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, 239 insertions, 0 deletions

diff --git a/lib/error-inject.c b/lib/error-inject.c
new file mode 100644
index 000000000..32f3d1ca9
--- /dev/null
+++ b/lib/error-inject.c
@@ -0,0 +1,239 @@
+// SPDX-License-Identifier: GPL-2.0
+// error-inject.c: Function-level error injection table
+#include <linux/error-injection.h>
+#include <linux/debugfs.h>
+#include <linux/kallsyms.h>
+#include <linux/kprobes.h>
+#include <linux/module.h>
+#include <linux/mutex.h>
+#include <linux/list.h>
+#include <linux/slab.h>
+#include <asm/sections.h>
+
+/* Whitelist of symbols that can be overridden for error injection. */
+static LIST_HEAD(error_injection_list);
+static DEFINE_MUTEX(ei_mutex);
+struct ei_entry {
+	struct list_head list;
+	unsigned long start_addr;
+	unsigned long end_addr;
+	int etype;
+	void *priv;
+};
+
+bool within_error_injection_list(unsigned long addr)
+{
+	struct ei_entry *ent;
+	bool ret = false;
+
+	mutex_lock(&ei_mutex);
+	list_for_each_entry(ent, &error_injection_list, list) {
+		if (addr >= ent->start_addr && addr < ent->end_addr) {
+			ret = true;
+			break;
+		}
+	}
+	mutex_unlock(&ei_mutex);
+	return ret;
+}
+
+int get_injectable_error_type(unsigned long addr)
+{
+	struct ei_entry *ent;
+	int ei_type = -EINVAL;
+
+	mutex_lock(&ei_mutex);
+	list_for_each_entry(ent, &error_injection_list, list) {
+		if (addr >= ent->start_addr && addr < ent->end_addr) {
+			ei_type = ent->etype;
+			break;
+		}
+	}
+	mutex_unlock(&ei_mutex);
+
+	return ei_type;
+}
+
+/*
+ * Lookup and populate the error_injection_list.
+ *
+ * For safety reasons we only allow certain functions to be overridden with
+ * bpf_error_injection, so we need to populate the list of the symbols that have
+ * been marked as safe for overriding.
+ */
+static void populate_error_injection_list(struct error_injection_entry *start,
+					  struct error_injection_entry *end,
+					  void *priv)
+{
+	struct error_injection_entry *iter;
+	struct ei_entry *ent;
+	unsigned long entry, offset = 0, size = 0;
+
+	mutex_lock(&ei_mutex);
+	for (iter = start; iter < end; iter++) {
+		entry = (unsigned long)dereference_symbol_descriptor((void *)iter->addr);
+
+		if (!kernel_text_address(entry) ||
+		    !kallsyms_lookup_size_offset(entry, &size, &offset)) {
+			pr_err("Failed to find error inject entry at %p\n",
+				(void *)entry);
+			continue;
+		}
+
+		ent = kmalloc_obj(*ent);
+		if (!ent)
+			break;
+		ent->start_addr = entry;
+		ent->end_addr = entry + size;
+		ent->etype = iter->etype;
+		ent->priv = priv;
+		INIT_LIST_HEAD(&ent->list);
+		list_add_tail(&ent->list, &error_injection_list);
+	}
+	mutex_unlock(&ei_mutex);
+}
+
+/* Markers of the _error_inject_whitelist section */
+extern struct error_injection_entry __start_error_injection_whitelist[];
+extern struct error_injection_entry __stop_error_injection_whitelist[];
+
+static void __init populate_kernel_ei_list(void)
+{
+	populate_error_injection_list(__start_error_injection_whitelist,
+				      __stop_error_injection_whitelist,
+				      NULL);
+}
+
+#ifdef CONFIG_MODULES
+static void module_load_ei_list(struct module *mod)
+{
+	if (!mod->num_ei_funcs)
+		return;
+
+	populate_error_injection_list(mod->ei_funcs,
+				      mod->ei_funcs + mod->num_ei_funcs, mod);
+}
+
+static void module_unload_ei_list(struct module *mod)
+{
+	struct ei_entry *ent, *n;
+
+	if (!mod->num_ei_funcs)
+		return;
+
+	mutex_lock(&ei_mutex);
+	list_for_each_entry_safe(ent, n, &error_injection_list, list) {
+		if (ent->priv == mod) {
+			list_del_init(&ent->list);
+			kfree(ent);
+		}
+	}
+	mutex_unlock(&ei_mutex);
+}
+
+/* Module notifier call back, checking error injection table on the module */
+static int ei_module_callback(struct notifier_block *nb,
+			      unsigned long val, void *data)
+{
+	struct module *mod = data;
+
+	if (val == MODULE_STATE_COMING)
+		module_load_ei_list(mod);
+	else if (val == MODULE_STATE_GOING)
+		module_unload_ei_list(mod);
+
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block ei_module_nb = {
+	.notifier_call = ei_module_callback,
+	.priority = 0
+};
+
+static __init int module_ei_init(void)
+{
+	return register_module_notifier(&ei_module_nb);
+}
+#else /* !CONFIG_MODULES */
+#define module_ei_init()	(0)
+#endif
+
+/*
+ * error_injection/whitelist -- shows which functions can be overridden for
+ * error injection.
+ */
+static void *ei_seq_start(struct seq_file *m, loff_t *pos)
+{
+	mutex_lock(&ei_mutex);
+	return seq_list_start(&error_injection_list, *pos);
+}
+
+static void ei_seq_stop(struct seq_file *m, void *v)
+{
+	mutex_unlock(&ei_mutex);
+}
+
+static void *ei_seq_next(struct seq_file *m, void *v, loff_t *pos)
+{
+	return seq_list_next(v, &error_injection_list, pos);
+}
+
+static const char *error_type_string(int etype)
+{
+	switch (etype) {
+	case EI_ETYPE_NULL:
+		return "NULL";
+	case EI_ETYPE_ERRNO:
+		return "ERRNO";
+	case EI_ETYPE_ERRNO_NULL:
+		return "ERRNO_NULL";
+	case EI_ETYPE_TRUE:
+		return "TRUE";
+	default:
+		return "(unknown)";
+	}
+}
+
+static int ei_seq_show(struct seq_file *m, void *v)
+{
+	struct ei_entry *ent = list_entry(v, struct ei_entry, list);
+
+	seq_printf(m, "%ps\t%s\n", (void *)ent->start_addr,
+		   error_type_string(ent->etype));
+	return 0;
+}
+
+static const struct seq_operations ei_sops = {
+	.start = ei_seq_start,
+	.next  = ei_seq_next,
+	.stop  = ei_seq_stop,
+	.show  = ei_seq_show,
+};
+
+DEFINE_SEQ_ATTRIBUTE(ei);
+
+static int __init ei_debugfs_init(void)
+{
+	struct dentry *dir, *file;
+
+	dir = debugfs_create_dir("error_injection", NULL);
+
+	file = debugfs_create_file("list", 0444, dir, NULL, &ei_fops);
+	if (IS_ERR(file)) {
+		debugfs_remove(dir);
+		return PTR_ERR(file);
+	}
+
+	return 0;
+}
+
+static int __init init_error_injection(void)
+{
+	populate_kernel_ei_list();
+
+	if (!module_ei_init())
+		ei_debugfs_init();
+
+	return 0;
+}
+late_initcall(init_error_injection);