[RFC PATCH v1 07/30] sysctl: use sysctl_field in pid sysctls

[Alexey Gladkov]

The pid namespace sysctl registration allocates a private ctl_table copy
only to replace the static pid_max data pointer with the value from the
registered namespace. This keeps otherwise shared sysctl metadata
writable at registration time and requires per-namespace table lifetime
management.

Use a sysctl_field accessor to derive pid_max from the registration
context instead. The pid sysctl table can remain static and const, while
registration no longer needs to clone, patch and free a ctl_table array.

Signed-off-by: Alexey Gladkov <legion@xxxxxxxxxx>
---
 include/linux/sysctl.h |  2 ++
 kernel/pid.c           | 41 +++++++++++++----------------------------
 2 files changed, 15 insertions(+), 28 deletions(-)

diff --git a/include/linux/sysctl.h b/include/linux/sysctl.h
index e2831190e484..056d10f4ab3d 100644
--- a/include/linux/sysctl.h
+++ b/include/linux/sysctl.h
@@ -39,6 +39,7 @@ struct ctl_table_root;
 struct ctl_table_header;
 struct ctl_dir;
 struct ipc_namespace;
+struct pid_namespace;
 struct user_namespace;
 
 /* Keep the same order as in fs/proc/proc_sysctl.c */
@@ -92,6 +93,7 @@ struct sysctl_context {
 	union {
 		struct user_namespace *user_ns;
 		struct ipc_namespace *ipc_ns;
+		struct pid_namespace *pid_ns;
 	} ns;
 };
 
diff --git a/kernel/pid.c b/kernel/pid.c
index fd5c2d4aa349..66991ee435fb 100644
--- a/kernel/pid.c
+++ b/kernel/pid.c
@@ -787,23 +787,15 @@ static int proc_do_cad_pid(const struct ctl_table *table, int write, void *buffe
 	return 0;
 }
 
-static const struct ctl_table pid_table[] = {
-	{
-		.procname	= "pid_max",
-		.data		= &init_pid_ns.pid_max,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec_minmax,
-		.extra1		= &pid_max_min,
-		.extra2		= &pid_max_max,
-	},
+static int *pid_max_data(const struct sysctl_context *ctx)
+{
+	return &ctx->ns.pid_ns->pid_max;
+}
+
+static const struct sysctl_field pid_table[] = {
+	SYSCTL_FIELD_STATIC_INT_MINMAX("pid_max", 0644, pid_max_data, &pid_max_min, &pid_max_max),
 #ifdef CONFIG_PROC_SYSCTL
-	{
-		.procname	= "cad_pid",
-		.maxlen		= sizeof(int),
-		.mode		= 0600,
-		.proc_handler	= proc_do_cad_pid,
-	},
+	SYSCTL_FIELD_CUSTOM("cad_pid", 0600, sizeof(int), NULL, proc_do_cad_pid),
 #endif
 };
 #endif
@@ -811,21 +803,18 @@ static const struct ctl_table pid_table[] = {
 int register_pidns_sysctls(struct pid_namespace *pidns)
 {
 #ifdef CONFIG_SYSCTL
-	struct ctl_table *tbl;
+	struct sysctl_context ctx = {
+		.ns.pid_ns = pidns,
+	};
 
 	setup_sysctl_set(&pidns->set, &pid_table_root, set_is_seen);
 
-	tbl = kmemdup(pid_table, sizeof(pid_table), GFP_KERNEL);
-	if (!tbl)
-		return -ENOMEM;
-	tbl->data = &pidns->pid_max;
 	pidns->pid_max = min(pid_max_max, max_t(int, pidns->pid_max,
 			     PIDS_PER_CPU_DEFAULT * num_possible_cpus()));
 
-	pidns->sysctls = __register_sysctl_table(&pidns->set, "kernel", tbl,
-						 ARRAY_SIZE(pid_table));
+	pidns->sysctls = register_sysctl_fields(&pidns->set, "kernel",
+						pid_table, &ctx);
 	if (!pidns->sysctls) {
-		kfree(tbl);
 		retire_sysctl_set(&pidns->set);
 		return -ENOMEM;
 	}
@@ -836,12 +825,8 @@ int register_pidns_sysctls(struct pid_namespace *pidns)
 void unregister_pidns_sysctls(struct pid_namespace *pidns)
 {
 #ifdef CONFIG_SYSCTL
-	const struct ctl_table *tbl;
-
-	tbl = pidns->sysctls->ctl_table_arg;
 	unregister_sysctl_table(pidns->sysctls);
 	retire_sysctl_set(&pidns->set);
-	kfree(tbl);
 #endif
 }
 
-- 
2.55.0