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>

1 files changed, 115 insertions, 0 deletions

diff --git a/tools/perf/tests/util.c b/tools/perf/tests/util.c
new file mode 100644
index 000000000..f9abd9911
--- /dev/null
+++ b/tools/perf/tests/util.c
@@ -0,0 +1,115 @@
+// SPDX-License-Identifier: GPL-2.0
+#include "tests.h"
+#include "util/blake2s.h"
+#include "util/debug.h"
+
+#include <linux/compiler.h>
+#include <stdlib.h>
+#include <string2.h>
+
+static int test_strreplace(char needle, const char *haystack,
+			   const char *replace, const char *expected)
+{
+	char *new = strreplace_chars(needle, haystack, replace);
+	int ret = strcmp(new, expected);
+
+	free(new);
+	return ret == 0;
+}
+
+/* Maximum data length tested by test_blake2s() */
+#define MAX_DATA_LEN 512
+
+/*
+ * Hash length tested by test_blake2s().  BLAKE2s supports variable-length
+ * hashes.  However, the only user of BLAKE2s in 'perf' uses 20-byte hashes,
+ * matching the length of the ELF build ID field.  So that's the length we test.
+ */
+#define HASH_LEN 20
+
+/* Test the implementation of the BLAKE2s hash algorithm. */
+static int test_blake2s(void)
+{
+	u8 data[MAX_DATA_LEN];
+	u8 hash[HASH_LEN];
+	u8 hash2[HASH_LEN];
+	struct blake2s_ctx main_ctx;
+	/*
+	 * This value was generated by the following Python code:
+	 *
+	 * import hashlib
+	 *
+	 * data = bytes(i % 256 for i in range(513))
+	 * h = hashlib.blake2s(digest_size=20)
+	 * for i in range(513):
+	 *     h.update(hashlib.blake2s(data=data[:i], digest_size=20).digest())
+	 * print(h.hexdigest())
+	 */
+	static const u8 expected_hash_of_hashes[20] = {
+		0xef, 0x9b, 0x13, 0x98, 0x78, 0x8e, 0x74, 0x59, 0x9c, 0xd5,
+		0x0c, 0xf0, 0x33, 0x97, 0x79, 0x3d, 0x3e, 0xd0, 0x95, 0xa6
+	};
+	size_t i;
+
+	/* Generate MAX_DATA_LEN bytes of data. */
+	for (i = 0; i < MAX_DATA_LEN; i++)
+		data[i] = i;
+
+	blake2s_init(&main_ctx, sizeof(hash));
+	for (i = 0; i <= MAX_DATA_LEN; i++) {
+		struct blake2s_ctx ctx;
+
+		/* Compute the BLAKE2s hash of 'i' data bytes. */
+		blake2s_init(&ctx, HASH_LEN);
+		blake2s_update(&ctx, data, i);
+		blake2s_final(&ctx, hash);
+
+		/* Verify that multiple updates produce the same result. */
+		blake2s_init(&ctx, HASH_LEN);
+		blake2s_update(&ctx, data, i / 2);
+		blake2s_update(&ctx, &data[i / 2], i - (i / 2));
+		blake2s_final(&ctx, hash2);
+		TEST_ASSERT_VAL("inconsistent BLAKE2s hashes",
+				memcmp(hash, hash2, HASH_LEN) == 0);
+
+		/*
+		 * Pass the hash to another BLAKE2s context, so that we
+		 * incrementally compute the hash of all the hashes.
+		 */
+		blake2s_update(&main_ctx, hash, HASH_LEN);
+	}
+
+	/* Verify the hash of all the hashes. */
+	blake2s_final(&main_ctx, hash);
+	TEST_ASSERT_VAL("wrong BLAKE2s hashes",
+			memcmp(hash, expected_hash_of_hashes, HASH_LEN) == 0);
+	return 0;
+}
+
+static int test__blake2s_case(struct test_suite *t __maybe_unused, int subtest __maybe_unused)
+{
+	return test_blake2s();
+}
+
+static int test__strreplace(struct test_suite *t __maybe_unused, int subtest __maybe_unused)
+{
+	TEST_ASSERT_VAL("empty string", test_strreplace(' ', "", "123", ""));
+	TEST_ASSERT_VAL("no match", test_strreplace('5', "123", "4", "123"));
+	TEST_ASSERT_VAL("replace 1", test_strreplace('3', "123", "4", "124"));
+	TEST_ASSERT_VAL("replace 2", test_strreplace('a', "abcabc", "ef", "efbcefbc"));
+	TEST_ASSERT_VAL("replace long", test_strreplace('a', "abcabc", "longlong",
+							"longlongbclonglongbc"));
+
+	return 0;
+}
+
+static struct test_case tests__util[] = {
+	TEST_CASE("String replacement", strreplace),
+	TEST_CASE("BLAKE2s hash", blake2s_case),
+	{ .name = NULL, }
+};
+
+struct test_suite suite__util = {
+	.desc = "util",
+	.test_cases = tests__util,
+};