Return-Path: From: Steve Abner Subject: Re: bluez iphone connection To: linux-bluetooth@vger.kernel.org References: <5662D1DF.4000601@att.net> Cc: Luiz Augusto von Dentz Message-ID: <56664C25.8040308@att.net> Date: Mon, 7 Dec 2015 22:19:01 -0500 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8; format=flowed Sender: linux-bluetooth-owner@vger.kernel.org List-ID: On 12/07/2015 09:24 AM, Luiz Augusto von Dentz wrote: > What kind of setup is this, can you use a different adapter per > system, Bluetooth pair work in relation to the adapter address so the > iPhone has no idea you are switching between systems. One computer, two operating systems, one common controller. I really like the idea that the phone or trackpad shouldn't care about switching systems. Problem I have is Bluez has to set pairing, not user, so one system sets with pin of xxxx, other with yyyy, device seems to say machine A, so I'll except pin xxxx, but not yyyy. So copy xxxx to replace yyyy. If bluez asked for pin then could tell it instead of copy it, but copy easier then remembering hundreds of pins. >> > Both systems I get: >> >Bluetooth: Unexpected continuation frame (len 0) >> >connect error: Too many levels of symbolic links (40) > Im afraid there is something wrong which the way you copy pairing > details from one system to the other, anyway it is probably not > recommended to do that in the first place. This only occurs with iphone, not a trackpad. So copying the pin should not be an issue. The pin is encoded, don't know by which crypto method. Plus, one way, phone to pc works, pc to phone doesn't. >> > Honestly forget which applies to connection from pc to phone, vs. phone to >> >pc. >> >If someone could direct me to documentation, which I cant find through web >> >searching, >> >or answer this, I'd be thankful. Hopefully, an automated process for end >> >users to simply connect to devices, would be nice. > There is plenty of example on how to pair, gnome has done that if you > need a graphical interface and bluetoothctl is also capable of doing > it, but Im afraid the problem with your setup so I suggest you use a > different adapter per system if possible or mount the same partition > in the place of /var/lib/bluetooth so the system are synchronized when > you reboot/ ArchLinux normally has most info, debian not too bad, one guy's article helped me to create a systemd service, based on his BT keyboard issues. And I don't see why one controller must be dedicated per OS. Binding to one partition, not an option. When LFS moves to another machine it will never see that partition. At which time it will be a different controller, but I am not at that development stage. I assume same hostname, yet different controller will provide a new connection slot on phone, thereby not creating an issue of pins. This would be a different /var/lib/bluetooth///info. But all said, this points to bluez. I haven't even asked about its other issues. Both systems: Failed to obtain handles for "Service Changed" characteristic One with multiple: Not enough free handles to register service Was just working out connecting. I have gotten it to where I can switch between OSes and connect by phone issuing the connect command. Both are accompanied by: Bluetooth: Unexpected continuation frame (len 0) The ubuntu can issue, through graphical means (and NetworkManager), connection. The LFS system, no graphical yet, just fb/gpu terminal, can not issue connection. Bluetoothctl has the "info", trusted, paired, etc, but refuses to connect. I have connected by "remove", restart phone and computer, issue new pin. Of course the means that ubuntu will no longer work until I re-pair, new pin, which without the copy means I must, remove, reboot, re-pair the LFS, and around I go. Now if someone knows exact sequence of connection of a paired device using either hcitool or bluetoothctl this would be helpful. I've tried agent on, pairable on, discoverable on, pair , (already paired it complains), connect , different sequences of those commands. Some is new, just added in sound today (so not researched yet): bluetoothd[269]: a2dp-source profile connect failed for 70:3E:AC:50:D8:89: Protocol not available kernel: Bluetooth: Unexpected continuation frame (len 0) bluetoothd[269]: Can't open input device: No such file or directory (2) bluetoothd[269]: AVRCP: failed to init uinput for 70:3E:AC:50:D8:89 bluetoothd[269]: Invalid folder length I am by no means an expert, but logically and process of elimination don't point to pin, nor the concept of one controller per OS, thou broadcom might like that idea. A missing module, a support library maybe? Steve