Merge branch 'for-3.9-async' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq

Pull async changes from Tejun Heo:
 "These are followups for the earlier deadlock issue involving async
  ending up waiting for itself through block requesting module[1].  The
  following changes are made by these commits.

   - Instead of requesting default elevator on each request_queue init,
     block now requests it once early during boot.

   - Kmod triggers warning if invoked from an async worker.

   - Async synchronization implementation has been reimplemented.  It's
     a lot simpler now."

* 'for-3.9-async' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
  async: initialise list heads to fix crash
  async: replace list of active domains with global list of pending items
  async: keep pending tasks on async_domain and remove async_pending
  async: use ULLONG_MAX for infinity cookie value
  async: bring sanity to the use of words domain and running
  async, kmod: warn on synchronous request_module() from async workers
  block: don't request module during elevator init
  init, block: try to load default elevator module early during boot

9 files changed, 130 insertions, 109 deletions

diff --git a/block/elevator.c b/block/elevator.c
index 9edba1b..603b2c1 100644
--- a/block/elevator.c
+++ b/block/elevator.c
@@ -100,14 +100,14 @@ static void elevator_put(struct elevator_type *e)
 	module_put(e->elevator_owner);
 }
 
-static struct elevator_type *elevator_get(const char *name)
+static struct elevator_type *elevator_get(const char *name, bool try_loading)
 {
 	struct elevator_type *e;
 
 	spin_lock(&elv_list_lock);
 
 	e = elevator_find(name);
-	if (!e) {
+	if (!e && try_loading) {
 		spin_unlock(&elv_list_lock);
 		request_module("%s-iosched", name);
 		spin_lock(&elv_list_lock);
@@ -136,6 +136,22 @@ static int __init elevator_setup(char *str)
 
 __setup("elevator=", elevator_setup);
 
+/* called during boot to load the elevator chosen by the elevator param */
+void __init load_default_elevator_module(void)
+{
+	struct elevator_type *e;
+
+	if (!chosen_elevator[0])
+		return;
+
+	spin_lock(&elv_list_lock);
+	e = elevator_find(chosen_elevator);
+	spin_unlock(&elv_list_lock);
+
+	if (!e)
+		request_module("%s-iosched", chosen_elevator);
+}
+
 static struct kobj_type elv_ktype;
 
 static struct elevator_queue *elevator_alloc(struct request_queue *q,
@@ -191,25 +207,30 @@ int elevator_init(struct request_queue *q, char *name)
 	q->boundary_rq = NULL;
 
 	if (name) {
-		e = elevator_get(name);
+		e = elevator_get(name, true);
 		if (!e)
 			return -EINVAL;
 	}
 
+	/*
+	 * Use the default elevator specified by config boot param or
+	 * config option.  Don't try to load modules as we could be running
+	 * off async and request_module() isn't allowed from async.
+	 */
 	if (!e && *chosen_elevator) {
-		e = elevator_get(chosen_elevator);
+		e = elevator_get(chosen_elevator, false);
 		if (!e)
 			printk(KERN_ERR "I/O scheduler %s not found\n",
 							chosen_elevator);
 	}
 
 	if (!e) {
-		e = elevator_get(CONFIG_DEFAULT_IOSCHED);
+		e = elevator_get(CONFIG_DEFAULT_IOSCHED, false);
 		if (!e) {
 			printk(KERN_ERR
 				"Default I/O scheduler not found. " \
 				"Using noop.\n");
-			e = elevator_get("noop");
+			e = elevator_get("noop", false);
 		}
 	}
 
@@ -951,7 +972,7 @@ int elevator_change(struct request_queue *q, const char *name)
 		return -ENXIO;
 
 	strlcpy(elevator_name, name, sizeof(elevator_name));
-	e = elevator_get(strstrip(elevator_name));
+	e = elevator_get(strstrip(elevator_name), true);
 	if (!e) {
 		printk(KERN_ERR "elevator: type %s not found\n", elevator_name);
 		return -EINVAL;
diff --git a/include/linux/async.h b/include/linux/async.h
index 345169c..a2e3f18 100644
--- a/include/linux/async.h
+++ b/include/linux/async.h
@@ -19,8 +19,7 @@ typedef u64 async_cookie_t;
 typedef void (async_func_ptr) (void *data, async_cookie_t cookie);
 struct async_domain {
 	struct list_head node;
-	struct list_head domain;
-	int count;
+	struct list_head pending;
 	unsigned registered:1;
 };
 
@@ -29,8 +28,7 @@ struct async_domain {
  */
 #define ASYNC_DOMAIN(_name) \
 	struct async_domain _name = { .node = LIST_HEAD_INIT(_name.node), \
-				      .domain = LIST_HEAD_INIT(_name.domain), \
-				      .count = 0, \
+				      .pending = LIST_HEAD_INIT(_name.pending), \
 				      .registered = 1 }
 
 /*
@@ -39,8 +37,7 @@ struct async_domain {
  */
 #define ASYNC_DOMAIN_EXCLUSIVE(_name) \
 	struct async_domain _name = { .node = LIST_HEAD_INIT(_name.node), \
-				      .domain = LIST_HEAD_INIT(_name.domain), \
-				      .count = 0, \
+				      .pending = LIST_HEAD_INIT(_name.pending), \
 				      .registered = 0 }
 
 extern async_cookie_t async_schedule(async_func_ptr *ptr, void *data);
diff --git a/include/linux/elevator.h b/include/linux/elevator.h
index c03af76..1866206 100644
--- a/include/linux/elevator.h
+++ b/include/linux/elevator.h
@@ -138,6 +138,7 @@ extern void elv_drain_elevator(struct request_queue *);
 /*
  * io scheduler registration
  */
+extern void __init load_default_elevator_module(void);
 extern int elv_register(struct elevator_type *);
 extern void elv_unregister(struct elevator_type *);
 
@@ -206,5 +207,9 @@ enum {
 	INIT_LIST_HEAD(&(rq)->csd.list);	\
 	} while (0)
 
+#else /* CONFIG_BLOCK */
+
+static inline void load_default_elevator_module(void) { }
+
 #endif /* CONFIG_BLOCK */
 #endif
diff --git a/include/linux/init.h b/include/linux/init.h
index 10ed4f4..8618147 100644
--- a/include/linux/init.h
+++ b/include/linux/init.h
@@ -153,6 +153,7 @@ extern unsigned int reset_devices;
 /* used by init/main.c */
 void setup_arch(char **);
 void prepare_namespace(void);
+void __init load_default_modules(void);
 
 extern void (*late_time_init)(void);
 
diff --git a/init/do_mounts_initrd.c b/init/do_mounts_initrd.c
index f9acf71..a32ec1c 100644
--- a/init/do_mounts_initrd.c
+++ b/init/do_mounts_initrd.c
@@ -61,6 +61,9 @@ static void __init handle_initrd(void)
 	sys_mkdir("/old", 0700);
 	sys_chdir("/old");
 
+	/* try loading default modules from initrd */
+	load_default_modules();
+
 	/*
 	 * In case that a resume from disk is carried out by linuxrc or one of
 	 * its children, we need to tell the freezer not to wait for us.
diff --git a/init/initramfs.c b/init/initramfs.c
index 84c6bf1..a67ef9d 100644
--- a/init/initramfs.c
+++ b/init/initramfs.c
@@ -592,7 +592,7 @@ static int __init populate_rootfs(void)
 			initrd_end - initrd_start);
 		if (!err) {
 			free_initrd();
-			return 0;
+			goto done;
 		} else {
 			clean_rootfs();
 			unpack_to_rootfs(__initramfs_start, __initramfs_size);
@@ -607,6 +607,7 @@ static int __init populate_rootfs(void)
 			sys_close(fd);
 			free_initrd();
 		}
+	done:
 #else
 		printk(KERN_INFO "Unpacking initramfs...\n");
 		err = unpack_to_rootfs((char *)initrd_start,
@@ -615,6 +616,11 @@ static int __init populate_rootfs(void)
 			printk(KERN_EMERG "Initramfs unpacking failed: %s\n", err);
 		free_initrd();
 #endif
+		/*
+		 * Try loading default modules from initramfs.  This gives
+		 * us a chance to load before device_initcalls.
+		 */
+		load_default_modules();
 	}
 	return 0;
 }
diff --git a/init/main.c b/init/main.c
index cee4b5c..63534a1 100644
--- a/init/main.c
+++ b/init/main.c
@@ -70,6 +70,8 @@
 #include <linux/perf_event.h>
 #include <linux/file.h>
 #include <linux/ptrace.h>
+#include <linux/blkdev.h>
+#include <linux/elevator.h>
 
 #include <asm/io.h>
 #include <asm/bugs.h>
@@ -794,6 +796,17 @@ static void __init do_pre_smp_initcalls(void)
 		do_one_initcall(*fn);
 }
 
+/*
+ * This function requests modules which should be loaded by default and is
+ * called twice right after initrd is mounted and right before init is
+ * exec'd.  If such modules are on either initrd or rootfs, they will be
+ * loaded before control is passed to userland.
+ */
+void __init load_default_modules(void)
+{
+	load_default_elevator_module();
+}
+
 static int run_init_process(const char *init_filename)
 {
 	argv_init[0] = init_filename;
@@ -900,4 +913,7 @@ static noinline void __init kernel_init_freeable(void)
 	 * we're essentially up and running. Get rid of the
 	 * initmem segments and start the user-mode stuff..
 	 */
+
+	/* rootfs is available now, try loading default modules */
+	load_default_modules();
 }
diff --git a/kernel/async.c b/kernel/async.c
index 6c68fc3..8ddee2c 100644
--- a/kernel/async.c
+++ b/kernel/async.c
@@ -61,63 +61,48 @@ asynchronous and synchronous parts of the kernel.
 
 static async_cookie_t next_cookie = 1;
 
-#define MAX_WORK	32768
+#define MAX_WORK		32768
+#define ASYNC_COOKIE_MAX	ULLONG_MAX	/* infinity cookie */
 
-static LIST_HEAD(async_pending);
-static ASYNC_DOMAIN(async_running);
-static LIST_HEAD(async_domains);
+static LIST_HEAD(async_global_pending);	/* pending from all registered doms */
+static ASYNC_DOMAIN(async_dfl_domain);
 static DEFINE_SPINLOCK(async_lock);
-static DEFINE_MUTEX(async_register_mutex);
 
 struct async_entry {
-	struct list_head	list;
+	struct list_head	domain_list;
+	struct list_head	global_list;
 	struct work_struct	work;
 	async_cookie_t		cookie;
 	async_func_ptr		*func;
 	void			*data;
-	struct async_domain	*running;
+	struct async_domain	*domain;
 };
 
 static DECLARE_WAIT_QUEUE_HEAD(async_done);
 
 static atomic_t entry_count;
 
-
-/*
- * MUST be called with the lock held!
- */
-static async_cookie_t  __lowest_in_progress(struct async_domain *running)
+static async_cookie_t lowest_in_progress(struct async_domain *domain)
 {
-	async_cookie_t first_running = next_cookie;	/* infinity value */
-	async_cookie_t first_pending = next_cookie;	/* ditto */
-	struct async_entry *entry;
+	struct async_entry *first = NULL;
+	async_cookie_t ret = ASYNC_COOKIE_MAX;
+	unsigned long flags;
 
-	/*
-	 * Both running and pending lists are sorted but not disjoint.
-	 * Take the first cookies from both and return the min.
-	 */
-	if (!list_empty(&running->domain)) {
-		entry = list_first_entry(&running->domain, typeof(*entry), list);
-		first_running = entry->cookie;
-	}
+	spin_lock_irqsave(&async_lock, flags);
 
-	list_for_each_entry(entry, &async_pending, list) {
-		if (entry->running == running) {
-			first_pending = entry->cookie;
-			break;
-		}
+	if (domain) {
+		if (!list_empty(&domain->pending))
+			first = list_first_entry(&domain->pending,
+					struct async_entry, domain_list);
+	} else {
+		if (!list_empty(&async_global_pending))
+			first = list_first_entry(&async_global_pending,
+					struct async_entry, global_list);
 	}
 
-	return min(first_running, first_pending);
-}
+	if (first)
+		ret = first->cookie;
 
-static async_cookie_t  lowest_in_progress(struct async_domain *running)
-{
-	unsigned long flags;
-	async_cookie_t ret;
-
-	spin_lock_irqsave(&async_lock, flags);
-	ret = __lowest_in_progress(running);
 	spin_unlock_irqrestore(&async_lock, flags);
 	return ret;
 }
@@ -129,20 +114,10 @@ static void async_run_entry_fn(struct work_struct *work)
 {
 	struct async_entry *entry =
 		container_of(work, struct async_entry, work);
-	struct async_entry *pos;
 	unsigned long flags;
 	ktime_t uninitialized_var(calltime), delta, rettime;
-	struct async_domain *running = entry->running;
-
-	/* 1) move self to the running queue, make sure it stays sorted */
-	spin_lock_irqsave(&async_lock, flags);
-	list_for_each_entry_reverse(pos, &running->domain, list)
-		if (entry->cookie < pos->cookie)
-			break;
-	list_move_tail(&entry->list, &pos->list);
-	spin_unlock_irqrestore(&async_lock, flags);
 
-	/* 2) run (and print duration) */
+	/* 1) run (and print duration) */
 	if (initcall_debug && system_state == SYSTEM_BOOTING) {
 		printk(KERN_DEBUG "calling  %lli_%pF @ %i\n",
 			(long long)entry->cookie,
@@ -159,23 +134,22 @@ static void async_run_entry_fn(struct work_struct *work)
 			(long long)ktime_to_ns(delta) >> 10);
 	}
 
-	/* 3) remove self from the running queue */
+	/* 2) remove self from the pending queues */
 	spin_lock_irqsave(&async_lock, flags);
-	list_del(&entry->list);
-	if (running->registered && --running->count == 0)
-		list_del_init(&running->node);
+	list_del_init(&entry->domain_list);
+	list_del_init(&entry->global_list);
 
-	/* 4) free the entry */
+	/* 3) free the entry */
 	kfree(entry);
 	atomic_dec(&entry_count);
 
 	spin_unlock_irqrestore(&async_lock, flags);
 
-	/* 5) wake up any waiters */
+	/* 4) wake up any waiters */
 	wake_up(&async_done);
 }
 
-static async_cookie_t __async_schedule(async_func_ptr *ptr, void *data, struct async_domain *running)
+static async_cookie_t __async_schedule(async_func_ptr *ptr, void *data, struct async_domain *domain)
 {
 	struct async_entry *entry;
 	unsigned long flags;
@@ -198,16 +172,22 @@ static async_cookie_t __async_schedule(async_func_ptr *ptr, void *data, struct a
 		ptr(data, newcookie);
 		return newcookie;
 	}
+	INIT_LIST_HEAD(&entry->domain_list);
+	INIT_LIST_HEAD(&entry->global_list);
 	INIT_WORK(&entry->work, async_run_entry_fn);
 	entry->func = ptr;
 	entry->data = data;
-	entry->running = running;
+	entry->domain = domain;
 
 	spin_lock_irqsave(&async_lock, flags);
+
+	/* allocate cookie and queue */
 	newcookie = entry->cookie = next_cookie++;
-	list_add_tail(&entry->list, &async_pending);
-	if (running->registered && running->count++ == 0)
-		list_add_tail(&running->node, &async_domains);
+
+	list_add_tail(&entry->domain_list, &domain->pending);
+	if (domain->registered)
+		list_add_tail(&entry->global_list, &async_global_pending);
+
 	atomic_inc(&entry_count);
 	spin_unlock_irqrestore(&async_lock, flags);
 
@@ -230,7 +210,7 @@ static async_cookie_t __async_schedule(async_func_ptr *ptr, void *data, struct a
  */
 async_cookie_t async_schedule(async_func_ptr *ptr, void *data)
 {
-	return __async_schedule(ptr, data, &async_running);
+	return __async_schedule(ptr, data, &async_dfl_domain);
 }
 EXPORT_SYMBOL_GPL(async_schedule);
 
@@ -238,18 +218,18 @@ EXPORT_SYMBOL_GPL(async_schedule);
  * async_schedule_domain - schedule a function for asynchronous execution within a certain domain
  * @ptr: function to execute asynchronously
  * @data: data pointer to pass to the function
- * @running: running list for the domain
+ * @domain: the domain
  *
  * Returns an async_cookie_t that may be used for checkpointing later.
- * @running may be used in the async_synchronize_*_domain() functions
- * to wait within a certain synchronization domain rather than globally.
- * A synchronization domain is specified via the running queue @running to use.
- * Note: This function may be called from atomic or non-atomic contexts.
+ * @domain may be used in the async_synchronize_*_domain() functions to
+ * wait within a certain synchronization domain rather than globally.  A
+ * synchronization domain is specified via @domain.  Note: This function
+ * may be called from atomic or non-atomic contexts.
  */
 async_cookie_t async_schedule_domain(async_func_ptr *ptr, void *data,
-				     struct async_domain *running)
+				     struct async_domain *domain)
 {
-	return __async_schedule(ptr, data, running);
+	return __async_schedule(ptr, data, domain);
 }
 EXPORT_SYMBOL_GPL(async_schedule_domain);
 
@@ -260,18 +240,7 @@ EXPORT_SYMBOL_GPL(async_schedule_domain);
  */
 void async_synchronize_full(void)
 {
-	mutex_lock(&async_register_mutex);
-	do {
-		struct async_domain *domain = NULL;
-
-		spin_lock_irq(&async_lock);
-		if (!list_empty(&async_domains))
-			domain = list_first_entry(&async_domains, typeof(*domain), node);
-		spin_unlock_irq(&async_lock);
-
-		async_synchronize_cookie_domain(next_cookie, domain);
-	} while (!list_empty(&async_domains));
-	mutex_unlock(&async_register_mutex);
+	async_synchronize_full_domain(NULL);
 }
 EXPORT_SYMBOL_GPL(async_synchronize_full);
 
@@ -286,51 +255,45 @@ EXPORT_SYMBOL_GPL(async_synchronize_full);
  */
 void async_unregister_domain(struct async_domain *domain)
 {
-	mutex_lock(&async_register_mutex);
 	spin_lock_irq(&async_lock);
-	WARN_ON(!domain->registered || !list_empty(&domain->node) ||
-		!list_empty(&domain->domain));
+	WARN_ON(!domain->registered || !list_empty(&domain->pending));
 	domain->registered = 0;
 	spin_unlock_irq(&async_lock);
-	mutex_unlock(&async_register_mutex);
 }
 EXPORT_SYMBOL_GPL(async_unregister_domain);
 
 /**
  * async_synchronize_full_domain - synchronize all asynchronous function within a certain domain
- * @domain: running list to synchronize on
+ * @domain: the domain to synchronize
  *
  * This function waits until all asynchronous function calls for the
- * synchronization domain specified by the running list @domain have been done.
+ * synchronization domain specified by @domain have been done.
  */
 void async_synchronize_full_domain(struct async_domain *domain)
 {
-	async_synchronize_cookie_domain(next_cookie, domain);
+	async_synchronize_cookie_domain(ASYNC_COOKIE_MAX, domain);
 }
 EXPORT_SYMBOL_GPL(async_synchronize_full_domain);
 
 /**
  * async_synchronize_cookie_domain - synchronize asynchronous function calls within a certain domain with cookie checkpointing
  * @cookie: async_cookie_t to use as checkpoint
- * @running: running list to synchronize on
+ * @domain: the domain to synchronize (%NULL for all registered domains)
  *
  * This function waits until all asynchronous function calls for the
- * synchronization domain specified by running list @running submitted
- * prior to @cookie have been done.
+ * synchronization domain specified by @domain submitted prior to @cookie
+ * have been done.
  */
-void async_synchronize_cookie_domain(async_cookie_t cookie, struct async_domain *running)
+void async_synchronize_cookie_domain(async_cookie_t cookie, struct async_domain *domain)
 {
 	ktime_t uninitialized_var(starttime), delta, endtime;
 
-	if (!running)
-		return;
-
 	if (initcall_debug && system_state == SYSTEM_BOOTING) {
 		printk(KERN_DEBUG "async_waiting @ %i\n", task_pid_nr(current));
 		starttime = ktime_get();
 	}
 
-	wait_event(async_done, lowest_in_progress(running) >= cookie);
+	wait_event(async_done, lowest_in_progress(domain) >= cookie);
 
 	if (initcall_debug && system_state == SYSTEM_BOOTING) {
 		endtime = ktime_get();
@@ -352,7 +315,7 @@ EXPORT_SYMBOL_GPL(async_synchronize_cookie_domain);
  */
 void async_synchronize_cookie(async_cookie_t cookie)
 {
-	async_synchronize_cookie_domain(cookie, &async_running);
+	async_synchronize_cookie_domain(cookie, &async_dfl_domain);
 }
 EXPORT_SYMBOL_GPL(async_synchronize_cookie);
 
diff --git a/kernel/kmod.c b/kernel/kmod.c
index 0023a87..56dd349 100644
--- a/kernel/kmod.c
+++ b/kernel/kmod.c
@@ -38,6 +38,7 @@
 #include <linux/suspend.h>
 #include <linux/rwsem.h>
 #include <linux/ptrace.h>
+#include <linux/async.h>
 #include <asm/uaccess.h>
 
 #include <trace/events/module.h>
@@ -130,6 +131,14 @@ int __request_module(bool wait, const char *fmt, ...)
 #define MAX_KMOD_CONCURRENT 50	/* Completely arbitrary value - KAO */
 	static int kmod_loop_msg;
 
+	/*
+	 * We don't allow synchronous module loading from async.  Module
+	 * init may invoke async_synchronize_full() which will end up
+	 * waiting for this task which already is waiting for the module
+	 * loading to complete, leading to a deadlock.
+	 */
+	WARN_ON_ONCE(wait && current_is_async());
+
 	va_start(args, fmt);
 	ret = vsnprintf(module_name, MODULE_NAME_LEN, fmt, args);
 	va_end(args);