[Devel] [PATCH RHEL9 COMMIT] pid_ns: Virtualize pid_max

[Konstantin Khorenko]

The commit is pushed to "branch-rh9-5.14.vz9.1.x-ovz" and will appear at https://src.openvz.org/scm/ovz/vzkernel.git
after rh9-5.14.0-4.vz9.10.12
------>
commit 4bf315c05027fecd13657a9209cabe232d79b580
Author: Kirill Tkhai <ktkhai at virtuozzo.com>
Date:   Wed Oct 20 11:40:43 2021 +0300

    pid_ns: Virtualize pid_max
    
    Introduce per pid namespace pid_max.
    
    Extracted from "Initial patch".
    
    Signed-off-by: Kirill Tkhai <ktkhai at virtuozzo.com>
    
    (cherry-picked from vz7 commit 80da36469321 ("pid_ns: Virtualize pid_max"))
    
    https://jira.sw.ru/browse/PSBM-127857
    Signed-off-by: Valeriy Vdovin <valeriy.vdovin at virtuozzo.com>
    
    +++
    pid: Use proper ns in proc_dointvec_pidmax()
    
    Use current task pid ns instead of pid ns for future children
    https://jira.sw.ru/browse/PSBM-121530
    Signed-off-by: Kirill Tkhai <ktkhai at virtuozzo.com>
    
    (cherry-picked from vz7 commit d91a0dc98453 ("pid: Use proper ns in
    proc_dointvec_pidmax()"))
    
    https://jira.sw.ru/browse/PSBM-127857
    Signed-off-by: Valeriy Vdovin <valeriy.vdovin at virtuozzo.com>
    
    (cherry picked from vz8 commit c3299968d4753f7fdf13a792d04795605951c533)
    Signed-off-by: Andrey Zhadchenko <andrey.zhadchenko at virtuozzo.com>
---
 include/linux/pid_namespace.h |  1 +
 include/linux/threads.h       |  1 +
 kernel/pid.c                  |  2 +-
 kernel/pid_namespace.c        |  4 ++--
 kernel/sysctl.c               | 15 ++++++++++++---
 5 files changed, 17 insertions(+), 6 deletions(-)

diff --git a/include/linux/pid_namespace.h b/include/linux/pid_namespace.h
index 7c7e627503d2..f6b68fb1b190 100644
--- a/include/linux/pid_namespace.h
+++ b/include/linux/pid_namespace.h
@@ -20,6 +20,7 @@ struct pid_namespace {
 	struct idr idr;
 	struct rcu_head rcu;
 	unsigned int pid_allocated;
+	int pid_max;
 	struct task_struct *child_reaper;
 	struct kmem_cache *pid_cachep;
 	unsigned int level;
diff --git a/include/linux/threads.h b/include/linux/threads.h
index 18d5a74bcc3d..f7dd16fc5606 100644
--- a/include/linux/threads.h
+++ b/include/linux/threads.h
@@ -26,6 +26,7 @@
  * This controls the default maximum pid allocated to a process
  */
 #define PID_MAX_DEFAULT (CONFIG_BASE_SMALL ? 0x1000 : 0x8000)
+#define PID_MAX_NS_DEFAULT	(PID_MAX_DEFAULT)
 
 /*
  * A maximum of 4 million PIDs should be enough for a while.
diff --git a/kernel/pid.c b/kernel/pid.c
index 8d0b41e4f735..6b1a86863d5e 100644
--- a/kernel/pid.c
+++ b/kernel/pid.c
@@ -231,7 +231,7 @@ struct pid *alloc_pid(struct pid_namespace *ns, pid_t *set_tid,
 			 * a partially initialized PID (see below).
 			 */
 			nr = idr_alloc_cyclic(&tmp->idr, NULL, pid_min,
-					      pid_max, GFP_ATOMIC);
+					      ns->pid_max, GFP_ATOMIC);
 		}
 		spin_unlock_irq(&pidmap_lock);
 		idr_preload_end();
diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c
index 5c6555bbb300..450d613ba8d5 100644
--- a/kernel/pid_namespace.c
+++ b/kernel/pid_namespace.c
@@ -110,6 +110,7 @@ static struct pid_namespace *create_pid_namespace(struct user_namespace *user_ns
 	ns->user_ns = get_user_ns(user_ns);
 	ns->ucounts = ucounts;
 	ns->pid_allocated = PIDNS_ADDING;
+	ns->pid_max = PID_MAX_NS_DEFAULT;
 
 	return ns;
 
@@ -281,6 +282,7 @@ static int pid_ns_ctl_handler(struct ctl_table *table, int write,
 	next = idr_get_cursor(&pid_ns->idr) - 1;
 
 	tmp.data = &next;
+	tmp.extra2 = &pid_ns->pid_max;
 	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
 	if (!ret && write)
 		idr_set_cursor(&pid_ns->idr, next + 1);
@@ -288,7 +290,6 @@ static int pid_ns_ctl_handler(struct ctl_table *table, int write,
 	return ret;
 }
 
-extern int pid_max;
 static struct ctl_table pid_ns_ctl_table[] = {
 	{
 		.procname = "ns_last_pid",
@@ -296,7 +297,6 @@ static struct ctl_table pid_ns_ctl_table[] = {
 		.mode = 0666 | S_ISVTX, /* permissions are checked in the handler */
 		.proc_handler = pid_ns_ctl_handler,
 		.extra1 = SYSCTL_ZERO,
-		.extra2 = &pid_max,
 	},
 	{ }
 };
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
index 53090d656dec..9bb8935f9b55 100644
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -1808,6 +1808,16 @@ int proc_do_static_key(struct ctl_table *table, int write,
 	return ret;
 }
 
+static int proc_dointvec_pidmax(struct ctl_table *table, int write,
+		void __user *buffer, size_t *lenp, loff_t *ppos)
+{
+	struct ctl_table tmp;
+
+	tmp = *table;
+	tmp.data = &task_active_pid_ns(current)->pid_max;
+	return proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
+}
+
 #ifdef CONFIG_COREDUMP
 sysctl_virtual(proc_dostring_coredump);
 #endif
@@ -2290,10 +2300,9 @@ static struct ctl_table kern_table[] = {
 #endif /* CONFIG_SMP */
 	{
 		.procname	= "pid_max",
-		.data		= &pid_max,
 		.maxlen		= sizeof (int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec_minmax,
+		.mode		= 0644 | S_ISVTX,
+		.proc_handler	= proc_dointvec_pidmax,
 		.extra1		= &pid_max_min,
 		.extra2		= &pid_max_max,
 	},