bluez regression: Bluetooth audio fails to reconnect after resume

Just a correction here, <https://git.kernel.org/pub/scm/bluetooth/bluez.git/commit/?id=d6e9539e31c6bb5afd39ec6f09c518d232e6345d> is the patch that broke this. Fix is here <https://marc.info/?l=linux-bluetooth&m=151801178209679&w=2>.

Thanks for the info.

I think we should patch this into Ubuntu 18.04 _after_ it gets released in 3 weeks.

The reason is that I want to prioritize 18.04 releasing with mature bluez and pulseaudio packages that have not changed in a long time on release day. So there can't be any last-minute surprises. Then after release day we can patch in bug fixes like this one, but will also have the release version to fall back on should anything go wrong.

Also, this bug is not major enough to justify patching this close to release. Even most users of Bluetooth audio I suspect will never notice it.

I can't find any fix that exists upstream yet. It sounds like Red Hat just reverted the problematic commit.

If you can't find out where an upstream fix exists (in git.kernel.org) then please start by creating an upstream bug report for it here:
https://bugzilla.kernel.org/enter_bug.cgi?product=Drivers&component=Bluetooth

And then let us know here what the ID of the new bug is.

Found it. The fix is bluez commit 1873096352f518d3247f8efb3c2e0aa8804e50ac

> Also, this bug is not major enough to justify patching this close to release. Even most users of Bluetooth audio I suspect will never notice it.

I've definitely noticed it. And manually running "sudo systemctl restart bluetooth" everytime I open my laptop is annoying. Is there a way to automate the workaround? (Putting a script in /lib/systemd/system-sleep/ does not seem to work).

I agree with Guillaume, I can not understand how releasing 18.04 with broken bluetooth and calling it mature version. I would assume that most user are running ubuntu on laptop and will notice it.
I also was under the impression that the purpose of the beta release was to shake out the bugs be for the final release!

As much as I love to have this fixed in the final release, I understand Daniel's hesitation. The point is to try and keep things as stable as possible prior to releasing. Good news it is fixed and it is just a matter of time before it is released.

"Mature" means it hasn't changed in a while and we know what bugs it contains. "Mature" does not mean zero bugs, but at least by not changing it at the last minute we don't risk introducing bigger bugs. And this bug is not big at all.

As an example, just today the "gjs" package has blown up with lots of new crashes that didn't exist last week. This was caused by an unnecessary upgrade of a previously mature package, and it looks like it may risk the whole 18.04 release. That is why you shouldn't change things late in the release cycle.

Now that 18.04 out now, can we have new build of the packet?

I just need to find out if we're allowed to try and push v5.49 into bionic, or if we have to cherry pick the fix alone.

I take it back. Version 5.49 adds "notable new features also".

So we will push 5.49 to Ubuntu 18.10, and only fix this bug in 18.04.

So far, I can't wait for 5.49 when 5.46 is fixing my trouble. So I downgraded to have my audio chain usable, which was not the case at with 5.48 : Can't connect at ll my UE BOOM with Dell 5540.

I'm not suggesting waiting for 5.49, kind of. This bug will be fixed in an update to 5.48.

Policy however requires that we always fix the development release first. So Ubuntu 18.10 will probably need to receive bluez 5.49 before we can get approval for a bluez 5.48 update to Ubuntu 18.04.

Now available for testing:

https://launchpad.net/~bluetooth/+archive/ubuntu/bluez

Only cosmic is expected to get 5.49 but the above PPA also provides it for bionic if you want to try it out without changing the whole OS :)

Awesome!... Thanks Daniel.

On Wed, May 9, 2018 at 6:01 AM Daniel van Vugt <
<email address hidden>> wrote:

> Now available for testing:
>
> https://launchpad.net/~bluetooth/+archive/ubuntu/bluez
>
> Only cosmic is expected to get 5.49 but the above PPA also provides it
> for bionic if you want to try it out without changing the whole OS :)
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1759628
>
> Title:
> bluez regression: Bluetooth audio fails to reconnect after resume
>
> Status in bluez package in Ubuntu:
> In Progress
> Status in bluez package in Fedora:
> Fix Released
> Status in Suse:
> Fix Released
>
> Bug description:
> STEPS TO REPRODUCE:
> 1. Connect to Bluetooth audio device
> 2. Suspend & Resume
> 3. Reconnect to Bluetooth device.
>
> This regression in bluez 5.48 has already been identified and fixed
> upstream. Report is here <https://marc.info/?l=linux-
> bluetooth&m=151616078627045&w=2> and patch is here
> <
> https://git.kernel.org/pub/scm/bluetooth/bluez.git/commit/?id=d6e9539e31c6bb5afd39ec6f09c518d232e6345d
> >.
>
> Syslog reports messages as follows when this issue is happening (I have
> replaced my device's MAC address with [MAC]):
> Mar 28 12:34:29 cue pulseaudio[1859]: [pulseaudio] bluez5-util.c:
> Information about device /org/bluez/hci0/dev_[MAC] is invalid
> Mar 28 12:34:29 cue bluetoothd[984]: Endpoint replied with an error:
> org.bluez.Error.InvalidArguments
> Mar 28 12:34:33 cue pulseaudio[1859]: [pulseaudio] bluez5-util.c:
> Information about device /org/bluez/hci0/dev_[MAC] is invalid
> Mar 28 12:34:33 cue bluetoothd[984]: Endpoint replied with an error:
> org.bluez.Error.InvalidArguments
>
> Workaround is to run sudo systemctl restart bluetooth after resume.
>
> ProblemType: Bug
> DistroRelease: Ubuntu 18.04
> Package: bluez 5.48-0ubuntu3
> ProcVersionSignature: Ubuntu 4.15.0-12.13-generic 4.15.7
> Uname: Linux 4.15.0-12-generic x86_64
> ApportVersion: 2.20.9-0ubuntu1
> Architecture: amd64
> CurrentDesktop: KDE
> Date: Wed Mar 28 12:37:11 2018
> InstallationDate: Installed on 2018-03-23 (5 days ago)
> InstallationMedia: Kubuntu 18.04 LTS "Bionic Beaver" - Alpha amd64
> (20180306.1)
> InterestingModules: rfcomm bnep btusb bluetooth
> MachineType: Hewlett-Packard HP ProBook 640 G1
> ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-4.15.0-12-generic
> root=/dev/mapper/internal-root ro quiet splash vt.handoff=1
> SourcePackage: bluez
> UpgradeStatus: No upgrade log present (probably fresh install)
> dmi.bios.date: 01/25/2018
> dmi.bios.vendor: Hewlett-Packard
> dmi.bios.version: L78 Ver. 01.43
> dmi.board.name: 2101
> dmi.board.vendor: Hewlett-Packard
> dmi.board.version: KBC Version 16.3C
> dmi.chassis.type: 10
> dmi.chassis.vendor: Hewlett-Packard
> dmi.modalias:
> dmi:bvnHewlett-Packard:bvrL78Ver.01.43:bd01/25/2018:svnHewlett-Packard:pnHPProBook640G1:pvrA3009DD10303:rvnHewlett-Packard:rn2101:rvrKBCVersion16.3C:cvnHewlett-Packard:ct10:cvr:
> dmi.product.family: 103C_5336AN G=N L=BUS B=HP S=PRO
> dmi.product.name: HP ProBook 640 G1
> dmi.produc...

Works fine for me with the update from bluez PPA, thanks #29

Thanks for the PPA. Working fine in Kubuntu 18.04.

Fixed for me using the above mentioned PPA.
Thx!

And I agree with Guillaume that, for me, this was a major annoyance nonetheless...

> Also, this bug is not major enough to justify patching this close to release. Even most users of Bluetooth audio I suspect will never notice it.

This affects anyone using a laptop and bluetooth headset. I'm not sure how you can come to the conclusion that most end users will never notice it. Before I found the "sudo systemctl restart bluetooth" solution I was adding removing my bluetooth device, restarting pulse and so on. I've installed bluez 5.49.4 from debian which works flawlessy. Thank you for fixing it soon in ubuntu as well.

I've been using the bluez ppa for a while now and I haven't had any issues with it. Before, I'd regularly have connection issues to my bose qc 35. Now, I can always connect immediately, haven't had an issue since.

I also noticed that my headset now connects a lot quicker than with the stock 18.04 bluez.

So I'm really hoping that this can get included in mainline 18.04 as soon as possible, this will make many users happy :)

The upstream commit to fix this issues is this one:
https://git.kernel.org/pub/scm/bluetooth/bluez.git/commit/src/advertising.c?id=1873096352f518d3247f8efb3c2e0aa8804e50ac

I've applied it to Ubuntu's bluez 5.48 and confirmed that it works, as expected.

Please backport!

We know. See comment #19.

Any backport to 18.04 is blocked by the requirement that we release to 18.10 first.

And the release to 18.10 is blocked waiting on sponsorship in bug 1759961.

When you say "sponsorship", are you looking for testing? Or some more official process?

"Sponsorship" means I have finished preparing the release but don't have permission to put it into 18.10. So that last step is waiting for someone who has permission to do it.

This bug was fixed in the package bluez - 5.50-0ubuntu1

---------------
bluez (5.50-0ubuntu1) cosmic; urgency=medium

  * New upstream release 5.50:
    - Fix issue with GATT and reading long values.
    - Fix issue with GATT and reading multiple includes.
    - Fix issue with GATT and service changes when offline.
    - Fix issue with handling secondary service discovery.
    - Fix issue with handling persistency of CCC values.
    - Fix issue with handling Mesh session on disconnection.
    - Fix issue with handling Mesh proxy PDU SAR message length.
    - Fix issue with handling Mesh default heartbeat TTL value.
    - Add support for Mesh node-reset operation handling.
    - Add support for GATT authorization request handling.
    - Add support for GATT minimum key size requirements.
    - Add support for GATT server and included services.
    - Add support for handling additional advertising data.
    - Add support for handling separate discoverable state.
    - Add support for enabling HFP version 1.7 features.
    - Add support for dedicated Bluetooth logging daemon.
  * Refreshed debian/patches to work with 5.50 and apply cleanly.
  * Fixed lintian errors:
    - dpkg-maintscript-helper missing in postinst and postrm.
    - Synchronized (some) man pages from Debian.
  * New upstream release 5.49 (LP: #1759961):
    - Fix issue with configuring discoverable advertising flag.
    - Fix issue with bearer selection and single mode controllers.
    - Fix issue with Connect and ConnectProfile returning in progress.
    - Fix issue with missing Paired property change when not bonded.
    - Fix issue with storage for controllers without public address.
    - Fix issue with handling AVCTP disconnecting the channel queue.
    - Fix issue with not clearing connectable setting on power off.
    - Fix issue with creating multiple mgmt socket instances.
    - Fix issue with GATT server and BR/EDR only devices.
    - Fix issue with InterfaceAdded event ordering. (LP: #1759628)
    - Add support for generic ConnectDevice method call.
    - Add support for Mesh heartbeat client functionality.
  * Drop upstreamed patch:
    - fix-lp1750308-upstream-commit-1bf03369.patch

 -- Daniel van Vugt <email address hidden> Wed, 06 Jun 2018 16:26:40 +0800

Will this fix (aka the new package version) be released to the 18.04 repos?

Here's a patch for 18.04.

Next: We await sponsorship of the patch, which means it will enter "bionic-proposed" for testing.

Then: You will all be asked to test the proposed update to verify it works before it reaches other users.

SRU upload to the bionic queue now

Hello Ian, or anyone else affected,

Accepted bluez into bionic-proposed. The package will build now and be available at https://launchpad.net/ubuntu/+source/bluez/5.48-0ubuntu3.1 in a few hours, and then in the -proposed repository.

Please help us by testing this new package. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation on how to enable and use -proposed.Your feedback will aid us getting this update out to other Ubuntu users.

If this package fixes the bug for you, please add a comment to this bug, mentioning the version of the package you tested and change the tag from verification-needed-bionic to verification-done-bionic. If it does not fix the bug for you, please add a comment stating that, and change the tag to verification-failed-bionic. In either case, details of your testing will help us make a better decision.

Further information regarding the verification process can be found at https://wiki.ubuntu.com/QATeam/PerformingSRUVerification . Thank you in advance!

The fix for 18.04 is now available for testing, in bionic proposed.

If you would like to just download and test the debs directly they are here:
https://launchpad.net/ubuntu/+source/bluez/5.48-0ubuntu3.1/+build/15084683

I downloaded and installed bluez_5.48-0ubuntu3.1_amd64.deb. It seems that the bug is fixed. :)
Thx!

I confirm

Bug is fixed. All working fine. automatic reconnect without turning on/off
on my headphones.

On Thu, Jul 5, 2018 at 10:11 AM Alexander Lochmann <
<email address hidden>> wrote:

> I downloaded and installed bluez_5.48-0ubuntu3.1_amd64.deb. It seems that
> the bug is fixed. :)
> Thx!
>
> --
> You received this bug notification because you are subscribed to a
> duplicate bug report (1778506).
> https://bugs.launchpad.net/bugs/1759628
>
> Title:
> bluez regression: Bluetooth audio fails to reconnect after resume
>
> Status in bluez package in Ubuntu:
> Fix Released
> Status in bluez source package in Bionic:
> Fix Committed
> Status in bluez package in Fedora:
> Fix Released
> Status in Suse:
> Fix Released
>
> Bug description:
> [Impact]
>
> Users of Bluetooth audio have no sound after they suspend and resume
> the system.
>
> [Test Case]
>
> 1. Connect to Bluetooth audio device
> 2. Suspend & Resume
> 3. Reconnect to Bluetooth device.
>
> [Regression Potential]
>
> Low. Although common Bluetooth code is modified in the fix, it has
> been released as a patch in other distros for some time already.
>
> [Other Info]
>
> Already released to cosmic as part of bluez version 5.50.
>
> This regression in bluez 5.48 has already been identified and fixed
> upstream. Report is here <https://marc.info/?l=linux-
> bluetooth&m=151616078627045&w=2> and patch is here
> <
> https://git.kernel.org/pub/scm/bluetooth/bluez.git/commit/?id=d6e9539e31c6bb5afd39ec6f09c518d232e6345d
> >.
>
> Syslog reports messages as follows when this issue is happening (I have
> replaced my device's MAC address with [MAC]):
> Mar 28 12:34:29 cue pulseaudio[1859]: [pulseaudio] bluez5-util.c:
> Information about device /org/bluez/hci0/dev_[MAC] is invalid
> Mar 28 12:34:29 cue bluetoothd[984]: Endpoint replied with an error:
> org.bluez.Error.InvalidArguments
> Mar 28 12:34:33 cue pulseaudio[1859]: [pulseaudio] bluez5-util.c:
> Information about device /org/bluez/hci0/dev_[MAC] is invalid
> Mar 28 12:34:33 cue bluetoothd[984]: Endpoint replied with an error:
> org.bluez.Error.InvalidArguments
>
> Workaround is to run sudo systemctl restart bluetooth after resume.
>
> ProblemType: Bug
> DistroRelease: Ubuntu 18.04
> Package: bluez 5.48-0ubuntu3
> ProcVersionSignature: Ubuntu 4.15.0-12.13-generic 4.15.7
> Uname: Linux 4.15.0-12-generic x86_64
> ApportVersion: 2.20.9-0ubuntu1
> Architecture: amd64
> CurrentDesktop: KDE
> Date: Wed Mar 28 12:37:11 2018
> InstallationDate: Installed on 2018-03-23 (5 days ago)
> InstallationMedia: Kubuntu 18.04 LTS "Bionic Beaver" - Alpha amd64
> (20180306.1)
> InterestingModules: rfcomm bnep btusb bluetooth
> MachineType: Hewlett-Packard HP ProBook 640 G1
> ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-4.15.0-12-generic
> root=/dev/mapper/internal-root ro quiet splash vt.handoff=1
> SourcePackage: bluez
> UpgradeStatus: No upgrade log present (probably fresh install)
> dmi.bios.date: 01/25/2018
> dmi.bios.vendor: Hewlett-Packard
> dmi.bios.version: L78 Ver. 01.43
> dmi.board.name: 2101
> dmi.board.vendor: Hewlett-Packard
> dmi.board.version: KBC ...

Thanks. Someone also needs to scroll up and change the tag from verification-needed-bionic to verification-done-bionic. Ideally someone who has verified the fix.

Cant find where?

On Fri, Jul 6, 2018 at 4:16 AM Daniel van Vugt <
<email address hidden>> wrote:

> Thanks. Someone also needs to scroll up and change the tag from
> verification-needed-bionic to verification-done-bionic. Ideally someone
> who has verified the fix.
>
> --
> You received this bug notification because you are subscribed to a
> duplicate bug report (1778506).
> https://bugs.launchpad.net/bugs/1759628
>
> Title:
> bluez regression: Bluetooth audio fails to reconnect after resume
>
> Status in bluez package in Ubuntu:
> Fix Released
> Status in bluez source package in Bionic:
> Fix Committed
> Status in bluez package in Fedora:
> Fix Released
> Status in Suse:
> Fix Released
>
> Bug description:
> [Impact]
>
> Users of Bluetooth audio have no sound after they suspend and resume
> the system.
>
> [Test Case]
>
> 1. Connect to Bluetooth audio device
> 2. Suspend & Resume
> 3. Reconnect to Bluetooth device.
>
> [Regression Potential]
>
> Low. Although common Bluetooth code is modified in the fix, it has
> been released as a patch in other distros for some time already.
>
> [Other Info]
>
> Already released to cosmic as part of bluez version 5.50.
>
> This regression in bluez 5.48 has already been identified and fixed
> upstream. Report is here <https://marc.info/?l=linux-
> bluetooth&m=151616078627045&w=2> and patch is here
> <
> https://git.kernel.org/pub/scm/bluetooth/bluez.git/commit/?id=d6e9539e31c6bb5afd39ec6f09c518d232e6345d
> >.
>
> Syslog reports messages as follows when this issue is happening (I have
> replaced my device's MAC address with [MAC]):
> Mar 28 12:34:29 cue pulseaudio[1859]: [pulseaudio] bluez5-util.c:
> Information about device /org/bluez/hci0/dev_[MAC] is invalid
> Mar 28 12:34:29 cue bluetoothd[984]: Endpoint replied with an error:
> org.bluez.Error.InvalidArguments
> Mar 28 12:34:33 cue pulseaudio[1859]: [pulseaudio] bluez5-util.c:
> Information about device /org/bluez/hci0/dev_[MAC] is invalid
> Mar 28 12:34:33 cue bluetoothd[984]: Endpoint replied with an error:
> org.bluez.Error.InvalidArguments
>
> Workaround is to run sudo systemctl restart bluetooth after resume.
>
> ProblemType: Bug
> DistroRelease: Ubuntu 18.04
> Package: bluez 5.48-0ubuntu3
> ProcVersionSignature: Ubuntu 4.15.0-12.13-generic 4.15.7
> Uname: Linux 4.15.0-12-generic x86_64
> ApportVersion: 2.20.9-0ubuntu1
> Architecture: amd64
> CurrentDesktop: KDE
> Date: Wed Mar 28 12:37:11 2018
> InstallationDate: Installed on 2018-03-23 (5 days ago)
> InstallationMedia: Kubuntu 18.04 LTS "Bionic Beaver" - Alpha amd64
> (20180306.1)
> InterestingModules: rfcomm bnep btusb bluetooth
> MachineType: Hewlett-Packard HP ProBook 640 G1
> ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-4.15.0-12-generic
> root=/dev/mapper/internal-root ro quiet splash vt.handoff=1
> SourcePackage: bluez
> UpgradeStatus: No upgrade log present (probably fresh install)
> dmi.bios.date: 01/25/2018
> dmi.bios.vendor: Hewlett-Packard
> dmi.bios.version: L78 Ver. 01.43
> dmi.board.name: 2101
> dmi.board.vendor: Hewlett-Packard
> dmi.board.version: KBC Version 16.3C
> dmi.chas...

This issue is resolved for me in bionic-proposed.

#29 Fix confirmed on Ubuntu 18.04 on Miix 700 / Intel 8260

You can ignore comment #29 now. The proper official fix for 18.04 is:
https://launchpad.net/ubuntu/+source/bluez/5.48-0ubuntu3.1/+build/15084683

This bug was fixed in the package bluez - 5.48-0ubuntu3.1

---------------
bluez (5.48-0ubuntu3.1) bionic; urgency=medium

  * Add fix-lp1759628-upstream-commit-1873096352.patch to fix loss of
    audio on suspend/resume (LP: #1759628)

 -- Daniel van Vugt <email address hidden> Fri, 22 Jun 2018 17:27:15 +0800

The verification of the Stable Release Update for bluez has completed successfully and the package has now been released to -updates. Subsequently, the Ubuntu Stable Release Updates Team is being unsubscribed and will not receive messages about this bug report. In the event that you encounter a regression using the package from -updates please report a new bug using ubuntu-bug and tag the bug report regression-update so we can easily find any regressions.