On 05/25/2016 12:55 AM, [email protected] wrote:
+ spin_lock_irq(q->queue_lock);
+ wait_event_lock_irq(q->reset_wq,
+ !queue_flag_test_and_set(QUEUE_FLAG_RESETTING, q),
+ *q->queue_lock);
+ if (blk_queue_dead(q)) {
+ rc = -ENODEV;
+ spin_unlock_irq(q->queue_lock);
+ goto done;
+ }
+ spin_unlock_irq(q->queue_lock);
Is the flag QUEUE_FLAG_RESETTING used to serialize blk_reset_queue()
calls? If so, please consider using a mutex instead. Lockdep can detect
lock inversion for mutexes and spinlocks but not for serialization that
is based on wait_event_lock_irq() loops.
+ if (q->mq_ops) {
+ blk_mq_stop_hw_queues(q);
+ blk_mq_freeze_queue(q);
+
+ eh_rc = q->mq_ops->reset(q);
+
+ blk_mq_unfreeze_queue(q);
+ blk_mq_start_stopped_hw_queues(q, true);
+ } else if (q->reset_fn) {
+ spin_lock_irq(q->queue_lock);
+ blk_stop_queue(q);
+ spin_unlock_irq(q->queue_lock);
+
+ while (q->request_fn_active)
+ msleep(10);
+
+ eh_rc = q->reset_fn(q);
+
+ spin_lock_irq(q->queue_lock);
+ blk_start_queue(q);
+ spin_unlock_irq(q->queue_lock);
+ }
Some of the above code is similar to some of the code in
scsi_internal_device_block() and scsi_internal_device_unblock(). Have
you considered to avoid this duplication by introducing helpers in the
block layer for blocking and unblocking queues?
Thanks,
Bart.
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
