Hi Marcel,
Thanks for the review.
static int intel_recv_lpm(struct hci_dev *hdev, struct sk_buff *skb)
@@ -895,9 +936,24 @@ static int intel_recv(struct hci_uart *hu, const void *data, int count)
static int intel_enqueue(struct hci_uart *hu, struct sk_buff *skb)
{
struct intel_data *intel = hu->priv;
+ struct list_head *p;
BT_DBG("hu %p skb %p", hu, skb);
+ mutex_lock(&intel_device_list_lock);
+ list_for_each(p, &intel_device_list) {
+ struct intel_device *idev = list_entry(p, struct intel_device,
+ list);
+
+ if (hu->tty->dev->parent == idev->pdev->dev.parent) {
+ pm_runtime_get_sync(&idev->pdev->dev);
just for my own clarification. Why this is get_sync and the other just get?
I want to be sure resume is completed before enqueuing any packet in
order to be able to send the packet in dequeue.
Moreover, since LPM and 'standard' packets share the same queue, we
don't want to enqueue a 'standard' packet before our LPM one.
In the threaded IRQ , we are actually already LPM resumed since we have
send a LPM resume ack to the controller, I just use get and put to
notify the resume to the PM core and trigger autosuspend.
In the busy work, we don't need to be synced since we are already in a
async work. It is more a keep alive to delay the autosuspend.
Regards,
Loic
--
Intel Open Source Technology Center
http://oss.intel.com/
--
To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
