1 files changed, 38 insertions, 20 deletions

diff --git a/kernel/fork.c b/kernel/fork.c
index 89b8b6c08592..9f7fe3541897 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -756,8 +756,13 @@ static void check_mm(struct mm_struct *mm)
 			 "Please make sure 'struct resident_page_types[]' is updated as well");
 
 	for (i = 0; i < NR_MM_COUNTERS; i++) {
-		long x = atomic_long_read(&mm->rss_stat.count[i]);
+		long x = percpu_counter_sum(&mm->rss_stat[i]);
 
+		if (likely(!x))
+			continue;
+
+		/* Making sure this is not due to race with CPU offlining. */
+		x = percpu_counter_sum_all(&mm->rss_stat[i]);
 		if (unlikely(x))
 			pr_alert("BUG: Bad rss-counter state mm:%p type:%s val:%ld\n",
 				 mm, resident_page_types[i], x);
@@ -782,6 +787,8 @@ static void check_mm(struct mm_struct *mm)
  */
 void __mmdrop(struct mm_struct *mm)
 {
+	int i;
+
 	BUG_ON(mm == &init_mm);
 	WARN_ON_ONCE(mm == current->mm);
 	WARN_ON_ONCE(mm == current->active_mm);
@@ -791,6 +798,9 @@ void __mmdrop(struct mm_struct *mm)
 	check_mm(mm);
 	put_user_ns(mm->user_ns);
 	mm_pasid_drop(mm);
+
+	for (i = 0; i < NR_MM_COUNTERS; i++)
+		percpu_counter_destroy(&mm->rss_stat[i]);
 	free_mm(mm);
 }
 EXPORT_SYMBOL_GPL(__mmdrop);
@@ -1110,6 +1120,8 @@ static void mm_init_uprobes_state(struct mm_struct *mm)
 static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p,
 	struct user_namespace *user_ns)
 {
+	int i;
+
 	mt_init_flags(&mm->mm_mt, MM_MT_FLAGS);
 	mt_set_external_lock(&mm->mm_mt, &mm->mmap_lock);
 	atomic_set(&mm->mm_users, 1);
@@ -1151,10 +1163,17 @@ static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p,
 	if (init_new_context(p, mm))
 		goto fail_nocontext;
 
+	for (i = 0; i < NR_MM_COUNTERS; i++)
+		if (percpu_counter_init(&mm->rss_stat[i], 0, GFP_KERNEL_ACCOUNT))
+			goto fail_pcpu;
+
 	mm->user_ns = get_user_ns(user_ns);
 	lru_gen_init_mm(mm);
 	return mm;
 
+fail_pcpu:
+	while (i > 0)
+		percpu_counter_destroy(&mm->rss_stat[--i]);
 fail_nocontext:
 	mm_free_pgd(mm);
 fail_nopgd:
@@ -2588,11 +2607,6 @@ struct task_struct * __init fork_idle(int cpu)
 	return task;
 }
 
-struct mm_struct *copy_init_mm(void)
-{
-	return dup_mm(NULL, &init_mm);
-}
-
 /*
  * This is like kernel_clone(), but shaved down and tailored to just
  * creating io_uring workers. It returns a created task, or an error pointer.
@@ -3011,10 +3025,27 @@ static void sighand_ctor(void *data)
 	init_waitqueue_head(&sighand->signalfd_wqh);
 }
 
-void __init proc_caches_init(void)
+void __init mm_cache_init(void)
 {
 	unsigned int mm_size;
 
+	/*
+	 * The mm_cpumask is located at the end of mm_struct, and is
+	 * dynamically sized based on the maximum CPU number this system
+	 * can have, taking hotplug into account (nr_cpu_ids).
+	 */
+	mm_size = sizeof(struct mm_struct) + cpumask_size();
+
+	mm_cachep = kmem_cache_create_usercopy("mm_struct",
+			mm_size, ARCH_MIN_MMSTRUCT_ALIGN,
+			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_ACCOUNT,
+			offsetof(struct mm_struct, saved_auxv),
+			sizeof_field(struct mm_struct, saved_auxv),
+			NULL);
+}
+
+void __init proc_caches_init(void)
+{
 	sighand_cachep = kmem_cache_create("sighand_cache",
 			sizeof(struct sighand_struct), 0,
 			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_TYPESAFE_BY_RCU|
@@ -3032,19 +3063,6 @@ void __init proc_caches_init(void)
 			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_ACCOUNT,
 			NULL);
 
-	/*
-	 * The mm_cpumask is located at the end of mm_struct, and is
-	 * dynamically sized based on the maximum CPU number this system
-	 * can have, taking hotplug into account (nr_cpu_ids).
-	 */
-	mm_size = sizeof(struct mm_struct) + cpumask_size();
-
-	mm_cachep = kmem_cache_create_usercopy("mm_struct",
-			mm_size, ARCH_MIN_MMSTRUCT_ALIGN,
-			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_ACCOUNT,
-			offsetof(struct mm_struct, saved_auxv),
-			sizeof_field(struct mm_struct, saved_auxv),
-			NULL);
 	vm_area_cachep = KMEM_CACHE(vm_area_struct, SLAB_PANIC|SLAB_ACCOUNT);
 	mmap_init();
 	nsproxy_cache_init();