Hi Luiz,
>>> This makes btusb to queue ACL and events during a polling interval
>>> by using of a delayed work, with the interval working as a time window
>>> where frames received from different endpoints are considered to be
>>> arrived at same time and then attempt to resolve potential conflics by
>>> processing the events ahead of ACL packets.
>>>
>>> It worth noting though that priorizing events over ACL data may result
>>> in inverting the order compared to how they appeared over the air, for
>>> instance there may be packets received before a disconnect event that
>>> will be discarded and unencrypted packets received before encryption
>>> change which would considered encrypted, because of these potential
>>> changes on the order the support for queuing during the polling
>>> interval is not enabled by default so platforms have the following
>>> means to enable it:
>>>
>>> At build-time:
>>>
>>> CONFIG_BT_HCIBTUSB_INTERVAL=y
>>>
>>> At runtime with use of module option:
>>>
>>> enable_interval
>>>
>>> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@xxxxxxxxx>
>>> ---
>>> drivers/bluetooth/Kconfig | 7 ++++
>>> drivers/bluetooth/btusb.c | 88 ++++++++++++++++++++++++++++++++++-----
>>> 2 files changed, 84 insertions(+), 11 deletions(-)
>>>
>>> diff --git a/drivers/bluetooth/Kconfig b/drivers/bluetooth/Kconfig
>>> index 4e73a531b377..2f20a853d946 100644
>>> --- a/drivers/bluetooth/Kconfig
>>> +++ b/drivers/bluetooth/Kconfig
>>> @@ -41,6 +41,13 @@ config BT_HCIBTUSB_AUTOSUSPEND
>>> This can be overridden by passing btusb.enable_autosuspend=[y|n]
>>> on the kernel commandline.
>>>
>>> +config BT_HCIBTUSB_INTERVAL
>>> + bool "Enable notification of USB polling interval"
>>> + depends on BT_HCIBTUSB
>>> + help
>>> + Say Y here to enable notification of USB polling interval for
>>> + Bluetooth USB devices by default.
>>> +
>>> config BT_HCIBTUSB_BCM
>>> bool "Broadcom protocol support"
>>> depends on BT_HCIBTUSB
>>> diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c
>>> index b14102fba601..38cb5448fc69 100644
>>> --- a/drivers/bluetooth/btusb.c
>>> +++ b/drivers/bluetooth/btusb.c
>>> @@ -30,7 +30,7 @@
>>> static bool disable_scofix;
>>> static bool force_scofix;
>>> static bool enable_autosuspend = IS_ENABLED(CONFIG_BT_HCIBTUSB_AUTOSUSPEND);
>>> -
>>> +static bool enable_interval = IS_ENABLED(CONFIG_BT_HCIBTUSB_INTERVAL);
>>> static bool reset = true;
>>>
>>> static struct usb_driver btusb_driver;
>>> @@ -519,8 +519,12 @@ struct btusb_data {
>>>
>>> unsigned long flags;
>>>
>>> - struct work_struct work;
>>> - struct work_struct waker;
>>> + struct work_struct work;
>>> + struct work_struct waker;
>>> + struct delayed_work rx_work;
>>> +
>>> + struct sk_buff_head acl_q;
>>> + struct sk_buff_head evt_q;
>>
>> so does it make sense to keep them separate if we delay processing anyway.
>
>
> We need them to reorder in case they are received at the same time
> (within the same polling interval), we could reorder in place but then
> will need to iterate into the queue to find where to insert events,
> using 2 queues is a lot simpler to understand and probably perform
> better.
do we actually have to queue the HCI event as well. Or just make sure that the bulk endpoints are processed at the same interval as the interrupt endpoint?
Regards
Marcel
