Hi Lukasz,
On Thu, Aug 28, 2014 at 11:47 AM, Lukasz Rymanowski
<lukasz.rymanowski@xxxxxxxxx> wrote:
> Hi Luiz,
>
> On Thu, Aug 28, 2014 at 10:06 AM, Luiz Augusto von Dentz
> <luiz.dentz@xxxxxxxxx> wrote:
>> Hi Lukasz,
>>
>> On Wed, Aug 27, 2014 at 11:47 PM, Lukasz Rymanowski
>> <lukasz.rymanowski@xxxxxxxxx> wrote:
>>> GATT, HID, HOG, might be interested in the fact that some device has
>>> been unpaired in order to clear cache or similar. This patch adds means
>>> to register callback for unpaired event.
>>> ---
>>> android/bluetooth.c | 38 ++++++++++++++++++++++++++++++++++++++
>>> android/bluetooth.h | 3 +++
>>> 2 files changed, 41 insertions(+)
>>>
>>> diff --git a/android/bluetooth.c b/android/bluetooth.c
>>> index 984ecba..588ecd4 100644
>>> --- a/android/bluetooth.c
>>> +++ b/android/bluetooth.c
>>> @@ -41,6 +41,7 @@
>>> #include "lib/uuid.h"
>>> #include "src/shared/util.h"
>>> #include "src/shared/mgmt.h"
>>> +#include "src/shared/queue.h"
>>> #include "src/eir.h"
>>> #include "lib/sdp.h"
>>> #include "lib/sdp_lib.h"
>>> @@ -223,6 +224,8 @@ static struct ipc *hal_ipc = NULL;
>>>
>>> static bool kernel_conn_control = false;
>>>
>>> +static struct queue *unpaired_cb_list;
>>> +
>>> static void get_device_android_addr(struct device *dev, uint8_t *addr)
>>> {
>>> /*
>>> @@ -1688,6 +1691,22 @@ void bt_auto_connect_remove(const bdaddr_t *addr)
>>> error("Failed to remove device");
>>> }
>>>
>>> +static bool match_by_value(const void *data, const void *user_data)
>>> +{
>>> + return data == user_data;
>>> +}
>>> +
>>> +void bt_unpaired_register(bt_unpaired_device_cb cb)
>>> +{
>>> + if (!unpaired_cb_list)
>>> + return;
>>> +
>>> + if (queue_find(unpaired_cb_list, match_by_value, cb))
>>> + return;
>>> +
>>> + queue_push_head(unpaired_cb_list, cb);
>>> +}
>>
>> Shouldn't it take the actual device address as well? Otherwise each
>> callback has to do a lookup on its own which is not that great.
>>
>
> I thought about it but eventually drop this idea. It is because each
> module interested in this callback needs to find provided device on
> their lists anyway. It is in order to remove it.
> If we would like to provide address in register function we would have
> to do some additional checks before actually callback call - so double
> check which is not nice.
> Queue would be also bigger since it would have to be for each device
> etc. So I would keep that approach if you agree
I guess you don't want to pass the struct pointer as user data either,
that would save the extra lookup, anyway I guess we can deal with that
later if we found it does not do well in some cases.
--
Luiz Augusto von Dentz
--
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
