erofs: lazily initialize per-CPU workers and CPU hotplug hooks

Currently, when EROFS is built with per-CPU workers, the workers are
started and CPU hotplug hooks are registered during module initialization.
This leads to unnecessary worker start/stop cycles during CPU hotplug
events, particularly on Android devices that frequently suspend and resume.

This change defers the initialization of per-CPU workers and the
registration of CPU hotplug hooks until the first EROFS mount. This
ensures that these resources are only allocated and managed when EROFS is
actually in use.

The tear down of per-CPU workers and unregistration of CPU hotplug hooks
still occurs during z_erofs_exit_subsystem(), but only if they were
initialized.

Signed-off-by: Sandeep Dhavale <dhavale@google.com>
Reviewed-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Reviewed-by: Chao Yu <chao@kernel.org>
Link: https://lore.kernel.org/r/20250506225743.308517-1-dhavale@google.com
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>

1 files changed, 50 insertions, 20 deletions

diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index 29541e0787b87..b0f93b9be8ccf 100644
--- a/fs/erofs/zdata.c
+++ b/fs/erofs/zdata.c
@@ -288,6 +288,7 @@ static struct workqueue_struct *z_erofs_workqueue __read_mostly;
 
 #ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
 static struct kthread_worker __rcu **z_erofs_pcpu_workers;
+static atomic_t erofs_percpu_workers_initialized = ATOMIC_INIT(0);
 
 static void erofs_destroy_percpu_workers(void)
 {
@@ -333,12 +334,8 @@ static int erofs_init_percpu_workers(void)
 	}
 	return 0;
 }
-#else
-static inline void erofs_destroy_percpu_workers(void) {}
-static inline int erofs_init_percpu_workers(void) { return 0; }
-#endif
 
-#if defined(CONFIG_HOTPLUG_CPU) && defined(CONFIG_EROFS_FS_PCPU_KTHREAD)
+#ifdef CONFIG_HOTPLUG_CPU
 static DEFINE_SPINLOCK(z_erofs_pcpu_worker_lock);
 static enum cpuhp_state erofs_cpuhp_state;
 
@@ -395,15 +392,53 @@ static void erofs_cpu_hotplug_destroy(void)
 	if (erofs_cpuhp_state)
 		cpuhp_remove_state_nocalls(erofs_cpuhp_state);
 }
-#else /* !CONFIG_HOTPLUG_CPU || !CONFIG_EROFS_FS_PCPU_KTHREAD */
+#else /* !CONFIG_HOTPLUG_CPU  */
 static inline int erofs_cpu_hotplug_init(void) { return 0; }
 static inline void erofs_cpu_hotplug_destroy(void) {}
-#endif
+#endif/* CONFIG_HOTPLUG_CPU */
+static int z_erofs_init_pcpu_workers(struct super_block *sb)
+{
+	int err;
 
-void z_erofs_exit_subsystem(void)
+	if (atomic_xchg(&erofs_percpu_workers_initialized, 1))
+		return 0;
+
+	err = erofs_init_percpu_workers();
+	if (err) {
+		erofs_err(sb, "per-cpu workers: failed to allocate.");
+		goto err_init_percpu_workers;
+	}
+
+	err = erofs_cpu_hotplug_init();
+	if (err < 0) {
+		erofs_err(sb, "per-cpu workers: failed CPU hotplug init.");
+		goto err_cpuhp_init;
+	}
+	erofs_info(sb, "initialized per-cpu workers successfully.");
+	return err;
+
+err_cpuhp_init:
+	erofs_destroy_percpu_workers();
+err_init_percpu_workers:
+	atomic_set(&erofs_percpu_workers_initialized, 0);
+	return err;
+}
+
+static void z_erofs_destroy_pcpu_workers(void)
 {
+	if (!atomic_xchg(&erofs_percpu_workers_initialized, 0))
+		return;
 	erofs_cpu_hotplug_destroy();
 	erofs_destroy_percpu_workers();
+}
+#else /* !CONFIG_EROFS_FS_PCPU_KTHREAD */
+static inline int z_erofs_init_pcpu_workers(struct super_block *sb) { return 0; }
+static inline void z_erofs_destroy_pcpu_workers(void) {}
+#endif/* CONFIG_EROFS_FS_PCPU_KTHREAD */
+
+void z_erofs_exit_subsystem(void)
+{
+	z_erofs_destroy_pcpu_workers();
 	destroy_workqueue(z_erofs_workqueue);
 	z_erofs_destroy_pcluster_pool();
 	z_erofs_exit_decompressor();
@@ -427,19 +462,8 @@ int __init z_erofs_init_subsystem(void)
 		goto err_workqueue_init;
 	}
 
-	err = erofs_init_percpu_workers();
-	if (err)
-		goto err_pcpu_worker;
-
-	err = erofs_cpu_hotplug_init();
-	if (err < 0)
-		goto err_cpuhp_init;
 	return err;
 
-err_cpuhp_init:
-	erofs_destroy_percpu_workers();
-err_pcpu_worker:
-	destroy_workqueue(z_erofs_workqueue);
 err_workqueue_init:
 	z_erofs_destroy_pcluster_pool();
 err_pcluster_pool:
@@ -641,8 +665,14 @@ static const struct address_space_operations z_erofs_cache_aops = {
 
 int z_erofs_init_super(struct super_block *sb)
 {
-	struct inode *const inode = new_inode(sb);
+	struct inode *inode;
+	int err;
+
+	err = z_erofs_init_pcpu_workers(sb);
+	if (err)
+		return err;
 
+	inode = new_inode(sb);
 	if (!inode)
 		return -ENOMEM;
 	set_nlink(inode, 1);