Bug 1124996
Summary: | Bluetooth audio not connecting with kernel-3.16 | ||||||
---|---|---|---|---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Stuart D Gathman <stuart> | ||||
Component: | kernel | Assignee: | Kernel Maintainer List <kernel-maint> | ||||
Status: | CLOSED CURRENTRELEASE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> | ||||
Severity: | high | Docs Contact: | |||||
Priority: | unspecified | ||||||
Version: | 20 | CC: | extras-qa, gansalmon, grau.fran, itamar, jonathan, kernel-maint, lkundrak, lpoetter, madhu.chinakonda, mchehab, nforrester, rdieter, stuart, wtaymans | ||||
Target Milestone: | --- | Keywords: | Reopened | ||||
Target Release: | --- | ||||||
Hardware: | i686 | ||||||
OS: | Unspecified | ||||||
Whiteboard: | abrt_hash:fe7035aa348bc8b2fc1d9301f63f6cf3e540769d | ||||||
Fixed In Version: | Doc Type: | Bug Fix | |||||
Doc Text: | Story Points: | --- | |||||
Clone Of: | 981355 | ||||||
: | 1144875 (view as bug list) | Environment: | |||||
Last Closed: | 2014-10-22 02:38:38 UTC | Type: | --- | ||||
Regression: | --- | Mount Type: | --- | ||||
Documentation: | --- | CRM: | |||||
Verified Versions: | Category: | --- | |||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
Cloudforms Team: | --- | Target Upstream Version: | |||||
Embargoed: | |||||||
Attachments: |
|
Description
Stuart D Gathman
2014-07-30 19:31:56 UTC
On F20, here is what appears in journalctl -a when attempting to connect bluetooth audio: Jul 30 14:58:32 elissa.gathman.org gnome-session[1531]: ** Message: has_config_widget 94:23:6E:0B:BD:AF AudioSink Jul 30 14:58:33 elissa.gathman.org gnome-session[1531]: ** Message: has_config_widget 94:23:6E:0B:BD:AF A/V_RemoteControlTarget Jul 30 14:58:33 elissa.gathman.org gnome-session[1531]: ** Message: has_config_widget 94:23:6E:0B:BD:AF A/V_RemoteControl Jul 30 14:58:35 elissa.gathman.org bluetoothd[716]: connect error: Software caused connection abort (103) Jul 30 14:58:45 elissa.gathman.org fprintd[9347]: ** Message: No devices in use, exit Jul 30 14:58:56 elissa.gathman.org bluetoothd[716]: connect error: Connection timed out (110) pulseaudio-5.0-7.fc20.i686 pulseaudio-module-bluetooth-5.0-7.fc20.i686 On a second attempt, this is what we get: Jul 30 15:00:09 elissa.gathman.org gnome-session[1531]: available 0** Message: has_config_widget 94:23:6E:0B:BD:AF AudioSink Jul 30 15:00:09 elissa.gathman.org gnome-session[1531]: ** Message: has_config_widget 94:23:6E:0B:BD:AF A/V_RemoteControlTarget Jul 30 15:00:09 elissa.gathman.org gnome-session[1531]: ** Message: has_config_widget 94:23:6E:0B:BD:AF A/V_RemoteControl Jul 30 15:00:12 elissa.gathman.org bluetoothd[716]: connect error: Permission denied (13) Jul 30 15:00:14 elissa.gathman.org bluetoothd[716]: connect error: Connection refused (111) Jul 30 15:00:16 elissa.gathman.org bluetoothd[716]: connect error: Connection refused (111) Jul 30 15:00:17 elissa.gathman.org bluetoothd[716]: connect error: Connection refused (111) Jul 30 15:00:19 elissa.gathman.org bluetoothd[716]: connect error: Connection refused (111) This might be relevant: http://www.spinics.net/lists/linux-bluetooth/msg22729.html Can I just stop the bluetooth.service and manually run bluetoothd? Will try it for some debugging output. bluetoothd[24803]: src/device.c:connect_profiles() /org/bluez/hci0/dev_94_23_6E_0B_BD_AF (all), client :1.329 bluetoothd[24803]: profiles/audio/a2dp.c:a2dp_sink_connect() path /org/bluez/hci0/dev_94_23_6E_0B_BD_AF bluetoothd[24803]: a2dp-sink profile connect failed for 94:23:6E:0B:BD:AF: Protocol not available According to this random person on the internet, pulseaudio does not support bluez-5, and ALSA does not support bluez-5. And bluez-5 cannot support a2dp without pulseaudio. So this seems to be an integration total failure. :-( According to http://www.freedesktop.org/wiki/Software/PulseAudio/Notes/5.0/, the current [pulseaudio] 4.0 release does not support bluez 5, and it seems even 5.0 is not going to support HFP. And I heard ALSA cannot support bluez5 without pulseaudio... The version combo from he11: pulseaudio-5.0-7.fc20.i686 bluez-5.18-1.fc20.i686 Only HSP/HFP modes are not supported (yet), see also bug #1045548 and upstream http://bugs.freedesktop.org/show_bug.cgi?id=73325 This seems to be fixed now. It is not obvious which update fixed it. Ok, not quite fixed. 1) bluetooth audio stops working after suspend. A reboot fixes it. 2) with the last set of updates, it doesn't work again. packages that might be relevant: kernel-3.16.2-200.fc20.i686 gstreamer1-libav-1.2.4-2.fc20.i686 Using bluetoothctl (to avoid bugs in the gnome applet - 100% CPU during discovery), this is what now happens after connect (after fresh reboot): [bluetooth]# connect 94:23:6E:0B:BD:AF Attempting to connect to 94:23:6E:0B:BD:AF [CHG] Device 94:23:6E:0B:BD:AF Connected: yes Failed to connect: org.bluez.Error.Failed [CHG] Device 94:23:6E:0B:BD:AF Connected: no [CHG] Device 94:23:6E:0B:BD:AF Connected: yes [CHG] Device 94:23:6E:0B:BD:AF Connected: no [CHG] Device 94:23:6E:0B:BD:AF Connected: yes [CHG] Device 94:23:6E:0B:BD:AF Connected: no [CHG] Device 94:23:6E:0B:BD:AF Connected: yes [CHG] Device 94:23:6E:0B:BD:AF Connected: no [CHG] Device 94:23:6E:0B:BD:AF Connected: yes [CHG] Device 94:23:6E:0B:BD:AF Connected: no ... Created attachment 938040 [details]
journalctl -b0 after reboot with new kernel and attempted bluetooth connect
OK, if suspending is the trigger, that's most likely a driver issue (reassigning to kernel). Rebooting with previous kernel allows connect to work, so there is definitely a problem with kernel-3.16. However, pulseaudio still doesn't allow an audio connection. So that is broken also by the updates. Pulseaudio itself was not updated, so it is not clear what broke it. Should I open a new bug for the suspend problem? After suspend, bluetoothctl no longer discovers audio devices, although other computers are still discovered. A reboot restores the ability to discover audio devices. I considered it "working" when it was only broken by suspend. However, it is totally broken again - even after reboot. That is why I asked about creating a new bug for the suspend problem. Even rebooting with working kernel (kernel-3.15.10-201.fc20.i686), the audio stream is still broken. So there is also a problem somewhere in the stack from pulseaudio to kernel after the updates. It's usually best to track one issue per bug, so doing a new bug makes sense. With kernel-3.15, the bluetooth connects to BR-100. With kernel-3.16 the bluetooth fails to connect to BR-100. So there is clearly a kernel problem introduced in 3.16. I created another issue for the userland problem, where even after connecting with kernel-3.15, the device doesn't show up in pulseaudio after some unidentified update. I created another issue for the userland problems. *** Bug 1144875 has been marked as a duplicate of this bug. *** There are 3 separate issues. 1) the kernel breaks with 3.16 2) When the kernel was working, it would not work after suspend/resume 3) Even when the kernel is working, and the device can connect, there is still no bluetooth audio device in pulseaudio since some update after Sep 03. I was attempting to split into 3 issues as advised. After further research, the kernel-3.16 isn't broken. It has a new feature - real support for bt4 low power burst (I think that's what you call it). The Gnome Bluetooth config GUI looks like it isn't connected, but it really is. (Maybe the GUI can be improved.) The command line tool also shows it disconnected, but apparently if the initial connection succeeds, then you are good. So why wouldn't the audio connect? Well, apparently this change in protocol requires a power cycle of the bluetooth device before it will operate with the new regime. Or else, maybe pulseaudio was fixed again with the last set of updates. We may never know. |