Hi Archie, > According to the HID1.1 spec, part 5.3.4.9: > The HIDSDPDisable attribute is a Boolean value, which indicates > whether connection to the SDP channel and Control or Interrupt > channels are mutually exclusive. This feature supports Bluetooth > HID devices that have minimal resources, and multiplex those > resources between servicing the initialization (SDP) and runtime > (Control and Interrupt) channels. > > However, Bluez still tries to connect SDP upon HID connection, > regardless of the existence of the HIDSDPDisable attribute. > > This patch prevents the connection of SDP after HID has been > established, if the device has HIDSDPDisable attribute. out of curiosity, is a qualification test failing or do you have devices that really enforce the Disable SDP part of the specification. A long long long time ago (we are talking 15+ years) some HID devices only allowed 2 L2CAP channels to be open at the same time. So the PSM 1 for SDP needed to be closed before opening the second HID data channel. If this is failing devices, it would be good to include the btmon trace of failure in the commit message so that we archive this. Regards Marcel
