Hi Vinicius,
On Mon, May 6, 2013 at 7:23 PM, Vinicius Gomes
<vinicius.gomes@xxxxxxxxxxxxx> wrote:
> Hi Johan,
>
> On Fri, May 3, 2013 at 1:36 AM, Johan Hedberg <johan.hedberg@xxxxxxxxx> wrote:
>> Hi Vinicius,
>>
>> On Thu, May 02, 2013, Vinicius Costa Gomes wrote:
>>> diff --git a/src/profile.c b/src/profile.c
>>> index 0500983..2c35dc7 100644
>>> --- a/src/profile.c
>>> +++ b/src/profile.c
>>> @@ -1045,6 +1045,7 @@ static struct ext_io *create_conn(struct ext_io *server, GIOChannel *io,
>>> {
>>> struct btd_device *device;
>>> struct ext_io *conn;
>>> + struct btd_profile *profile;
>>> GIOCondition cond;
>>>
>>> conn = g_new0(struct ext_io, 1);
>>> @@ -1052,11 +1053,14 @@ static struct ext_io *create_conn(struct ext_io *server, GIOChannel *io,
>>> conn->proto = server->proto;
>>> conn->ext = server->ext;
>>> conn->adapter = btd_adapter_ref(server->adapter);
>>> + profile = &server->ext->p;
>>>
>>> device = adapter_find_device(server->adapter, dst);
>>>
>>> - if (device)
>>> + if (device) {
>>> conn->device = btd_device_ref(device);
>>> + conn->service = service_create(device, profile);
>>> + }
>>>
>>> cond = G_IO_HUP | G_IO_ERR | G_IO_NVAL;
>>> conn->io_id = g_io_add_watch(io, cond, ext_io_disconnected, conn);
>>
>> How would the service show up in the list of services for the device
>> object in question? To me it seems like src/device.c is the only place
>> that should be calling service_create because of this. What does seem to
>> be missing in profile.c is a call to btd_device_add_uuid for unexpected
>> connections from a device which we did not yet know to support a certain
>> UUID. The btd_device_add_uuid function should cause a new service to be
>> created, but it may not yet be enough to get the service to be assigned
>> to conn in profile.c.
>
> The problem in question seems that the service is already created, but
> it doesn't get associated with
> the conn in my particular case of incomming connections (testing the
> AG side of HFP, for the record).
The issue seems to be caused by a missing service search and
association in create_conn(), leaving conn->service unassigned and
leading to the issue you describe, affecting incoming connections for
external profiles.
I'll send a patchset to fix it.
Cheers,
Mikel
--
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
