IssueTracker

I've got a BLE peripheral that is capable of being connected as slave to multiple centrals and just tried what would happen if you use autoConnect parameter set to true and also have a pending connection to another device. What I expected happens:

On a Nexus 5 (that has a Broadcom Bluetooth 4.1 chip), running the latest Marshmallow update, as soon as the BLE peripheral starts an advertisement after it has connected, the previous link is dropped by the Broadcom chip. Since this goes on and on, it gets stuck in a "connect-disconnect loop". I've attached a new snoop log for this. See from packet 1865 and forward.

On an LG G3 phone (that has a Qualcomm Bluetooth 4.1 chip), running the latest Marshmallow update, as soon as the BLE peripheral starts an advertisement after it has connected, the bluetooth chip now initiates the second connection to this device, notifying the host about this, which as written above gets confused and gets stuck in a state where it will make no further connections until Bluetooth is restarted. I've attached a new snoop log for this as well. See from packet 593 and forward.

On a Nexus 5X phone (that has a Qualcomm Bluetooth 4.2 chip), no bug will be triggered and there are no problems. The reason is that since the Bluetooth 4.2 standard, some new sentences have been added: "There shall be only one connection between two LE device addresses. An initiator shall not send a connection request to an advertiser it is already connected to. If an advertiser receives a connection request from an initiator it is already connected to, it shall ignore that request." The bluetooth chip honours this and will therefore not establish a new connection, even though the peripheral advertises and its address is in the white list.

So, I really think the best fix would be to make sure that when there is a pending connection, the white list should not contain addresses to devices that already have an established connection. As newer BLE peripherals that support multiple centrals will eventually be more common on the market, I think it's good to fix this issue as soon as possible before people are stuck with non-upgradable old Marshmallow phones.

FYI Same test as in last comment with a Sony Xperia Z3 (Broadcom Bluetooth 4.1) with the latest Android N Developer Preview (NPD56N.1000106): the issue still persists. It behaves exactly the same as with the Nexus 5 running Marshmallow. I'm attaching a snoop log. See packet 1161 and forward.

Can you please check the issue on latest version of android N and let us know the result?
If issue reproduces, then please share the bugreport and bt snoop logs for the same.

Android bug report:
After reproducing the issue, navigate to developer settings, ensure ‘USB debugging’ is enabled, then enable ‘Bug report shortcut’. To take bug report, hold the power button and select the ‘Take bug report’ option.

Note: Please upload the files to google drive and share the folder to android-bugreport@google.com, then share the link here.

Hi. Nothing has changed in the new Android version (NBD90W), so the bluetooth hci captures and logcat logs above are still valid.

Android's Bluetooth team has already read this report and are currently working on a fix: https://android-review.googlesource.com/#/c/268213/

We have passed this defect on to the development team and will update this issue with more information as it becomes available.

I thought Qualcomm's Bluetooth 4.2 controllers did not suffer from this problem but apparently I was wrong.

Here attached is a hci snoop log from a Moto Z Play 2016 having a Qualcomm Bluetooth 4.2 controller. At packet 617 the device 80:e4:da:70:ec:66 gets connected. At packet 686 it gets connected again even though no Disconnect Complete event has been sent in between. 31 seconds later the first connection times out. I don't understand how that is possible however that the first connections stays alive for so long time since the applied supervision timeout is 8 seconds. The peripheral used cannot be in advertising state and connected state at the same time.

This means using "auto connect" with Android >= 6.0 for more than one device at the same time is very likely to fail if you just wait long enough.

I have also filed a new CL here: https://android-review.googlesource.com/#/c/315988/.

I am curious - what effect would setting `autoConnect` parameter to false have on this issue?  What is the necessity for `autoConnect` being set to true.

If the issue has already occurred, you won't be able to connect even if you set autoConnect to false. But the issue can only be triggered in the first place by a device that has a gatt connection with the autoConnect parameter set to true.

I can see https://android-review.googlesource.com/#/c/268213/ was never merged, but https://android-review.googlesource.com/#/c/315988/ was.
Does 315988 superseed 268213, or is 268213 still needed?

Yes, the last patch has fixed the issue. The first one should be marked as abandoned.

The status of this bug says that it is not in production yet - even though it was merged into AOSP master on May 13, 2017.
Does this mean that both Android 6, 7.x and 8.x still have this problem?

Thank you for your feedback. We assure you that we are doing our best to address the issue reported, however our product team has shifted work priority that doesn't include this issue. For now, we will be closing the issue as won't fix obsolete. If this issue currently still exists, we request that you log a new issue along with latest bug report here https://goo.gl/TbMiIO.

How this can be new. It has been a known issue since 5.1.1 when up. Nexus 5, in Canada, on Rogers, not a smartphone anymore !

I've opened up my Nexus 5 and the 3g antenna connector was not properly fitted. Might be worthed that you guys open up your N5s and check it too.
I've attached a jpg with the antenna connectors (top right one was not connected in my phone). After properly fixing the connector, phone is now connecting to 3g/4g + internet just fine. Question is now... why was the connector not properly fitted?

Flashed N5 to 5.0.1 (from 5.1.1) using Nexus Reboot Toolkit and just reactivated my data on Simple mobile (after a month off the data plan). I then rebooted and MY DATA IS BACK!! Anyone else have a similar experience going from 5.1.1 to 5.0.1? Hoping it stays good. Hiding the update notification for the time being.

I lose my data for hours at a time intermittently for no reason, after a text or call is made usually seems to be the trigger,  Nexus 5 with ATT Straight Talk

I have a nexus 5, latest update that google rolled out, on the straight talk data plan by AT&T.I have the wifi to data compatibility issue as well as a bug that makes it so that it takes about 2 minutes to connect to LTE after turning off data. There is also another bug that makes it so that if I stay on LTE for too long then it still displays that I have LTE compatibility but in actuality I'm not even connected to the network. The resolve is to turn off data and turn back which will then show that there is no network connected and the cycle repeats.

I had data connection issue with My N4 on 5.0.1 after each time I get voice call. After upgrading to 5.1.1 it's one hell.. My N4 forgot what is data connection :( :(

Same issues here. Cannot keep data for very long without having to either reboot the phone or turn flight mode on and off. Not improved with latest version.
Making me wonder why I have an android phone.

I am experiencing similar issues with one of my 2 Nexus 7 tablets (Android 5.1.1).  I have mobile data through Vini in French Polynesia on both tablets.  The older Nexus 7 has a Baseband Version of 1231_0.18.0_049 and Build Number LMY47V. This tablet drops the Vini Cell connection after a while and can only be reconnected with a reboot.  The newer Nexus 7 has a Baseband Version of DEB_G00_2.43.0_1210 and a Build Number of LMY48L. I have no cellular connection problems with this newer tablet.

Lo mismo para mi, no tengo señal después de que me alejó de mi WiFi en casa o el trabajo, debo cambiar el tipo de red de lte a 3g o viceversa para que funcione de nuevo, esto no me pasaba antes de lollipop. Cuando lo van a arreglar??? Nexus 5 con 5.1.1. Creó que se le tendría q cambiar la prioridad a este caso porque cientos de personas tienen el mismo problema.

I noticed this problem while troubleshooting another; resetting APN settings got the connection back up.

Nexus 5 Lolipop 5.1.1

I still have my other problem to deal with though: After going online once, the phone is using data even when I turn it back off.

Same problem with Moto G

Guys, seems related to Android issue 36907297.

WifiFixerFree from PlayStore seems to improve and nearly patch the problem.

https://play.google.com/store/apps/details?id=com.symsoft.wififixerfree

Can you please check the issue on latest version of android and let us know the result?
If issue reproduces, then please share the bugreport for the same.

Android bug report:
After reproducing the issue, navigate to developer settings, ensure ‘USB debugging’ is enabled, then enable ‘Bug report shortcut’. To take bug report, hold the power button and select the ‘Take bug report’ option.

Note: Please upload the files to google drive and share the folder to android-bugreport@google.com, then share the link here.

We are closing this issue as we don't have enough actionable information. If you are still facing this problem on latest android version, please open new issue and add the relevant information along with reference to earlier issue.