sched_ext: Maintain an online cid mask in the scheduler arenagrafted

Schedulers on the default cid mapping treat [0, nr_online_cids) as the
online set and restart on hotplug. Schedulers that install their own mapping
with scx_bpf_cid_override() have no way to learn which cids are online: the
count no longer identifies members and the CPU-form cpumask is unusable from
cid programs. This is an obvious hole in the cid API.

Add scx_bpf_online_cmask(), a kernel-maintained cmask in the scheduler's
arena, allocated alongside the per-CPU scratch masks and populated after the
cid mapping is finalized and before ops.init(), for child schedulers too.
The pointer stays valid through ops.exit() with no reference to take. It is
the arena offset as a void pointer, the same form struct_ops arena arguments
arrive in. The verifier types the void return as a scalar for the program's
arena cast.

The mask follows the SCX hotplug notifications: seeded from cpu_active_mask
and updated before ops.cid_online/offline() runs, so it lags cpu_online_mask
only inside a hotplug transition. Updates walk the scheduler list under the
lock that also serializes unlinking. Reads are live, not atomic snapshots.
Root initialization excludes hotplug.

v2: Reworded the getter kerneldoc (Andrea Righi).

Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Andrea Righi <arighi@nvidia.com>

3 files changed, 217 insertions, 0 deletions

diff --git a/tools/testing/selftests/fchmodat2/.gitignore b/tools/testing/selftests/fchmodat2/.gitignore
new file mode 100644
index 000000000..82a4846cb
--- /dev/null
+++ b/tools/testing/selftests/fchmodat2/.gitignore
@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0-only
+/*_test
diff --git a/tools/testing/selftests/fchmodat2/Makefile b/tools/testing/selftests/fchmodat2/Makefile
new file mode 100644
index 000000000..4373cea79
--- /dev/null
+++ b/tools/testing/selftests/fchmodat2/Makefile
@@ -0,0 +1,15 @@
+# SPDX-License-Identifier: GPL-2.0-or-later
+
+CFLAGS += -Wall -O2 -g -fsanitize=address -fsanitize=undefined $(KHDR_INCLUDES)
+
+# gcc requires -static-libasan in order to ensure that Address Sanitizer's
+# library is the first one loaded. However, clang already statically links the
+# Address Sanitizer if -fsanitize is specified. Therefore, simply omit
+# -static-libasan for clang builds.
+ifeq ($(LLVM),)
+    CFLAGS += -static-libasan
+endif
+
+TEST_GEN_PROGS := fchmodat2_test
+
+include ../lib.mk
diff --git a/tools/testing/selftests/fchmodat2/fchmodat2_test.c b/tools/testing/selftests/fchmodat2/fchmodat2_test.c
new file mode 100644
index 000000000..8e1b6384b
--- /dev/null
+++ b/tools/testing/selftests/fchmodat2/fchmodat2_test.c
@@ -0,0 +1,200 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#define _GNU_SOURCE
+#include <fcntl.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <syscall.h>
+#include <unistd.h>
+
+#include "kselftest.h"
+
+struct testdir {
+	char *dirname;
+	int dfd;
+};
+
+int sys_fchmodat2(int dfd, const char *filename, mode_t mode, int flags)
+{
+	int ret = syscall(__NR_fchmodat2, dfd, filename, mode, flags);
+
+	return ret >= 0 ? ret : -errno;
+}
+
+static void setup_testdir(struct testdir *testdir)
+{
+	int ret, dfd;
+	char dirname[] = "/tmp/ksft-fchmodat2.XXXXXX";
+
+	/* Make the top-level directory. */
+	if (!mkdtemp(dirname))
+		ksft_exit_fail_msg("%s: failed to create tmpdir\n", __func__);
+
+	dfd = open(dirname, O_PATH | O_DIRECTORY);
+	if (dfd < 0) {
+		ksft_perror("failed to open tmpdir");
+		goto err;
+	}
+
+	ret = openat(dfd, "regfile", O_CREAT | O_WRONLY | O_TRUNC, 0644);
+	if (ret < 0) {
+		ksft_perror("failed to create file in tmpdir");
+		goto err;
+	}
+	close(ret);
+
+	ret = symlinkat("regfile", dfd, "symlink");
+	if (ret < 0) {
+		ksft_perror("symlinkat() failed");
+		goto err_regfile;
+	}
+
+	testdir->dirname = strdup(dirname);
+	if (!testdir->dirname) {
+		ksft_perror("Out of memory");
+		goto err_symlink;
+	}
+	testdir->dfd = dfd;
+
+	return;
+
+err_symlink:
+	unlinkat(testdir->dfd, "symlink", 0);
+err_regfile:
+	unlinkat(testdir->dfd, "regfile", 0);
+err:
+	unlink(dirname);
+	ksft_exit_fail();
+}
+
+static void cleanup_testdir(struct testdir *testdir)
+{
+	unlinkat(testdir->dfd, "regfile", 0);
+	unlinkat(testdir->dfd, "symlink", 0);
+	rmdir(testdir->dirname);
+	free(testdir->dirname);
+}
+
+int expect_mode(int dfd, const char *filename, mode_t expect_mode)
+{
+	struct stat st;
+	int ret = fstatat(dfd, filename, &st, AT_SYMLINK_NOFOLLOW);
+
+	if (ret) {
+		ksft_perror("fstatat() failed\n");
+		return 0;
+	}
+
+	return (st.st_mode == expect_mode);
+}
+
+void test_regfile(void)
+{
+	struct testdir testdir;
+	int ret;
+
+	setup_testdir(&testdir);
+
+	ret = sys_fchmodat2(testdir.dfd, "regfile", 0640, 0);
+
+	if (ret < 0) {
+		ksft_perror("fchmodat2(noflag) failed");
+		goto out;
+	}
+
+	if (!expect_mode(testdir.dfd, "regfile", 0100640)) {
+		ksft_print_msg("%s: wrong file mode bits after fchmodat2\n",
+				__func__);
+		ret = 1;
+		goto out;
+	}
+
+	ret = sys_fchmodat2(testdir.dfd, "regfile", 0600, AT_SYMLINK_NOFOLLOW);
+
+	if (ret < 0) {
+		ksft_perror("fchmodat2(AT_SYMLINK_NOFOLLOW) failed");
+		goto out;
+	}
+
+	if (!expect_mode(testdir.dfd, "regfile", 0100600)) {
+		ksft_print_msg("%s: wrong file mode bits after fchmodat2 with nofollow\n",
+			       __func__);
+		ret = 1;
+	}
+
+out:
+	ksft_test_result(ret == 0, "fchmodat2(regfile)\n");
+	cleanup_testdir(&testdir);
+}
+
+void test_symlink(void)
+{
+	struct testdir testdir;
+	int ret;
+
+	setup_testdir(&testdir);
+
+	ret = sys_fchmodat2(testdir.dfd, "symlink", 0640, 0);
+
+	if (ret < 0) {
+		ksft_perror("fchmodat2(noflag) failed");
+		goto err;
+	}
+
+	if (!expect_mode(testdir.dfd, "regfile", 0100640)) {
+		ksft_print_msg("%s: wrong file mode bits after fchmodat2\n",
+			       __func__);
+		goto err;
+	}
+
+	if (!expect_mode(testdir.dfd, "symlink", 0120777)) {
+		ksft_print_msg("%s: wrong symlink mode bits after fchmodat2\n",
+			       __func__);
+		goto err;
+	}
+
+	ret = sys_fchmodat2(testdir.dfd, "symlink", 0600, AT_SYMLINK_NOFOLLOW);
+
+	/*
+	 * On certain filesystems (xfs or btrfs), chmod operation fails. So we
+	 * first check the symlink target but if the operation fails we mark the
+	 * test as skipped.
+	 *
+	 * https://sourceware.org/legacy-ml/libc-alpha/2020-02/msg00467.html
+	 */
+	if (ret == 0 && !expect_mode(testdir.dfd, "symlink", 0120600)) {
+		ksft_print_msg("%s: wrong symlink mode bits after fchmodat2 with nofollow\n",
+				__func__);
+		ret = 1;
+		goto err;
+	}
+
+	if (!expect_mode(testdir.dfd, "regfile", 0100640)) {
+		ksft_print_msg("%s: wrong file mode bits after fchmodat2 with nofollow\n",
+			       __func__);
+	}
+
+	if (ret != 0)
+		ksft_test_result_skip("fchmodat2(symlink)\n");
+	else
+		ksft_test_result_pass("fchmodat2(symlink)\n");
+	cleanup_testdir(&testdir);
+	return;
+
+err:
+	ksft_test_result_fail("fchmodat2(symlink)\n");
+	cleanup_testdir(&testdir);
+}
+
+#define NUM_TESTS 2
+
+int main(int argc, char **argv)
+{
+	ksft_print_header();
+	ksft_set_plan(NUM_TESTS);
+
+	test_regfile();
+	test_symlink();
+
+	ksft_finished();
+}