New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 700585 link

Starred by 5 users

Issue metadata

Status: WontFix
Owner: ----
Closed: Apr 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 3
Type: Bug-Regression


Show other hotlists

Hotlists containing this issue:
Hotlist-1


Sign in to add a comment

ARC++ BTLE broken for edgar.

Project Member Reported by pbath...@chromium.org, Mar 11 2017

Issue description

Moving from https://b.corp.google.com/issues/36095748


In short, Bluetooth LE is broken for Android apps running under ARC++.  Full service discovery does not report back complete list of services, as verified with nRF Connect app.  The BLE stack seems to be omitting UUID 0x1800, Generic Access, from services list.  Apps that try to communicate with BLE devices flat out don't work as a result.

UI language:
en-US

Version:
58.0.3029.6 dev

Product Specific Data (whitelisted):
CHROME VERSION: 58.0.3029.6 dev
CHROMEOS_AUSERVER: <URL: 1>
CHROMEOS_RELEASE_BOARD: edgar-signed-mp-v2keys
CHROMEOS_RELEASE_DESCRIPTION: 9334.2.0 (Official Build) dev-channel edgar
CHROMEOS_RELEASE_TRACK: dev-channel
CHROMEOS_RELEASE_VERSION: 9334.2.0
ENTERPRISE_ENROLLED: Not managed
expi:  3300110 3300132 3312841 3313324 3313326 3313362 3313401 
hardware_class: EDGAR D25-A4C-S8I-A7K
lsusb: Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 001 Device 003: ID 8087:0a2a Intel Corp. 
Bus 001 Device 006: ID 04f2:b567 Chicony Electronics Co., Ltd 
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub

routes: default via <IPv4: 10> dev wlan0 metric 1 
<IPv4: 30>/24 dev wlan0 proto kernel scope link src <IPv4: 1>0 
<IPv4: 11>/30 dev br0 proto kernel scope link src <IPv4: 13> 


Browser:
Chrome 58.0.3029.6
 
Comment #4 from the buganizer bug

This might related to http://b/36072163 Fitbit app stops working
Labels: -Type-Bug Type-Bug-Regression
Re #1 , this seems to be regression because the fitbit app was working on minnie on M-57 9202.0.0 build and i updated it to M-58 9334.2.0 build, the app was not working.
Description: Show this description

Comment 4 by nath...@gmail.com, Mar 16 2017

Howdy, I'm the one who made the posts to the chromium-os-dev list and also submitted the feedback that this bug was generated from.

Just to be clear, this only seems to affect apps running in ARC++.  If I take the BLE Sphero, pair it with the laptop, and then run Sphero's SPRK Lightning Lab app for Chrome (https://chrome.google.com/webstore/detail/sprk-lightning-lab/hfiocchbmngcelgfdcfbepgoipapddlh?hl=en), that works just fine.

This makes me think the problem is not with the core OS bluetooth subsystem, but rather with the ARC++ Bluetooth Bridge.  However, I am not familiar with the code, so you all would know better than I.  Just thought I'd clarify.  Thanks.

Comment 5 by nath...@gmail.com, Mar 27 2017

FYI, I recently tested a Fitbit Blaze in both BLE Scanner and nRF Connect, and unlike the Sphero, neither app showed any advertised services for the device.  So something is going on other than just the Generic Access service coming up missing.

Also, perhaps this bug should be re-titled?
Status: pbathinichromium.org (was: Untriaged)
Summary: ARC++ BTLE broken for edgar. (was: Please see full chromium-os-dev discussion thread, linked to via URL field, for complete details.)
According to b/36072163, Fitbit is working correctly in Caroline 9334.28

Pramod, can you help looking at this with edgar?


Labels: -Pri-2 Pri-3
Look like ARC++ for edgar is not actually launched.

Mark this as P3 for now
Status: (was: pbathinichromium.org)
As, i check on 9334.28.0 build, ARC++ is not enabled for edgar. 


Comment 9 by nath...@gmail.com, Apr 3 2017

Yes it's true that ARC++ is not enabled for Edgar by default.  I forced its launch by using the --arc-availability flag.  Other than BLE, ARC++ seems to work fine on this model.

The policies for filing bugs against models that aren't yet supported for ARC++ (but which are on the roadmap) are not clear to me.  Since Edgar will eventually be a supported model, is there any reason to not start hunting down this bug before that happens, so that when ARC++ is finally enabled, the bug does not exist for everyone who is using it?
Status: WontFix
> Is there any reason to not start hunting down this bug before that happens, so that when ARC++ is finally enabled, the bug does not exist for everyone who is using it?

Android N ARC++ code is branched from M for a while and the code is quite different now. There are test team that test ARC++ M branch for supported mode to make sure that it work. But for feature development, we want to add a new code in N or future release branch only to avoid duplication work.

I will mark this as WontFix as it is bug unreleased platform.

Comment 11 by nath...@gmail.com, Apr 3 2017

I see, thanks; that makes sense.

So it sounds like we need to wait for N to appear on Edgar dev or canary branch and then re-test to see if issue is gone after that.  Any idea when N might show up on remaining models that still don't have it?
> Any idea when N might show up on remaining models that still don't have it?

Sorry, no idea about that.

Sign in to add a comment