Merge tag 'kernel-7.2-rc1.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs

Pull misc kernel updates from Christian Brauner:
 "Fixes

   - rhashtable: give each instance its own lockdep class

     syzbot reported a circular locking dependency between ht->mutex and
     fs_reclaim via the simple_xattrs rhashtable being torn down during
     inode eviction.

     The predicted deadlock cannot occur: rhashtable_free_and_destroy()
     cancels the deferred worker before taking ht->mutex and
     acquisitions on distinct rhashtables are on distinct mutexes.

     Lockdep flags a cycle anyway because every ht->mutex in the kernel
     shared the single static lockdep class from
     rhashtable_init_noprof().

     The lockdep key is lifted to a per-call-site static key so every
     rhashtable instance gets its own class.

   - selftests/clone3: fix misuse of the libcap library interface in the
     cap_checkpoint_restore test and remove unused variables

   - selftests/pid_namespace: compute the pid_max test limits
     dynamically instead of hardcoding values below the kernel-enforced
     minimum of PIDS_PER_CPU_MIN * num_possible_cpus() which made the
     tests fail on machines with many possible CPUs

   - selftests: fix the Makefile TARGETS entry for nsfs which wasn't
     adjusted when the tests moved under filesystems/

  Cleanups

   - ipc/sem.c: use unsigned int for nsops to match the declaration in
     syscalls.h"

* tag 'kernel-7.2-rc1.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
  selftests/clone3: remove unused variables
  selftests/clone3: fix libcap interface usage
  ipc/sem.c: use unsigned int for nsops
  selftests: Fix Makefile target for nsfs
  rhashtable: give each instance its own lockdep class
  selftests/pid_namespace: compute pid_max test limits dynamically

3 files changed, 130 insertions, 52 deletions

diff --git a/tools/testing/selftests/Makefile b/tools/testing/selftests/Makefile
index 6e59b8f63e41..641a180fb35f 100644
--- a/tools/testing/selftests/Makefile
+++ b/tools/testing/selftests/Makefile
@@ -37,6 +37,7 @@ TARGETS += filesystems/fat
 TARGETS += filesystems/overlayfs
 TARGETS += filesystems/statmount
 TARGETS += filesystems/mount-notify
+TARGETS += filesystems/nsfs
 TARGETS += filesystems/fuse
 TARGETS += filesystems/move_mount
 TARGETS += filesystems/empty_mntns
@@ -85,7 +86,6 @@ TARGETS += net/ppp
 TARGETS += net/rds
 TARGETS += net/tcp_ao
 TARGETS += nolibc
-TARGETS += nsfs
 TARGETS += pci_endpoint
 TARGETS += pcie_bwctrl
 TARGETS += perf_events
diff --git a/tools/testing/selftests/clone3/clone3_cap_checkpoint_restore.c b/tools/testing/selftests/clone3/clone3_cap_checkpoint_restore.c
index e82281efa273..ab62bcf4107d 100644
--- a/tools/testing/selftests/clone3/clone3_cap_checkpoint_restore.c
+++ b/tools/testing/selftests/clone3/clone3_cap_checkpoint_restore.c
@@ -53,9 +53,6 @@ static int call_clone3_set_tid(struct __test_metadata *_metadata,
 	}
 
 	if (pid == 0) {
-		int ret;
-		char tmp = 0;
-
 		TH_LOG("I am the child, my PID is %d (expected %d)", getpid(), set_tid[0]);
 
 		if (set_tid[0] != getpid())
@@ -87,15 +84,11 @@ static int test_clone3_set_tid(struct __test_metadata *_metadata,
 	return ret;
 }
 
-struct libcap {
-	struct __user_cap_header_struct hdr;
-	struct __user_cap_data_struct data[2];
-};
-
 static int set_capability(void)
 {
-	cap_value_t cap_values[] = { CAP_SETUID, CAP_SETGID };
-	struct libcap *cap;
+	cap_value_t cap_values[] = {
+		CAP_SETUID, CAP_SETGID, CAP_CHECKPOINT_RESTORE
+	};
 	int ret = -1;
 	cap_t caps;
 
@@ -111,14 +104,8 @@ static int set_capability(void)
 		goto out;
 	}
 
-	cap_set_flag(caps, CAP_EFFECTIVE, 2, cap_values, CAP_SET);
-	cap_set_flag(caps, CAP_PERMITTED, 2, cap_values, CAP_SET);
-
-	cap = (struct libcap *) caps;
-
-	/* 40 -> CAP_CHECKPOINT_RESTORE */
-	cap->data[1].effective |= 1 << (40 - 32);
-	cap->data[1].permitted |= 1 << (40 - 32);
+	cap_set_flag(caps, CAP_EFFECTIVE, 3, cap_values, CAP_SET);
+	cap_set_flag(caps, CAP_PERMITTED, 3, cap_values, CAP_SET);
 
 	if (cap_set_proc(caps)) {
 		perror("cap_set_proc");
@@ -135,7 +122,6 @@ TEST(clone3_cap_checkpoint_restore)
 {
 	pid_t pid;
 	int status;
-	int ret = 0;
 	pid_t set_tid[1];
 
 	test_clone3_supported();
diff --git a/tools/testing/selftests/pid_namespace/pid_max.c b/tools/testing/selftests/pid_namespace/pid_max.c
index c9519e7385b6..5d686a09aa15 100644
--- a/tools/testing/selftests/pid_namespace/pid_max.c
+++ b/tools/testing/selftests/pid_namespace/pid_max.c
@@ -12,10 +12,74 @@
 #include <syscall.h>
 #include <sys/mount.h>
 #include <sys/wait.h>
+#include <unistd.h>
 
 #include "kselftest_harness.h"
 #include "../pidfd/pidfd.h"
 
+/*
+ * The kernel computes the minimum allowed pid_max as:
+ *   max(RESERVED_PIDS + 1, PIDS_PER_CPU_MIN * num_possible_cpus())
+ * Mirror that here so the test values are always valid.
+ *
+ * Note: glibc's get_nprocs_conf() returns the number of *configured*
+ * (present) CPUs, not *possible* CPUs.  The kernel uses
+ * num_possible_cpus() which corresponds to /sys/devices/system/cpu/possible.
+ * These can differ significantly (e.g. 16 configured vs 128 possible).
+ */
+#define RESERVED_PIDS		300
+#define PIDS_PER_CPU_MIN	8
+
+/* Count CPUs from a range list like "0-31" or "0-15,32-47". */
+static int num_possible_cpus(void)
+{
+	FILE *f;
+	int count = 0;
+	int lo, hi;
+
+	f = fopen("/sys/devices/system/cpu/possible", "r");
+	if (!f)
+		return 0;
+
+	while (fscanf(f, "%d", &lo) == 1) {
+		if (fscanf(f, "-%d", &hi) == 1)
+			count += hi - lo + 1;
+		else
+			count++;
+		/* skip comma separator */
+		fscanf(f, ",");
+	}
+
+	fclose(f);
+	return count;
+}
+
+static int pid_min(void)
+{
+	int cpu_min = PIDS_PER_CPU_MIN * num_possible_cpus();
+
+	return cpu_min > (RESERVED_PIDS + 1) ? cpu_min : (RESERVED_PIDS + 1);
+}
+
+/*
+ * Outer and inner pid_max limits used by the tests.  The outer limit is
+ * the more restrictive ancestor; the inner limit is set higher in a
+ * nested namespace but must still be capped by the outer limit.
+ * Both are derived from the kernel's minimum so they are always writable.
+ *
+ * Global so that clone callbacks can access them without parameter plumbing.
+ */
+static int outer_limit;
+static int inner_limit;
+
+static int write_int_to_fd(int fd, int val)
+{
+	char buf[12];
+	int len = snprintf(buf, sizeof(buf), "%d", val);
+
+	return write(fd, buf, len);
+}
+
 #define __STACK_SIZE (8 * 1024 * 1024)
 static pid_t do_clone(int (*fn)(void *), void *arg, int flags)
 {
@@ -60,18 +124,18 @@ static int pid_max_cb(void *data)
 		return -1;
 	}
 
-	ret = write(fd, "500", sizeof("500") - 1);
+	ret = write_int_to_fd(fd, inner_limit);
 	if (ret < 0) {
 		fprintf(stderr, "%m - Failed to write pid_max\n");
 		return -1;
 	}
 
-	for (int i = 0; i < 501; i++) {
+	for (int i = 0; i < inner_limit + 1; i++) {
 		pid = fork();
 		if (pid == 0)
 			exit(EXIT_SUCCESS);
 		wait_for_pid(pid);
-		if (pid > 500) {
+		if (pid > inner_limit) {
 			fprintf(stderr, "Managed to create pid number beyond limit\n");
 			return -1;
 		}
@@ -106,7 +170,7 @@ static int pid_max_nested_inner(void *data)
 		return fret;
 	}
 
-	ret = write(fd, "500", sizeof("500") - 1);
+	ret = write_int_to_fd(fd, inner_limit);
 	close(fd);
 	if (ret < 0) {
 		fprintf(stderr, "%m - Failed to write pid_max\n");
@@ -133,8 +197,8 @@ static int pid_max_nested_inner(void *data)
 		return fret;
 	}
 
-	/* Now make sure that we wrap pids at 400. */
-	for (i = 0; i < 510; i++) {
+	/* Now make sure that we wrap pids at outer_limit. */
+	for (i = 0; i < inner_limit + 10; i++) {
 		pid_t pid;
 
 		pid = fork();
@@ -145,7 +209,7 @@ static int pid_max_nested_inner(void *data)
 			exit(EXIT_SUCCESS);
 
 		wait_for_pid(pid);
-		if (pid >= 500) {
+		if (pid >= inner_limit) {
 			fprintf(stderr, "Managed to create process with pid %d beyond configured limit\n", pid);
 			return fret;
 		}
@@ -156,15 +220,19 @@ static int pid_max_nested_inner(void *data)
 
 static int pid_max_nested_outer(void *data)
 {
-	int fret = -1, nr_procs = 400;
-	pid_t pids[1000];
-	int fd, i, ret;
+	int fret = -1, nr_procs = 0;
+	pid_t *pids;
+	int fd, ret;
 	pid_t pid;
 
+	pids = malloc(outer_limit * sizeof(pid_t));
+	if (!pids)
+		return -1;
+
 	ret = mount("", "/", NULL, MS_PRIVATE | MS_REC, 0);
 	if (ret) {
 		fprintf(stderr, "%m - Failed to make rootfs private mount\n");
-		return fret;
+		goto out;
 	}
 
 	umount2("/proc", MNT_DETACH);
@@ -172,27 +240,28 @@ static int pid_max_nested_outer(void *data)
 	ret = mount("proc", "/proc", "proc", 0, NULL);
 	if (ret) {
 		fprintf(stderr, "%m - Failed to mount proc\n");
-		return fret;
+		goto out;
 	}
 
 	fd = open("/proc/sys/kernel/pid_max", O_RDWR | O_CLOEXEC | O_NOCTTY);
 	if (fd < 0) {
 		fprintf(stderr, "%m - Failed to open pid_max\n");
-		return fret;
+		goto out;
 	}
 
-	ret = write(fd, "400", sizeof("400") - 1);
+	ret = write_int_to_fd(fd, outer_limit);
 	close(fd);
 	if (ret < 0) {
 		fprintf(stderr, "%m - Failed to write pid_max\n");
-		return fret;
+		goto out;
 	}
 
 	/*
-	 * Create 397 processes. This leaves room for do_clone() (398) and
-	 * one more 399. So creating another process needs to fail.
+	 * Create (outer_limit - 4) processes. This leaves room for
+	 * do_clone() and one more. So creating another process needs
+	 * to fail.
 	 */
-	for (nr_procs = 0; nr_procs < 396; nr_procs++) {
+	for (nr_procs = 0; nr_procs < outer_limit - 4; nr_procs++) {
 		pid = fork();
 		if (pid < 0)
 			goto reap;
@@ -220,20 +289,26 @@ reap:
 	for (int i = 0; i < nr_procs; i++)
 		wait_for_pid(pids[i]);
 
+out:
+	free(pids);
 	return fret;
 }
 
 static int pid_max_nested_limit_inner(void *data)
 {
-	int fret = -1, nr_procs = 400;
+	int fret = -1, nr_procs = 0;
 	int fd, ret;
 	pid_t pid;
-	pid_t pids[1000];
+	pid_t *pids;
+
+	pids = malloc(inner_limit * sizeof(pid_t));
+	if (!pids)
+		return -1;
 
 	ret = mount("", "/", NULL, MS_PRIVATE | MS_REC, 0);
 	if (ret) {
 		fprintf(stderr, "%m - Failed to make rootfs private mount\n");
-		return fret;
+		goto out;
 	}
 
 	umount2("/proc", MNT_DETACH);
@@ -241,23 +316,23 @@ static int pid_max_nested_limit_inner(void *data)
 	ret = mount("proc", "/proc", "proc", 0, NULL);
 	if (ret) {
 		fprintf(stderr, "%m - Failed to mount proc\n");
-		return fret;
+		goto out;
 	}
 
 	fd = open("/proc/sys/kernel/pid_max", O_RDWR | O_CLOEXEC | O_NOCTTY);
 	if (fd < 0) {
 		fprintf(stderr, "%m - Failed to open pid_max\n");
-		return fret;
+		goto out;
 	}
 
-	ret = write(fd, "500", sizeof("500") - 1);
+	ret = write_int_to_fd(fd, inner_limit);
 	close(fd);
 	if (ret < 0) {
 		fprintf(stderr, "%m - Failed to write pid_max\n");
-		return fret;
+		goto out;
 	}
 
-	for (nr_procs = 0; nr_procs < 500; nr_procs++) {
+	for (nr_procs = 0; nr_procs < inner_limit; nr_procs++) {
 		pid = fork();
 		if (pid < 0)
 			break;
@@ -268,7 +343,7 @@ static int pid_max_nested_limit_inner(void *data)
 		pids[nr_procs] = pid;
 	}
 
-	if (nr_procs >= 400) {
+	if (nr_procs >= outer_limit) {
 		fprintf(stderr, "Managed to create processes beyond the configured outer limit\n");
 		goto reap;
 	}
@@ -279,6 +354,8 @@ reap:
 	for (int i = 0; i < nr_procs; i++)
 		wait_for_pid(pids[i]);
 
+out:
+	free(pids);
 	return fret;
 }
 
@@ -307,7 +384,7 @@ static int pid_max_nested_limit_outer(void *data)
 		return -1;
 	}
 
-	ret = write(fd, "400", sizeof("400") - 1);
+	ret = write_int_to_fd(fd, outer_limit);
 	close(fd);
 	if (ret < 0) {
 		fprintf(stderr, "%m - Failed to write pid_max\n");
@@ -328,17 +405,32 @@ static int pid_max_nested_limit_outer(void *data)
 	return 0;
 }
 
-TEST(pid_max_simple)
+FIXTURE(pid_max) {
+	int dummy;
+};
+
+FIXTURE_SETUP(pid_max)
 {
-	pid_t pid;
+	int min = pid_min();
 
+	outer_limit = min + 100;
+	inner_limit = min + 200;
+}
+
+FIXTURE_TEARDOWN(pid_max)
+{
+}
+
+TEST_F(pid_max, simple)
+{
+	pid_t pid;
 
 	pid = do_clone(pid_max_cb, NULL, CLONE_NEWPID | CLONE_NEWNS);
 	ASSERT_GT(pid, 0);
 	ASSERT_EQ(0, wait_for_pid(pid));
 }
 
-TEST(pid_max_nested_limit)
+TEST_F(pid_max, nested_limit)
 {
 	pid_t pid;
 
@@ -347,7 +439,7 @@ TEST(pid_max_nested_limit)
 	ASSERT_EQ(0, wait_for_pid(pid));
 }
 
-TEST(pid_max_nested)
+TEST_F(pid_max, nested)
 {
 	pid_t pid;