module, async: async_synchronize_full() on module init iff async is used

If the default iosched is built as module, the kernel may deadlock
while trying to load the iosched module on device probe if the probing
was running off async.  This is because async_synchronize_full() at
the end of module init ends up waiting for the async job which
initiated the module loading.

 async A				modprobe

 1. finds a device
 2. registers the block device
 3. request_module(default iosched)
					4. modprobe in userland
					5. load and init module
					6. async_synchronize_full()

Async A waits for modprobe to finish in request_module() and modprobe
waits for async A to finish in async_synchronize_full().

Because there's no easy to track dependency once control goes out to
userland, implementing properly nested flushing is difficult.  For
now, make module init perform async_synchronize_full() iff module init
has queued async jobs as suggested by Linus.

This avoids the described deadlock because iosched module doesn't use
async and thus wouldn't invoke async_synchronize_full().  This is
hacky and incomplete.  It will deadlock if async module loading nests;
however, this works around the known problem case and seems to be the
best of bad options.

For more details, please refer to the following thread.

  http://thread.gmane.org/gmane.linux.kernel/1420814

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Alex Riesen <raa.lkml@gmail.com>
Tested-by: Ming Lei <ming.lei@canonical.com>
Tested-by: Alex Riesen <raa.lkml@gmail.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 206bb089c06b5541a889291f7938e80b3b480678..6fc8f45de4e912f1196d07a5ace4d4cc752ebe3f 100644 (file)
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1810,6 +1810,7 @@ extern void thread_group_cputime_adjusted(struct task_struct *p, cputime_t *ut,
 #define PF_MEMALLOC    0x00000800      /* Allocating memory */
 #define PF_NPROC_EXCEEDED 0x00001000   /* set_user noticed that RLIMIT_NPROC was exceeded */
 #define PF_USED_MATH   0x00002000      /* if unset the fpu must be initialized before use */
+#define PF_USED_ASYNC  0x00004000      /* used async_schedule*(), used by module init */
 #define PF_NOFREEZE    0x00008000      /* this thread should not be frozen */
 #define PF_FROZEN      0x00010000      /* frozen for system suspend */
 #define PF_FSTRANS     0x00020000      /* inside a filesystem transaction */

diff --git a/kernel/async.c b/kernel/async.c
index 9d31183848582ca8da812d70d2d68364ff898f1d..a1d585c351d663f3cc64f0f874b0a38562797dc7 100644 (file)
--- a/kernel/async.c
+++ b/kernel/async.c
@@ -196,6 +196,9 @@ static async_cookie_t __async_schedule(async_func_ptr *ptr, void *data, struct a
        atomic_inc(&entry_count);
        spin_unlock_irqrestore(&async_lock, flags);
 
+       /* mark that this task has queued an async job, used by module init */
+       current->flags |= PF_USED_ASYNC;
+
        /* schedule for execution */
        queue_work(system_unbound_wq, &entry->work);
 

diff --git a/kernel/module.c b/kernel/module.c
index 250092c1d57d6cf582ef801f70a10969e19abab2..b10b048367e14b0e89f6b7bfc5e740afd666516b 100644 (file)
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -3013,6 +3013,12 @@ static int do_init_module(struct module *mod)
 {
        int ret = 0;
 
+       /*
+        * We want to find out whether @mod uses async during init.  Clear
+        * PF_USED_ASYNC.  async_schedule*() will set it.
+        */
+       current->flags &= ~PF_USED_ASYNC;
+
        blocking_notifier_call_chain(&module_notify_list,
                        MODULE_STATE_COMING, mod);
 
@@ -3058,8 +3064,25 @@ static int do_init_module(struct module *mod)
        blocking_notifier_call_chain(&module_notify_list,
                                     MODULE_STATE_LIVE, mod);
 
-       /* We need to finish all async code before the module init sequence is done */
-       async_synchronize_full();
+       /*
+        * We need to finish all async code before the module init sequence
+        * is done.  This has potential to deadlock.  For example, a newly
+        * detected block device can trigger request_module() of the
+        * default iosched from async probing task.  Once userland helper
+        * reaches here, async_synchronize_full() will wait on the async
+        * task waiting on request_module() and deadlock.
+        *
+        * This deadlock is avoided by perfomring async_synchronize_full()
+        * iff module init queued any async jobs.  This isn't a full
+        * solution as it will deadlock the same if module loading from
+        * async jobs nests more than once; however, due to the various
+        * constraints, this hack seems to be the best option for now.
+        * Please refer to the following thread for details.
+        *
+        * http://thread.gmane.org/gmane.linux.kernel/1420814
+        */
+       if (current->flags & PF_USED_ASYNC)
+               async_synchronize_full();
 
        mutex_lock(&module_mutex);
        /* Drop initial reference. */