu"), 0);
+}
+
+TEST_F_TIMEOUT(file_stressor, slab_typesafe_by_rcu, 900 * 2)
+{
+ for (int i = 0; i < self->nr_procs; i++) {
+ pid_t pid_self;
+
+ self->pids_openers[i] = fork();
+ ASSERT_GE(self->pids_openers[i], 0);
+
+ if (self->pids_openers[i] != 0)
+ continue;
+
+ self->pids_openers[i] = getpid();
+ for (;;) {
+ for (int i = 0; i < self->max_fds; i++) {
+ char path[PATH_MAX];
+ int fd;
+
+ sprintf(path, "/slab_typesafe_by_rcu/file-%d-%d", self->pids_openers[i], i);
+ fd = open(path, O_CREAT | O_RDONLY | O_CLOEXEC, 0644);
+ if (fd < 0)
+ continue;
+ }
+
+ close_range(3, ~0U, 0);
+ }
+
+ exit(0);
+ }
+
+ for (int i = 0; i < self->nr_procs; i++) {
+ char path[PATH_MAX];
+
+ sprintf(path, "/proc/%d/fd/", self->pids_openers[i]);
+ self->fd_proc_pid[i] = open(path, O_DIRECTORY | O_RDONLY | O_CLOEXEC);
+ ASSERT_GE(self->fd_proc_pid[i], 0);
+ }
+
+ for (int i = 0; i < self->nr_procs; i++) {
+ self->pids_getdents[i] = fork();
+ ASSERT_GE(self->pids_getdents[i], 0);
+
+ if (self->pids_getdents[i] != 0)
+ continue;
+
+ self->pids_getdents[i] = getpid();
+ for (;;) {
+ char ents[1024];
+ ssize_t nr_read;
+
+ /*
+ * Concurrently read /proc/<pid>/fd/ which rougly does:
+ *
+ * f = fget_task_next(p, &fd);
+ * if (!f)
+ * break;
+ * data.mode = f->f_mode;
+ * fput(f);
+ *
+ * Which means that it'll try to get a reference to a
+ * file in another task's file descriptor table.
+ *
+ * Under heavy file load it is increasingly likely that
+ * the other task will manage to close @file and @file
+ * is being recycled due to SLAB_TYPEAFE_BY_RCU
+ * concurrently. This will trigger various warnings in
+ * the file reference counting code.
+ */
+ do {
+ nr_read = syscall(SYS_getdents64, self->fd_proc_pid[i], ents, sizeof(ents));
+ } while (nr_read >= 0);
+
+ lseek(self->fd_proc_pid[i], 0, SEEK_SET);
+ }
+
+ exit(0);
+ }
+
+ ASSERT_EQ(clock_nanosleep(CLOCK_MONOTONIC, 0, &(struct timespec){ .tv_sec = 900 /* 15 min */ }, NULL), 0);
+
+ for (int i = 0; i < self->nr_procs; i++) {
+ kill(self->pids_openers[i], SIGKILL);
+ kill(self->pids_getdents[i], SIGKILL);
+ }
+}
+
+TEST_HARNESS_MAIN
diff --git a/tools/testing/selftests/filesystems/nsfs/.gitignore b/tools/testing/selftests/filesystems/nsfs/.gitignore
new file mode 100644
index 000000000000..92a8249006d1
--- /dev/null
+++ b/tools/testing/selftests/filesystems/nsfs/.gitignore
@@ -0,0 +1,4 @@
+# SPDX-License-Identifier: GPL-2.0-only
+owner
+pidns
+iterate_mntns
diff --git a/tools/testing/selftests/filesystems/nsfs/Makefile b/tools/testing/selftests/filesystems/nsfs/Makefile
new file mode 100644
index 000000000000..231aaa7dfd95
--- /dev/null
+++ b/tools/testing/selftests/filesystems/nsfs/Makefile
@@ -0,0 +1,6 @@
+# SPDX-License-Identifier: GPL-2.0-only
+TEST_GEN_PROGS := owner pidns iterate_mntns
+
+CFLAGS := -Wall -Werror
+
+include ../../lib.mk
diff --git a/tools/testing/selftests/filesystems/nsfs/config b/tools/testing/selftests/filesystems/nsfs/config
new file mode 100644
index 000000000000..598d0a225fc9
--- /dev/null
+++ b/tools/testing/selftests/filesystems/nsfs/config
@@ -0,0 +1,3 @@
+CONFIG_USER_NS=y
+CONFIG_UTS_NS=y
+CONFIG_PID_NS=y
diff --git a/tools/testing/selftests/filesystems/nsfs/iterate_mntns.c b/tools/testing/selftests/filesystems/nsfs/iterate_mntns.c
new file mode 100644
index 000000000000..457cf76f3c5f
--- /dev/null
+++ b/tools/testing/selftests/filesystems/nsfs/iterate_mntns.c
@@ -0,0 +1,149 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+// Copyright (c) 2024 Christian Brauner <brauner@kernel.org>
+
+#define _GNU_SOURCE
+#include <fcntl.h>
+#include <sched.h>
+#include <stdio.h>
+#include <string.h>
+#include <sys/stat.h>
+#include <sys/mount.h>
+#include <unistd.h>
+
+#include "../../kselftest_harness.h"
+
+#define MNT_NS_COUNT 11
+#define MNT_NS_LAST_INDEX 10
+
+struct mnt_ns_info {
+ __u32 size;
+ __u32 nr_mounts;
+ __u64 mnt_ns_id;
+};
+
+#define MNT_NS_INFO_SIZE_VER0 16 /* size of first published struct */
+
+/* Get information about namespace. */
+#define NS_MNT_GET_INFO _IOR(0xb7, 10, struct mnt_ns_info)
+/* Get next namespace. */
+#define NS_MNT_GET_NEXT _IOR(0xb7, 11, struct mnt_ns_info)
+/* Get previous namespace. */
+#define NS_MNT_GET_PREV _IOR(0xb7, 12, struct mnt_ns_info)
+
+FIXTURE(iterate_mount_namespaces) {
+ int fd_mnt_ns[MNT_NS_COUNT];
+ __u64 mnt_ns_id[MNT_NS_COUNT];
+};
+
+FIXTURE_SETUP(iterate_mount_namespaces)
+{
+ for (int i = 0; i < MNT_NS_COUNT; i++)
+ self->fd_mnt_ns[i] = -EBADF;
+
+ /*
+ * Creating a new user namespace let's us guarantee that we only see
+ * mount namespaces that we did actually create.
+ */
+ ASSERT_EQ(unshare(CLONE_NEWUSER), 0);
+
+ for (int i = 0; i < MNT_NS_COUNT; i++) {
+ struct mnt_ns_info info = {};
+
+ ASSERT_EQ(unshare(CLONE_NEWNS), 0);
+ self->fd_mnt_ns[i] = open("/proc/self/ns/mnt", O_RDONLY | O_CLOEXEC);
+ ASSERT_GE(self->fd_mnt_ns[i], 0);
+ ASSERT_EQ(ioctl(self->fd_mnt_ns[i], NS_MNT_GET_INFO, &info), 0);
+ self->mnt_ns_id[i] = info.mnt_ns_id;
+ }
+}
+
+FIXTURE_TEARDOWN(iterate_mount_namespaces)
+{
+ for (int i = 0; i < MNT_NS_COUNT; i++) {
+ if (self->fd_mnt_ns[i] < 0)
+ continue;
+ ASSERT_EQ(close(self->fd_mnt_ns[i]), 0);
+ }
+}
+
+TEST_F(iterate_mount_namespaces, iterate_all_forward)
+{
+ int fd_mnt_ns_cur, count = 0;
+
+ fd_mnt_ns_cur = fcntl(self->fd_mnt_ns[0], F_DUPFD_CLOEXEC);
+ ASSERT_GE(fd_mnt_ns_cur, 0);
+
+ for (;; count++) {
+ struct mnt_ns_info info = {};
+ int fd_mnt_ns_next;
+
+ fd_mnt_ns_next = ioctl(fd_mnt_ns_cur, NS_MNT_GET_NEXT, &info);
+ if (fd_mnt_ns_next < 0 && errno == ENOENT)
+ break;
+ ASSERT_GE(fd_mnt_ns_next, 0);
+ ASSERT_EQ(close(fd_mnt_ns_cur), 0);
+ fd_mnt_ns_cur = fd_mnt_ns_next;
+ }
+ ASSERT_EQ(count, MNT_NS_LAST_INDEX);
+}
+
+TEST_F(iterate_mount_namespaces, iterate_all_backwards)
+{
+ int fd_mnt_ns_cur, count = 0;
+
+ fd_mnt_ns_cur = fcntl(self->fd_mnt_ns[MNT_NS_LAST_INDEX], F_DUPFD_CLOEXEC);
+ ASSERT_GE(fd_mnt_ns_cur, 0);
+
+ for (;; count++) {
+ struct mnt_ns_info info = {};
+ int fd_mnt_ns_prev;
+
+ fd_mnt_ns_prev = ioctl(fd_mnt_ns_cur, NS_MNT_GET_PREV, &info);
+ if (fd_mnt_ns_prev < 0 && errno == ENOENT)
+ break;
+ ASSERT_GE(fd_mnt_ns_prev, 0);
+ ASSERT_EQ(close(fd_mnt_ns_cur), 0);
+ fd_mnt_ns_cur = fd_mnt_ns_prev;
+ }
+ ASSERT_EQ(count, MNT_NS_LAST_INDEX);
+}
+
+TEST_F(iterate_mount_namespaces, iterate_forward)
+{
+ int fd_mnt_ns_cur;
+
+ ASSERT_EQ(setns(self->fd_mnt_ns[0], CLONE_NEWNS), 0);
+
+ fd_mnt_ns_cur = self->fd_mnt_ns[0];
+ for (int i = 1; i < MNT_NS_COUNT; i++) {
+ struct mnt_ns_info info = {};
+ int fd_mnt_ns_next;
+
+ fd_mnt_ns_next = ioctl(fd_mnt_ns_cur, NS_MNT_GET_NEXT, &info);
+ ASSERT_GE(fd_mnt_ns_next, 0);
+ ASSERT_EQ(close(fd_mnt_ns_cur), 0);
+ fd_mnt_ns_cur = fd_mnt_ns_next;
+ ASSERT_EQ(info.mnt_ns_id, self->mnt_ns_id[i]);
+ }
+}
+
+TEST_F(iterate_mount_namespaces, iterate_backward)
+{
+ int fd_mnt_ns_cur;
+
+ ASSERT_EQ(setns(self->fd_mnt_ns[MNT_NS_LAST_INDEX], CLONE_NEWNS), 0);
+
+ fd_mnt_ns_cur = self->fd_mnt_ns[MNT_NS_LAST_INDEX];
+ for (int i = MNT_NS_LAST_INDEX - 1; i >= 0; i--) {
+ struct mnt_ns_info info = {};
+ int fd_mnt_ns_prev;
+
+ fd_mnt_ns_prev = ioctl(fd_mnt_ns_cur, NS_MNT_GET_PREV, &info);
+ ASSERT_GE(fd_mnt_ns_prev, 0);
+ ASSERT_EQ(close(fd_mnt_ns_cur), 0);
+ fd_mnt_ns_cur = fd_mnt_ns_prev;
+ ASSERT_EQ(info.mnt_ns_id, self->mnt_ns_id[i]);
+ }
+}
+
+TEST_HARNESS_MAIN
diff --git a/tools/testing/selftests/filesystems/nsfs/owner.c b/tools/testing/selftests/filesystems/nsfs/owner.c
new file mode 100644
index 000000000000..96a976c74550
--- /dev/null
+++ b/tools/testing/selftests/filesystems/nsfs/owner.c
@@ -0,0 +1,92 @@
+// SPDX-License-Identifier: GPL-2.0
+#define _GNU_SOURCE
+#include <sched.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <errno.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+#include <sys/ioctl.h>
+#include <sys/prctl.h>
+#include <sys/wait.h>
+
+#define NSIO 0xb7
+#define NS_GET_USERNS _IO(NSIO, 0x1)
+
+#define pr_err(fmt, ...) \
+ ({ \
+ fprintf(stderr, "%s:%d:" fmt ": %m\n", \
+ __func__, __LINE__, ##__VA_ARGS__); \
+ 1; \
+ })
+
+int main(int argc, char *argvp[])
+{
+ int pfd[2], ns, uns, init_uns;
+ struct stat st1, st2;
+ char path[128];
+ pid_t pid;
+ char c;
+
+ if (pipe(pfd))
+ return 1;
+
+ pid = fork();
+ if (pid < 0)
+ return pr_err("fork");
+ if (pid == 0) {
+ prctl(PR_SET_PDEATHSIG, SIGKILL);
+ if (unshare(CLONE_NEWUTS | CLONE_NEWUSER))
+ return pr_err("unshare");
+ close(pfd[0]);
+ close(pfd[1]);
+ while (1)
+ sleep(1);
+ return 0;
+ }
+ close(pfd[1]);
+ if (read(pfd[0], &c, 1) != 0)
+ return pr_err("Unable to read from pipe");
+ close(pfd[0]);
+
+ snprintf(path, sizeof(path), "/proc/%d/ns/uts", pid);
+ ns = open(path, O_RDONLY);
+ if (ns < 0)
+ return pr_err("Unable to open %s", path);
+
+ uns = ioctl(ns, NS_GET_USERNS);
+ if (uns < 0)
+ return pr_err("Unable to get an owning user namespace");
+
+ if (fstat(uns, &st1))
+ return pr_err("fstat");
+
+ snprintf(path, sizeof(path), "/proc/%d/ns/user", pid);
+ if (stat(path, &st2))
+ return pr_err("stat");
+
+ if (st1.st_ino != st2.st_ino)
+ return pr_err("NS_GET_USERNS returned a wrong namespace");
+
+ init_uns = ioctl(uns, NS_GET_USERNS);
+ if (uns < 0)
+ return pr_err("Unable to get an owning user namespace");
+
+ if (ioctl(init_uns, NS_GET_USERNS) >= 0 || errno != EPERM)
+ return pr_err("Don't get EPERM");
+
+ if (unshare(CLONE_NEWUSER))
+ return pr_err("unshare");
+
+ if (ioctl(ns, NS_GET_USERNS) >= 0 || errno != EPERM)
+ return pr_err("Don't get EPERM");
+ if (ioctl(init_uns, NS_GET_USERNS) >= 0 || errno != EPERM)
+ return pr_err("Don't get EPERM");
+
+ kill(pid, SIGKILL);
+ wait(NULL);
+ return 0;
+}
diff --git a/tools/testing/selftests/filesystems/nsfs/pidns.c b/tools/testing/selftests/filesystems/nsfs/pidns.c
new file mode 100644
index 000000000000..e3c772c6a7c7
--- /dev/null
+++ b/tools/testing/selftests/filesystems/nsfs/pidns.c
@@ -0,0 +1,79 @@
+// SPDX-License-Identifier: GPL-2.0
+#define _GNU_SOURCE
+#include <sched.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <errno.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+#include <sys/ioctl.h>
+#include <sys/prctl.h>
+#include <sys/wait.h>
+
+#define pr_err(fmt, ...) \
+ ({ \
+ fprintf(stderr, "%s:%d:" fmt ": %m\n", \
+ __func__, __LINE__, ##__VA_ARGS__); \
+ 1; \
+ })
+
+#define NSIO 0xb7
+#define NS_GET_USERNS _IO(NSIO, 0x1)
+#define NS_GET_PARENT _IO(NSIO, 0x2)
+
+#define __stack_aligned__ __attribute__((aligned(16)))
+struct cr_clone_arg {
+ char stack[128] __stack_aligned__;
+ char stack_ptr[];
+};
+
+static int child(void *args)
+{
+ prctl(PR_SET_PDEATHSIG, SIGKILL);
+ while (1)
+ sleep(1);
+ exit(0);
+}
+
+int main(int argc, char *argv[])
+{
+ char *ns_strs[] = {"pid", "user"};
+ char path[] = "/proc/0123456789/ns/pid";
+ struct cr_clone_arg ca;
+ struct stat st1, st2;
+ int ns, pns, i;
+ pid_t pid;
+
+ pid = clone(child, ca.stack_ptr, CLONE_NEWUSER | CLONE_NEWPID | SIGCHLD, NULL);
+ if (pid < 0)
+ return pr_err("clone");
+
+ for (i = 0; i < 2; i++) {
+ snprintf(path, sizeof(path), "/proc/%d/ns/%s", pid, ns_strs[i]);
+ ns = open(path, O_RDONLY);
+ if (ns < 0)
+ return pr_err("Unable to open %s", path);
+
+ pns = ioctl(ns, NS_GET_PARENT);
+ if (pns < 0)
+ return pr_err("Unable to get a parent pidns");
+
+ snprintf(path, sizeof(path), "/proc/self/ns/%s", ns_strs[i]);
+ if (stat(path, &st2))
+ return pr_err("Unable to stat %s", path);
+ if (fstat(pns, &st1))
+ return pr_err("Unable to stat the parent pidns");
+ if (st1.st_ino != st2.st_ino)
+ return pr_err("NS_GET_PARENT returned a wrong namespace");
+
+ if (ioctl(pns, NS_GET_PARENT) >= 0 || errno != EPERM)
+ return pr_err("Don't get EPERM");
+ }
+
+ kill(pid, SIGKILL);
+ wait(NULL);
+ return 0;
+}
diff --git a/tools/testing/selftests/filesystems/overlayfs/.gitignore b/tools/testing/selftests/filesystems/overlayfs/.gitignore
index 52ae618fdd98..e23a18c8b37f 100644
--- a/tools/testing/selftests/filesystems/overlayfs/.gitignore
+++ b/tools/testing/selftests/filesystems/overlayfs/.gitignore
@@ -1,2 +1,3 @@
# SPDX-License-Identifier: GPL-2.0-only
dev_in_maps
+set_layers_via_fds
diff --git a/tools/testing/selftests/filesystems/overlayfs/Makefile b/tools/testing/selftests/filesystems/overlayfs/Makefile
index 56b2b48a765b..e8d1adb021af 100644
--- a/tools/testing/selftests/filesystems/overlayfs/Makefile
+++ b/tools/testing/selftests/filesystems/overlayfs/Makefile
@@ -1,6 +1,6 @@
# SPDX-License-Identifier: GPL-2.0
-TEST_GEN_PROGS := dev_in_maps
+TEST_GEN_PROGS := dev_in_maps set_layers_via_fds
CFLAGS := -Wall -Werror
diff --git a/tools/testing/selftests/filesystems/overlayfs/dev_in_maps.c b/tools/testing/selftests/filesystems/overlayfs/dev_in_maps.c
index e19ab0e85709..3b796264223f 100644
--- a/tools/testing/selftests/filesystems/overlayfs/dev_in_maps.c
+++ b/tools/testing/selftests/filesystems/overlayfs/dev_in_maps.c
@@ -1,5 +1,6 @@
// SPDX-License-Identifier: GPL-2.0
#define _GNU_SOURCE
+#define __SANE_USERSPACE_TYPES__ // Use ll64
#include <inttypes.h>
#include <unistd.h>
@@ -10,35 +11,13 @@
#include <linux/mount.h>
#include <sys/syscall.h>
#include <sys/stat.h>
-#include <sys/mount.h>
#include <sys/mman.h>
#include <sched.h>
#include <fcntl.h>
#include "../../kselftest.h"
#include "log.h"
-
-static int sys_fsopen(const char *fsname, unsigned int flags)
-{
- return syscall(__NR_fsopen, fsname, flags);
-}
-
-static int sys_fsconfig(int fd, unsigned int cmd, const char *key, const char *value, int aux)
-{
- return syscall(__NR_fsconfig, fd, cmd, key, value, aux);
-}
-
-static int sys_fsmount(int fd, unsigned int flags, unsigned int attr_flags)
-{
- return syscall(__NR_fsmount, fd, flags, attr_flags);
-}
-
-static int sys_move_mount(int from_dfd, const char *from_pathname,
- int to_dfd, const char *to_pathname,
- unsigned int flags)
-{
- return syscall(__NR_move_mount, from_dfd, from_pathname, to_dfd, to_pathname, flags);
-}
+#include "wrappers.h"
static long get_file_dev_and_inode(void *addr, struct statx *stx)
{
@@ -166,8 +145,7 @@ int main(int argc, char **argv)
ksft_test_result_skip("unable to create a new mount namespace\n");
return 1;
}
-
- if (mount(NULL, "/", NULL, MS_SLAVE | MS_REC, NULL) == -1) {
+ if (sys_mount(NULL, "/", NULL, MS_SLAVE | MS_REC, NULL) == -1) {
pr_perror("mount");
return 1;
}
diff --git a/tools/testing/selftests/filesystems/overlayfs/set_layers_via_fds.c b/tools/testing/selftests/filesystems/overlayfs/set_layers_via_fds.c
new file mode 100644
index 000000000000..1d0ae785a667
--- /dev/null
+++ b/tools/testing/selftests/filesystems/overlayfs/set_layers_via_fds.c
@@ -0,0 +1,217 @@
+// SPDX-License-Identifier: GPL-2.0
+#define _GNU_SOURCE
+#define __SANE_USERSPACE_TYPES__ // Use ll64
+
+#include <fcntl.h>
+#include <sched.h>
+#include <stdio.h>
+#include <string.h>
+#include <sys/stat.h>
+#include <sys/mount.h>
+#include <unistd.h>
+
+#include "../../kselftest_harness.h"
+#include "log.h"
+#include "wrappers.h"
+
+FIXTURE(set_layers_via_fds) {
+};
+
+FIXTURE_SETUP(set_layers_via_fds)
+{
+ ASSERT_EQ(mkdir("/set_layers_via_fds", 0755), 0);
+}
+
+FIXTURE_TEARDOWN(set_layers_via_fds)
+{
+ umount2("/set_layers_via_fds", 0);
+ ASSERT_EQ(rmdir("/set_layers_via_fds"), 0);
+}
+
+TEST_F(set_layers_via_fds, set_layers_via_fds)
+{
+ int fd_context, fd_tmpfs, fd_overlay;
+ int layer_fds[] = { [0 ... 8] = -EBADF };
+ bool layers_found[] = { [0 ... 8] = false };
+ size_t len = 0;
+ char *line = NULL;
+ FILE *f_mountinfo;
+
+ ASSERT_EQ(unshare(CLONE_NEWNS), 0);
+ ASSERT_EQ(sys_mount(NULL, "/", NULL, MS_SLAVE | MS_REC, NULL), 0);
+
+ fd_context = sys_fsopen("tmpfs", 0);
+ ASSERT_GE(fd_context, 0);
+
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_CMD_CREATE, NULL, NULL, 0), 0);
+ fd_tmpfs = sys_fsmount(fd_context, 0, 0);
+ ASSERT_GE(fd_tmpfs, 0);
+ ASSERT_EQ(close(fd_context), 0);
+
+ ASSERT_EQ(mkdirat(fd_tmpfs, "w", 0755), 0);
+ ASSERT_EQ(mkdirat(fd_tmpfs, "u", 0755), 0);
+ ASSERT_EQ(mkdirat(fd_tmpfs, "l1", 0755), 0);
+ ASSERT_EQ(mkdirat(fd_tmpfs, "l2", 0755), 0);
+ ASSERT_EQ(mkdirat(fd_tmpfs, "l3", 0755), 0);
+ ASSERT_EQ(mkdirat(fd_tmpfs, "l4", 0755), 0);
+ ASSERT_EQ(mkdirat(fd_tmpfs, "d1", 0755), 0);
+ ASSERT_EQ(mkdirat(fd_tmpfs, "d2", 0755), 0);
+ ASSERT_EQ(mkdirat(fd_tmpfs, "d3", 0755), 0);
+
+ layer_fds[0] = openat(fd_tmpfs, "w", O_DIRECTORY);
+ ASSERT_GE(layer_fds[0], 0);
+
+ layer_fds[1] = openat(fd_tmpfs, "u", O_DIRECTORY);
+ ASSERT_GE(layer_fds[1], 0);
+
+ layer_fds[2] = openat(fd_tmpfs, "l1", O_DIRECTORY);
+ ASSERT_GE(layer_fds[2], 0);
+
+ layer_fds[3] = openat(fd_tmpfs, "l2", O_DIRECTORY);
+ ASSERT_GE(layer_fds[3], 0);
+
+ layer_fds[4] = openat(fd_tmpfs, "l3", O_DIRECTORY);
+ ASSERT_GE(layer_fds[4], 0);
+
+ layer_fds[5] = openat(fd_tmpfs, "l4", O_DIRECTORY);
+ ASSERT_GE(layer_fds[5], 0);
+
+ layer_fds[6] = openat(fd_tmpfs, "d1", O_DIRECTORY);
+ ASSERT_GE(layer_fds[6], 0);
+
+ layer_fds[7] = openat(fd_tmpfs, "d2", O_DIRECTORY);
+ ASSERT_GE(layer_fds[7], 0);
+
+ layer_fds[8] = openat(fd_tmpfs, "d3", O_DIRECTORY);
+ ASSERT_GE(layer_fds[8], 0);
+
+ ASSERT_EQ(sys_move_mount(fd_tmpfs, "", -EBADF, "/tmp", MOVE_MOUNT_F_EMPTY_PATH), 0);
+ ASSERT_EQ(close(fd_tmpfs), 0);
+
+ fd_context = sys_fsopen("overlay", 0);
+ ASSERT_GE(fd_context, 0);
+
+ ASSERT_NE(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "lowerdir", NULL, layer_fds[2]), 0);
+
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "workdir", NULL, layer_fds[0]), 0);
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "upperdir", NULL, layer_fds[1]), 0);
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "lowerdir+", NULL, layer_fds[2]), 0);
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "lowerdir+", NULL, layer_fds[3]), 0);
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "lowerdir+", NULL, layer_fds[4]), 0);
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "lowerdir+", NULL, layer_fds[5]), 0);
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "datadir+", NULL, layer_fds[6]), 0);
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "datadir+", NULL, layer_fds[7]), 0);
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "datadir+", NULL, layer_fds[8]), 0);
+
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_SET_STRING, "metacopy", "on", 0), 0);
+
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_CMD_CREATE, NULL, NULL, 0), 0);
+
+ fd_overlay = sys_fsmount(fd_context, 0, 0);
+ ASSERT_GE(fd_overlay, 0);
+
+ ASSERT_EQ(sys_move_mount(fd_overlay, "", -EBADF, "/set_layers_via_fds", MOVE_MOUNT_F_EMPTY_PATH), 0);
+
+ f_mountinfo = fopen("/proc/self/mountinfo", "r");
+ ASSERT_NE(f_mountinfo, NULL);
+
+ while (getline(&line, &len, f_mountinfo) != -1) {
+ char *haystack = line;
+
+ if (strstr(haystack, "workdir=/tmp/w"))
+ layers_found[0] = true;
+ if (strstr(haystack, "upperdir=/tmp/u"))
+ layers_found[1] = true;
+ if (strstr(haystack, "lowerdir+=/tmp/l1"))
+ layers_found[2] = true;
+ if (strstr(haystack, "lowerdir+=/tmp/l2"))
+ layers_found[3] = true;
+ if (strstr(haystack, "lowerdir+=/tmp/l3"))
+ layers_found[4] = true;
+ if (strstr(haystack, "lowerdir+=/tmp/l4"))
+ layers_found[5] = true;
+ if (strstr(haystack, "datadir+=/tmp/d1"))
+ layers_found[6] = true;
+ if (strstr(haystack, "datadir+=/tmp/d2"))
+ layers_found[7] = true;
+ if (strstr(haystack, "datadir+=/tmp/d3"))
+ layers_found[8] = true;
+ }
+ free(line);
+
+ for (int i = 0; i < ARRAY_SIZE(layer_fds); i++) {
+ ASSERT_EQ(layers_found[i], true);
+ ASSERT_EQ(close(layer_fds[i]), 0);
+ }
+
+ ASSERT_EQ(close(fd_context), 0);
+ ASSERT_EQ(close(fd_overlay), 0);
+ ASSERT_EQ(fclose(f_mountinfo), 0);
+}
+
+TEST_F(set_layers_via_fds, set_500_layers_via_fds)
+{
+ int fd_context, fd_tmpfs, fd_overlay, fd_work, fd_upper, fd_lower;
+ int layer_fds[500] = { [0 ... 499] = -EBADF };
+
+ ASSERT_EQ(unshare(CLONE_NEWNS), 0);
+ ASSERT_EQ(sys_mount(NULL, "/", NULL, MS_SLAVE | MS_REC, NULL), 0);
+
+ fd_context = sys_fsopen("tmpfs", 0);
+ ASSERT_GE(fd_context, 0);
+
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_CMD_CREATE, NULL, NULL, 0), 0);
+ fd_tmpfs = sys_fsmount(fd_context, 0, 0);
+ ASSERT_GE(fd_tmpfs, 0);
+ ASSERT_EQ(close(fd_context), 0);
+
+ for (int i = 0; i < ARRAY_SIZE(layer_fds); i++) {
+ char path[100];
+
+ sprintf(path, "l%d", i);
+ ASSERT_EQ(mkdirat(fd_tmpfs, path, 0755), 0);
+ layer_fds[i] = openat(fd_tmpfs, path, O_DIRECTORY);
+ ASSERT_GE(layer_fds[i], 0);
+ }
+
+ ASSERT_EQ(mkdirat(fd_tmpfs, "w", 0755), 0);
+ fd_work = openat(fd_tmpfs, "w", O_DIRECTORY);
+ ASSERT_GE(fd_work, 0);
+
+ ASSERT_EQ(mkdirat(fd_tmpfs, "u", 0755), 0);
+ fd_upper = openat(fd_tmpfs, "u", O_DIRECTORY);
+ ASSERT_GE(fd_upper, 0);
+
+ ASSERT_EQ(mkdirat(fd_tmpfs, "l501", 0755), 0);
+ fd_lower = openat(fd_tmpfs, "l501", O_DIRECTORY);
+ ASSERT_GE(fd_lower, 0);
+
+ ASSERT_EQ(sys_move_mount(fd_tmpfs, "", -EBADF, "/tmp", MOVE_MOUNT_F_EMPTY_PATH), 0);
+ ASSERT_EQ(close(fd_tmpfs), 0);
+
+ fd_context = sys_fsopen("overlay", 0);
+ ASSERT_GE(fd_context, 0);
+
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "workdir", NULL, fd_work), 0);
+ ASSERT_EQ(close(fd_work), 0);
+
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "upperdir", NULL, fd_upper), 0);
+ ASSERT_EQ(close(fd_upper), 0);
+
+ for (int i = 0; i < ARRAY_SIZE(layer_fds); i++) {
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "lowerdir+", NULL, layer_fds[i]), 0);
+ ASSERT_EQ(close(layer_fds[i]), 0);
+ }
+
+ ASSERT_NE(sys_fsconfig(fd_context, FSCONFIG_SET_FD, "lowerdir+", NULL, fd_lower), 0);
+ ASSERT_EQ(close(fd_lower), 0);
+
+ ASSERT_EQ(sys_fsconfig(fd_context, FSCONFIG_CMD_CREATE, NULL, NULL, 0), 0);
+
+ fd_overlay = sys_fsmount(fd_context, 0, 0);
+ ASSERT_GE(fd_overlay, 0);
+ ASSERT_EQ(close(fd_context), 0);
+ ASSERT_EQ(close(fd_overlay), 0);
+}
+
+TEST_HARNESS_MAIN
diff --git a/tools/testing/selftests/filesystems/overlayfs/wrappers.h b/tools/testing/selftests/filesystems/overlayfs/wrappers.h
new file mode 100644
index 000000000000..071b95fd2ac0
--- /dev/null
+++ b/tools/testing/selftests/filesystems/overlayfs/wrappers.h
@@ -0,0 +1,47 @@
+// SPDX-License-Identifier: GPL-2.0
+//
+#ifndef __SELFTEST_OVERLAYFS_WRAPPERS_H__
+#define __SELFTEST_OVERLAYFS_WRAPPERS_H__
+
+#define _GNU_SOURCE
+
+#include <linux/types.h>
+#include <linux/mount.h>
+#include <sys/syscall.h>
+
+static inline int sys_fsopen(const char *fsname, unsigned int flags)
+{
+ return syscall(__NR_fsopen, fsname, flags);
+}
+
+static inline int sys_fsconfig(int fd, unsigned int cmd, const char *key,
+ const char *value, int aux)
+{
+ return syscall(__NR_fsconfig, fd, cmd, key, value, aux);
+}
+
+static inline int sys_fsmount(int fd, unsigned int flags,
+ unsigned int attr_flags)
+{
+ return syscall(__NR_fsmount, fd, flags, attr_flags);
+}
+
+static inline int sys_mount(const char *src, const char *tgt, const char *fst,
+ unsigned long flags, const void *data)
+{
+ return syscall(__NR_mount, src, tgt, fst, flags, data);
+}
+
+#ifndef MOVE_MOUNT_F_EMPTY_PATH
+#define MOVE_MOUNT_F_EMPTY_PATH 0x00000004 /* Empty from path permitted */
+#endif
+
+static inline int sys_move_mount(int from_dfd, const char *from_pathname,
+ int to_dfd, const char *to_pathname,
+ unsigned int flags)
+{
+ return syscall(__NR_move_mount, from_dfd, from_pathname, to_dfd,
+ to_pathname, flags);
+}
+
+#endif
diff --git a/tools/testing/selftests/filesystems/statmount/.gitignore b/tools/testing/selftests/filesystems/statmount/.gitignore
index 82a4846cbc4b..973363ad66a2 100644
--- a/tools/testing/selftests/filesystems/statmount/.gitignore
+++ b/tools/testing/selftests/filesystems/statmount/.gitignore
@@ -1,2 +1,3 @@
# SPDX-License-Identifier: GPL-2.0-only
+statmount_test_ns
/*_test
diff --git a/tools/testing/selftests/filesystems/statmount/Makefile b/tools/testing/selftests/filesystems/statmount/Makefile
index 07a0d5b545ca..14ee91a41650 100644
--- a/tools/testing/selftests/filesystems/statmount/Makefile
+++ b/tools/testing/selftests/filesystems/statmount/Makefile
@@ -1,6 +1,6 @@
# SPDX-License-Identifier: GPL-2.0-or-later
CFLAGS += -Wall -O2 -g $(KHDR_INCLUDES)
-TEST_GEN_PROGS := statmount_test
+TEST_GEN_PROGS := statmount_test statmount_test_ns listmount_test
include ../../lib.mk
diff --git a/tools/testing/selftests/filesystems/statmount/listmount_test.c b/tools/testing/selftests/filesystems/statmount/listmount_test.c
new file mode 100644
index 000000000000..15f0834f7557
--- /dev/null
+++ b/tools/testing/selftests/filesystems/statmount/listmount_test.c
@@ -0,0 +1,66 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+// Copyright (c) 2024 Christian Brauner <brauner@kernel.org>
+
+#define _GNU_SOURCE
+#include <fcntl.h>
+#include <sched.h>
+#include <stdio.h>
+#include <string.h>
+#include <sys/stat.h>
+#include <sys/mount.h>
+#include <unistd.h>
+
+#include "statmount.h"
+#include "../../kselftest_harness.h"
+
+#ifndef LISTMOUNT_REVERSE
+#define LISTMOUNT_REVERSE (1 << 0) /* List later mounts first */
+#endif
+
+#define LISTMNT_BUFFER 10
+
+/* Check that all mount ids are in increasing order. */
+TEST(listmount_forward)
+{
+ uint64_t list[LISTMNT_BUFFER], last_mnt_id = 0;
+
+ for (;;) {
+ ssize_t nr_mounts;
+
+ nr_mounts = listmount(LSMT_ROOT, 0, last_mnt_id,
+ list, LISTMNT_BUFFER, 0);
+ ASSERT_GE(nr_mounts, 0);
+ if (nr_mounts == 0)
+ break;
+
+ for (size_t cur = 0; cur < nr_mounts; cur++) {
+ if (cur < nr_mounts - 1)
+ ASSERT_LT(list[cur], list[cur + 1]);
+ last_mnt_id = list[cur];
+ }
+ }
+}
+
+/* Check that all mount ids are in decreasing order. */
+TEST(listmount_backward)
+{
+ uint64_t list[LISTMNT_BUFFER], last_mnt_id = 0;
+
+ for (;;) {
+ ssize_t nr_mounts;
+
+ nr_mounts = listmount(LSMT_ROOT, 0, last_mnt_id,
+ list, LISTMNT_BUFFER, LISTMOUNT_REVERSE);
+ ASSERT_GE(nr_mounts, 0);
+ if (nr_mounts == 0)
+ break;
+
+ for (size_t cur = 0; cur < nr_mounts; cur++) {
+ if (cur < nr_mounts - 1)
+ ASSERT_GT(list[cur], list[cur + 1]);
+ last_mnt_id = list[cur];
+ }
+ }
+}
+
+TEST_HARNESS_MAIN
diff --git a/tools/testing/selftests/filesystems/statmount/statmount.h b/tools/testing/selftests/filesystems/statmount/statmount.h
new file mode 100644
index 000000000000..f4294bab9d73
--- /dev/null
+++ b/tools/testing/selftests/filesystems/statmount/statmount.h
@@ -0,0 +1,46 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __STATMOUNT_H
+#define __STATMOUNT_H
+
+#include <stdint.h>
+#include <linux/mount.h>
+#include <asm/unistd.h>
+
+static inline int statmount(uint64_t mnt_id, uint64_t mnt_ns_id, uint64_t mask,
+ struct statmount *buf, size_t bufsize,
+ unsigned int flags)
+{
+ struct mnt_id_req req = {
+ .size = MNT_ID_REQ_SIZE_VER0,
+ .mnt_id = mnt_id,
+ .param = mask,
+ };
+
+ if (mnt_ns_id) {
+ req.size = MNT_ID_REQ_SIZE_VER1;
+ req.mnt_ns_id = mnt_ns_id;
+ }
+
+ return syscall(__NR_statmount, &req, buf, bufsize, flags);
+}
+
+static ssize_t listmount(uint64_t mnt_id, uint64_t mnt_ns_id,
+ uint64_t last_mnt_id, uint64_t list[], size_t num,
+ unsigned int flags)
+{
+ struct mnt_id_req req = {
+ .size = MNT_ID_REQ_SIZE_VER0,
+ .mnt_id = mnt_id,
+ .param = last_mnt_id,
+ };
+
+ if (mnt_ns_id) {
+ req.size = MNT_ID_REQ_SIZE_VER1;
+ req.mnt_ns_id = mnt_ns_id;
+ }
+
+ return syscall(__NR_listmount, &req, list, num, flags);
+}
+
+#endif /* __STATMOUNT_H */
diff --git a/tools/testing/selftests/filesystems/statmount/statmount_test.c b/tools/testing/selftests/filesystems/statmount/statmount_test.c
index 3eafd7da58e2..46d289611ce8 100644
--- a/tools/testing/selftests/filesystems/statmount/statmount_test.c
+++ b/tools/testing/selftests/filesystems/statmount/statmount_test.c
@@ -3,17 +3,16 @@
#define _GNU_SOURCE
#include <assert.h>
-#include <stdint.h>
+#include <stddef.h>
#include <sched.h>
#include <fcntl.h>
#include <sys/param.h>
#include <sys/mount.h>
#include <sys/stat.h>
#include <sys/statfs.h>
-#include <linux/mount.h>
#include <linux/stat.h>
-#include <asm/unistd.h>
+#include "statmount.h"
#include "../../kselftest.h"
static const char *const known_fs[] = {
@@ -28,25 +27,13 @@ static const char *const known_fs[] = {
"ipathfs", "iso9660", "jffs2", "jfs", "minix", "mqueue", "msdos",
"nfs", "nfs4", "nfsd", "nilfs2", "nsfs", "ntfs", "ntfs3", "ocfs2",
"ocfs2_dlmfs", "ocxlflash", "omfs", "openpromfs", "overlay", "pipefs",
- "proc", "pstore", "pvfs2", "qnx4", "qnx6", "ramfs", "reiserfs",
+ "proc", "pstore", "pvfs2", "qnx4", "qnx6", "ramfs",
"resctrl", "romfs", "rootfs", "rpc_pipefs", "s390_hypfs", "secretmem",
"securityfs", "selinuxfs", "smackfs", "smb3", "sockfs", "spufs",
"squashfs", "sysfs", "sysv", "tmpfs", "tracefs", "ubifs", "udf",
"ufs", "v7", "vboxsf", "vfat", "virtiofs", "vxfs", "xenfs", "xfs",
"zonefs", NULL };
-static int statmount(uint64_t mnt_id, uint64_t mask, struct statmount *buf,
- size_t bufsize, unsigned int flags)
-{
- struct mnt_id_req req = {
- .size = MNT_ID_REQ_SIZE_VER0,
- .mnt_id = mnt_id,
- .param = mask,
- };
-
- return syscall(__NR_statmount, &req, buf, bufsize, flags);
-}
-
static struct statmount *statmount_alloc(uint64_t mnt_id, uint64_t mask, unsigned int flags)
{
size_t bufsize = 1 << 15;
@@ -55,7 +42,7 @@ static struct statmount *statmount_alloc(uint64_t mnt_id, uint64_t mask, unsigne
int ret;
for (;;) {
- ret = statmount(mnt_id, mask, tmp, bufsize, flags);
+ ret = statmount(mnt_id, 0, mask, tmp, bufsize, flags);
if (ret != -1)
break;
if (tofree)
@@ -120,12 +107,20 @@ static char root_mntpoint[] = "/tmp/statmount_test_root.XXXXXX";
static int orig_root;
static uint64_t root_id, parent_id;
static uint32_t old_root_id, old_parent_id;
-
+static FILE *f_mountinfo;
static void cleanup_namespace(void)
{
- fchdir(orig_root);
- chroot(".");
+ int ret;
+
+ ret = fchdir(orig_root);
+ if (ret == -1)
+ ksft_perror("fchdir to original root");
+
+ ret = chroot(".");
+ if (ret == -1)
+ ksft_perror("chroot to original root");
+
umount2(root_mntpoint, MNT_DETACH);
rmdir(root_mntpoint);
}
@@ -137,7 +132,7 @@ static void setup_namespace(void)
uid_t uid = getuid();
gid_t gid = getgid();
- ret = unshare(CLONE_NEWNS|CLONE_NEWUSER);
+ ret = unshare(CLONE_NEWNS|CLONE_NEWUSER|CLONE_NEWPID);
if (ret == -1)
ksft_exit_fail_msg("unsharing mountns and userns: %s\n",
strerror(errno));
@@ -148,6 +143,11 @@ static void setup_namespace(void)
sprintf(buf, "0 %d 1", gid);
write_file("/proc/self/gid_map", buf);
+ f_mountinfo = fopen("/proc/self/mountinfo", "re");
+ if (!f_mountinfo)
+ ksft_exit_fail_msg("failed to open mountinfo: %s\n",
+ strerror(errno));
+
ret = mount("", "/", NULL, MS_REC|MS_PRIVATE, NULL);
if (ret == -1)
ksft_exit_fail_msg("making mount tree private: %s\n",
@@ -207,25 +207,13 @@ static int setup_mount_tree(int log2_num)
return 0;
}
-static ssize_t listmount(uint64_t mnt_id, uint64_t last_mnt_id,
- uint64_t list[], size_t num, unsigned int flags)
-{
- struct mnt_id_req req = {
- .size = MNT_ID_REQ_SIZE_VER0,
- .mnt_id = mnt_id,
- .param = last_mnt_id,
- };
-
- return syscall(__NR_listmount, &req, list, num, flags);
-}
-
static void test_listmount_empty_root(void)
{
ssize_t res;
const unsigned int size = 32;
uint64_t list[size];
- res = listmount(LSMT_ROOT, 0, list, size, 0);
+ res = listmount(LSMT_ROOT, 0, 0, list, size, 0);
if (res == -1) {
ksft_test_result_fail("listmount: %s\n", strerror(errno));
return;
@@ -250,7 +238,7 @@ static void test_statmount_zero_mask(void)
struct statmount sm;
int ret;
- ret = statmount(root_id, 0, &sm, sizeof(sm), 0);
+ ret = statmount(root_id, 0, 0, &sm, sizeof(sm), 0);
if (ret == -1) {
ksft_test_result_fail("statmount zero mask: %s\n",
strerror(errno));
@@ -276,7 +264,7 @@ static void test_statmount_mnt_basic(void)
int ret;
uint64_t mask = STATMOUNT_MNT_BASIC;
- ret = statmount(root_id, mask, &sm, sizeof(sm), 0);
+ ret = statmount(root_id, 0, mask, &sm, sizeof(sm), 0);
if (ret == -1) {
ksft_test_result_fail("statmount mnt basic: %s\n",
strerror(errno));
@@ -336,7 +324,7 @@ static void test_statmount_sb_basic(void)
struct statx sx;
struct statfs sf;
- ret = statmount(root_id, mask, &sm, sizeof(sm), 0);
+ ret = statmount(root_id, 0, mask, &sm, sizeof(sm), 0);
if (ret == -1) {
ksft_test_result_fail("statmount sb basic: %s\n",
strerror(errno));
@@ -395,6 +383,10 @@ static void test_statmount_mnt_point(void)
return;
}
+ if (!(sm->mask & STATMOUNT_MNT_POINT)) {
+ ksft_test_result_fail("missing STATMOUNT_MNT_POINT in mask\n");
+ return;
+ }
if (strcmp(sm->str + sm->mnt_point, "/") != 0) {
ksft_test_result_fail("unexpected mount point: '%s' != '/'\n",
sm->str + sm->mnt_point);
@@ -420,6 +412,10 @@ static void test_statmount_mnt_root(void)
strerror(errno));
return;
}
+ if (!(sm->mask & STATMOUNT_MNT_ROOT)) {
+ ksft_test_result_fail("missing STATMOUNT_MNT_ROOT in mask\n");
+ return;
+ }
mnt_root = sm->str + sm->mnt_root;
last_root = strrchr(mnt_root, '/');
if (last_root)
@@ -449,6 +445,10 @@ static void test_statmount_fs_type(void)
strerror(errno));
return;
}
+ if (!(sm->mask & STATMOUNT_FS_TYPE)) {
+ ksft_test_result_fail("missing STATMOUNT_FS_TYPE in mask\n");
+ return;
+ }
fs_type = sm->str + sm->fs_type;
for (s = known_fs; s != NULL; s++) {
if (strcmp(fs_type, *s) == 0)
@@ -461,6 +461,96 @@ static void test_statmount_fs_type(void)
free(sm);
}
+static void test_statmount_mnt_opts(void)
+{
+ struct statmount *sm;
+ const char *statmount_opts;
+ char *line = NULL;
+ size_t len = 0;
+
+ sm = statmount_alloc(root_id, STATMOUNT_MNT_BASIC | STATMOUNT_MNT_OPTS,
+ 0);
+ if (!sm) {
+ ksft_test_result_fail("statmount mnt opts: %s\n",
+ strerror(errno));
+ return;
+ }
+
+ if (!(sm->mask & STATMOUNT_MNT_BASIC)) {
+ ksft_test_result_fail("missing STATMOUNT_MNT_BASIC in mask\n");
+ return;
+ }
+
+ while (getline(&line, &len, f_mountinfo) != -1) {
+ int i;
+ char *p, *p2;
+ unsigned int old_mnt_id;
+
+ old_mnt_id = atoi(line);
+ if (old_mnt_id != sm->mnt_id_old)
+ continue;
+
+ for (p = line, i = 0; p && i < 5; i++)
+ p = strchr(p + 1, ' ');
+ if (!p)
+ continue;
+
+ p2 = strchr(p + 1, ' ');
+ if (!p2)
+ continue;
+ *p2 = '\0';
+ p = strchr(p2 + 1, '-');
+ if (!p)
+ continue;
+ for (p++, i = 0; p && i < 2; i++)
+ p = strchr(p + 1, ' ');
+ if (!p)
+ continue;
+ p++;
+
+ /* skip generic superblock options */
+ if (strncmp(p, "ro", 2) == 0)
+ p += 2;
+ else if (strncmp(p, "rw", 2) == 0)
+ p += 2;
+ if (*p == ',')
+ p++;
+ if (strncmp(p, "sync", 4) == 0)
+ p += 4;
+ if (*p == ',')
+ p++;
+ if (strncmp(p, "dirsync", 7) == 0)
+ p += 7;
+ if (*p == ',')
+ p++;
+ if (strncmp(p, "lazytime", 8) == 0)
+ p += 8;
+ if (*p == ',')
+ p++;
+ p2 = strrchr(p, '\n');
+ if (p2)
+ *p2 = '\0';
+
+ if (sm->mask & STATMOUNT_MNT_OPTS)
+ statmount_opts = sm->str + sm->mnt_opts;
+ else
+ statmount_opts = "";
+ if (strcmp(statmount_opts, p) != 0)
+ ksft_test_result_fail(
+ "unexpected mount options: '%s' != '%s'\n",
+ statmount_opts, p);
+ else
+ ksft_test_result_pass("statmount mount options\n");
+ free(sm);
+ free(line);
+ return;
+ }
+
+ ksft_test_result_fail("didnt't find mount entry\n");
+ free(sm);
+ free(line);
+}
+
static void test_statmount_string(uint64_t mask, size_t off, const char *name)
{
struct statmount *sm;
@@ -497,14 +587,14 @@ static void test_statmount_string(uint64_t mask, size_t off, const char *name)
exactsize = sm->size;
shortsize = sizeof(*sm) + i;
- ret = statmount(root_id, mask, sm, exactsize, 0);
+ ret = statmount(root_id, 0, mask, sm, exactsize, 0);
if (ret == -1) {
ksft_test_result_fail("statmount exact size: %s\n",
strerror(errno));
goto out;
}
errno = 0;
- ret = statmount(root_id, mask, sm, shortsize, 0);
+ ret = statmount(root_id, 0, mask, sm, shortsize, 0);
if (ret != -1 || errno != EOVERFLOW) {
ksft_test_result_fail("should have failed with EOVERFLOW: %s\n",
strerror(errno));
@@ -532,7 +622,7 @@ static void test_listmount_tree(void)
if (res == -1)
return;
- num = res = listmount(LSMT_ROOT, 0, list, size, 0);
+ num = res = listmount(LSMT_ROOT, 0, 0, list, size, 0);
if (res == -1) {
ksft_test_result_fail("listmount: %s\n", strerror(errno));
return;
@@ -544,7 +634,7 @@ static void test_listmount_tree(void)
}
for (i = 0; i < size - step;) {
- res = listmount(LSMT_ROOT, i ? list2[i - 1] : 0, list2 + i, step, 0);
+ res = listmount(LSMT_ROOT, 0, i ? list2[i - 1] : 0, list2 + i, step, 0);
if (res == -1)
ksft_test_result_fail("short listmount: %s\n",
strerror(errno));
@@ -576,18 +666,18 @@ int main(void)
int ret;
uint64_t all_mask = STATMOUNT_SB_BASIC | STATMOUNT_MNT_BASIC |
STATMOUNT_PROPAGATE_FROM | STATMOUNT_MNT_ROOT |
- STATMOUNT_MNT_POINT | STATMOUNT_FS_TYPE;
+ STATMOUNT_MNT_POINT | STATMOUNT_FS_TYPE | STATMOUNT_MNT_NS_ID;
ksft_print_header();
- ret = statmount(0, 0, NULL, 0, 0);
+ ret = statmount(0, 0, 0, NULL, 0, 0);
assert(ret == -1);
if (errno == ENOSYS)
ksft_exit_skip("statmount() syscall not supported\n");
setup_namespace();
- ksft_set_plan(14);
+ ksft_set_plan(15);
test_listmount_empty_root();
test_statmount_zero_mask();
test_statmount_mnt_basic();
@@ -595,6 +685,7 @@ int main(void)
test_statmount_mnt_root();
test_statmount_mnt_point();
test_statmount_fs_type();
+ test_statmount_mnt_opts();
test_statmount_string(STATMOUNT_MNT_ROOT, str_off(mnt_root), "mount root");
test_statmount_string(STATMOUNT_MNT_POINT, str_off(mnt_point), "mount point");
test_statmount_string(STATMOUNT_FS_TYPE, str_off(fs_type), "fs type");
diff --git a/tools/testing/selftests/filesystems/statmount/statmount_test_ns.c b/tools/testing/selftests/filesystems/statmount/statmount_test_ns.c
new file mode 100644
index 000000000000..70cb0c8b21cf
--- /dev/null
+++ b/tools/testing/selftests/filesystems/statmount/statmount_test_ns.c
@@ -0,0 +1,367 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#define _GNU_SOURCE
+
+#include <assert.h>
+#include <fcntl.h>
+#include <limits.h>
+#include <sched.h>
+#include <stdlib.h>
+#include <sys/mount.h>
+#include <sys/stat.h>
+#include <sys/wait.h>
+#include <linux/nsfs.h>
+#include <linux/stat.h>
+
+#include "statmount.h"
+#include "../../kselftest.h"
+
+#define NSID_PASS 0
+#define NSID_FAIL 1
+#define NSID_SKIP 2
+#define NSID_ERROR 3
+
+static void handle_result(int ret, const char *testname)
+{
+ if (ret == NSID_PASS)
+ ksft_test_result_pass("%s\n", testname);
+ else if (ret == NSID_FAIL)
+ ksft_test_result_fail("%s\n", testname);
+ else if (ret == NSID_ERROR)
+ ksft_exit_fail_msg("%s\n", testname);
+ else
+ ksft_test_result_skip("%s\n", testname);
+}
+
+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);
+ return ret;
+}
+
+static int get_mnt_ns_id(const char *mnt_ns, uint64_t *mnt_ns_id)
+{
+ int fd = open(mnt_ns, O_RDONLY);
+
+ if (fd < 0) {
+ ksft_print_msg("failed to open for ns %s: %s\n",
+ mnt_ns, strerror(errno));
+ sleep(60);
+ return NSID_ERROR;
+ }
+
+ if (ioctl(fd, NS_GET_MNTNS_ID, mnt_ns_id) < 0) {
+ ksft_print_msg("failed to get the nsid for ns %s: %s\n",
+ mnt_ns, strerror(errno));
+ return NSID_ERROR;
+ }
+ close(fd);
+ return NSID_PASS;
+}
+
+static int get_mnt_id(const char *path, uint64_t *mnt_id)
+{
+ struct statx sx;
+ int ret;
+
+ ret = statx(AT_FDCWD, path, 0, STATX_MNT_ID_UNIQUE, &sx);
+ if (ret == -1) {
+ ksft_print_msg("retrieving unique mount ID for %s: %s\n", path,
+ strerror(errno));
+ return NSID_ERROR;
+ }
+
+ if (!(sx.stx_mask & STATX_MNT_ID_UNIQUE)) {
+ ksft_print_msg("no unique mount ID available for %s\n", path);
+ return NSID_ERROR;
+ }
+
+ *mnt_id = sx.stx_mnt_id;
+ return NSID_PASS;
+}
+
+static int write_file(const char *path, const char *val)
+{
+ int fd = open(path, O_WRONLY);
+ size_t len = strlen(val);
+ int ret;
+
+ if (fd == -1) {
+ ksft_print_msg("opening %s for write: %s\n", path, strerror(errno));
+ return NSID_ERROR;
+ }
+
+ ret = write(fd, val, len);
+ if (ret == -1) {
+ ksft_print_msg("writing to %s: %s\n", path, strerror(errno));
+ return NSID_ERROR;
+ }
+ if (ret != len) {
+ ksft_print_msg("short write to %s\n", path);
+ return NSID_ERROR;
+ }
+
+ ret = close(fd);
+ if (ret == -1) {
+ ksft_print_msg("closing %s\n", path);
+ return NSID_ERROR;
+ }
+
+ return NSID_PASS;
+}
+
+static int setup_namespace(void)
+{
+ int ret;
+ char buf[32];
+ uid_t uid = getuid();
+ gid_t gid = getgid();
+
+ ret = unshare(CLONE_NEWNS|CLONE_NEWUSER|CLONE_NEWPID);
+ if (ret == -1)
+ ksft_exit_fail_msg("unsharing mountns and userns: %s\n",
+ strerror(errno));
+
+ sprintf(buf, "0 %d 1", uid);
+ ret = write_file("/proc/self/uid_map", buf);
+ if (ret != NSID_PASS)
+ return ret;
+ ret = write_file("/proc/self/setgroups", "deny");
+ if (ret != NSID_PASS)
+ return ret;
+ sprintf(buf, "0 %d 1", gid);
+ ret = write_file("/proc/self/gid_map", buf);
+ if (ret != NSID_PASS)
+ return ret;
+
+ ret = mount("", "/", NULL, MS_REC|MS_PRIVATE, NULL);
+ if (ret == -1) {
+ ksft_print_msg("making mount tree private: %s\n",
+ strerror(errno));
+ return NSID_ERROR;
+ }
+
+ return NSID_PASS;
+}
+
+static int _test_statmount_mnt_ns_id(void)
+{
+ struct statmount sm;
+ uint64_t mnt_ns_id;
+ uint64_t root_id;
+ int ret;
+
+ ret = get_mnt_ns_id("/proc/self/ns/mnt", &mnt_ns_id);
+ if (ret != NSID_PASS)
+ return ret;
+
+ ret = get_mnt_id("/", &root_id);
+ if (ret != NSID_PASS)
+ return ret;
+
+ ret = statmount(root_id, 0, STATMOUNT_MNT_NS_ID, &sm, sizeof(sm), 0);
+ if (ret == -1) {
+ ksft_print_msg("statmount mnt ns id: %s\n", strerror(errno));
+ return NSID_ERROR;
+ }
+
+ if (sm.size != sizeof(sm)) {
+ ksft_print_msg("unexpected size: %u != %u\n", sm.size,
+ (uint32_t)sizeof(sm));
+ return NSID_FAIL;
+ }
+ if (sm.mask != STATMOUNT_MNT_NS_ID) {
+ ksft_print_msg("statmount mnt ns id unavailable\n");
+ return NSID_SKIP;
+ }
+
+ if (sm.mnt_ns_id != mnt_ns_id) {
+ ksft_print_msg("unexpected mnt ns ID: 0x%llx != 0x%llx\n",
+ (unsigned long long)sm.mnt_ns_id,
+ (unsigned long long)mnt_ns_id);
+ return NSID_FAIL;
+ }
+
+ return NSID_PASS;
+}
+
+static void test_statmount_mnt_ns_id(void)
+{
+ pid_t pid;
+ int ret;
+
+ pid = fork();
+ if (pid < 0)
+ ksft_exit_fail_msg("failed to fork: %s\n", strerror(errno));
+
+ /* We're the original pid, wait for the result. */
+ if (pid != 0) {
+ ret = wait_for_pid(pid);
+ handle_result(ret, "test statmount ns id");
+ return;
+ }
+
+ ret = setup_namespace();
+ if (ret != NSID_PASS)
+ exit(ret);
+ ret = _test_statmount_mnt_ns_id();
+ exit(ret);
+}
+
+static int validate_external_listmount(pid_t pid, uint64_t child_nr_mounts)
+{
+ uint64_t list[256];
+ uint64_t mnt_ns_id;
+ uint64_t nr_mounts;
+ char buf[256];
+ int ret;
+
+ /* Get the mount ns id for our child. */
+ snprintf(buf, sizeof(buf), "/proc/%lu/ns/mnt", (unsigned long)pid);
+ ret = get_mnt_ns_id(buf, &mnt_ns_id);
+
+ nr_mounts = listmount(LSMT_ROOT, mnt_ns_id, 0, list, 256, 0);
+ if (nr_mounts == (uint64_t)-1) {
+ ksft_print_msg("listmount: %s\n", strerror(errno));
+ return NSID_ERROR;
+ }
+
+ if (nr_mounts != child_nr_mounts) {
+ ksft_print_msg("listmount results is %zi != %zi\n", nr_mounts,
+ child_nr_mounts);
+ return NSID_FAIL;
+ }
+
+ /* Validate that all of our entries match our mnt_ns_id. */
+ for (int i = 0; i < nr_mounts; i++) {
+ struct statmount sm;
+
+ ret = statmount(list[i], mnt_ns_id, STATMOUNT_MNT_NS_ID, &sm,
+ sizeof(sm), 0);
+ if (ret < 0) {
+ ksft_print_msg("statmount mnt ns id: %s\n", strerror(errno));
+ return NSID_ERROR;
+ }
+
+ if (sm.mask != STATMOUNT_MNT_NS_ID) {
+ ksft_print_msg("statmount mnt ns id unavailable\n");
+ return NSID_SKIP;
+ }
+
+ if (sm.mnt_ns_id != mnt_ns_id) {
+ ksft_print_msg("listmount gave us the wrong ns id: 0x%llx != 0x%llx\n",
+ (unsigned long long)sm.mnt_ns_id,
+ (unsigned long long)mnt_ns_id);
+ return NSID_FAIL;
+ }
+ }
+
+ return NSID_PASS;
+}
+
+static void test_listmount_ns(void)
+{
+ uint64_t nr_mounts;
+ char pval;
+ int child_ready_pipe[2];
+ int parent_ready_pipe[2];
+ pid_t pid;
+ int ret, child_ret;
+
+ if (pipe(child_ready_pipe) < 0)
+ ksft_exit_fail_msg("failed to create the child pipe: %s\n",
+ strerror(errno));
+ if (pipe(parent_ready_pipe) < 0)
+ ksft_exit_fail_msg("failed to create the parent pipe: %s\n",
+ strerror(errno));
+
+ pid = fork();
+ if (pid < 0)
+ ksft_exit_fail_msg("failed to fork: %s\n", strerror(errno));
+
+ if (pid == 0) {
+ char cval;
+ uint64_t list[256];
+
+ close(child_ready_pipe[0]);
+ close(parent_ready_pipe[1]);
+
+ ret = setup_namespace();
+ if (ret != NSID_PASS)
+ exit(ret);
+
+ nr_mounts = listmount(LSMT_ROOT, 0, 0, list, 256, 0);
+ if (nr_mounts == (uint64_t)-1) {
+ ksft_print_msg("listmount: %s\n", strerror(errno));
+ exit(NSID_FAIL);
+ }
+
+ /*
+ * Tell our parent how many mounts we have, and then wait for it
+ * to tell us we're done.
+ */
+ if (write(child_ready_pipe[1], &nr_mounts, sizeof(nr_mounts)) !=
+ sizeof(nr_mounts))
+ ret = NSID_ERROR;
+ if (read(parent_ready_pipe[0], &cval, sizeof(cval)) != sizeof(cval))
+ ret = NSID_ERROR;
+ exit(NSID_PASS);
+ }
+
+ close(child_ready_pipe[1]);
+ close(parent_ready_pipe[0]);
+
+ /* Wait until the child has created everything. */
+ if (read(child_ready_pipe[0], &nr_mounts, sizeof(nr_mounts)) !=
+ sizeof(nr_mounts))
+ ret = NSID_ERROR;
+
+ ret = validate_external_listmount(pid, nr_mounts);
+
+ if (write(parent_ready_pipe[1], &pval, sizeof(pval)) != sizeof(pval))
+ ret = NSID_ERROR;
+
+ child_ret = wait_for_pid(pid);
+ if (child_ret != NSID_PASS)
+ ret = child_ret;
+ handle_result(ret, "test listmount ns id");
+}
+
+int main(void)
+{
+ int ret;
+
+ ksft_print_header();
+ ret = statmount(0, 0, 0, NULL, 0, 0);
+ assert(ret == -1);
+ if (errno == ENOSYS)
+ ksft_exit_skip("statmount() syscall not supported\n");
+
+ ksft_set_plan(2);
+ test_statmount_mnt_ns_id();
+ test_listmount_ns();
+
+ if (ksft_get_fail_cnt() + ksft_get_error_cnt() > 0)
+ ksft_exit_fail();
+ else
+ ksft_exit_pass();
+}