995860 - [tarako] b2g process restart when make a call with BT headset connected

Status: RESOLVED FIXED

(Firefox OS Graveyard :: Bluetooth, defect, P1)

Description

[Peipei Cheng (needinfo if you need my action)]

Attachment: BT_Dialer_adb.log

Steps to reproduce
------------------------------------------------------------------
1. Pair and connect tarako to a BT headset
2. Make a MO call
   --> You will see b2g process restarts.

Attached is adb logcat. Test starts: 04-14 03:50:56.430

Test build
--------------------------------------------------------------------
URL: https://pvtbuilds.mozilla.org/pvt/mozilla.org/b2gotoro/nightly/mozilla-b2g/28_v1_3t-tarako/latest/
Gaia b2802627a974795ccba989cede0540f20fadc633
Gecko https://hg.mozilla.org/releases/mozilla-b2g28_v1_3t/rev/c8491a42d4e8
BuildID 20140413164001
Version 28.1
ro.build.version.incremental=eng.cltbld.20140413.210357

Comment 1

[Joe Cheng [:jcheng] (please needinfo)]

triage: 1.3T+, this is a common use case

Comment 2

[Peipei Cheng (needinfo if you need my action)]

It seems to be specific to BT headset. 
The one I used to reproduce this problem is Jabra BT2045. I also tried on another one Jabra Talk, it could not reproduce this issue.

Comment 3

[Peipei Cheng (needinfo if you need my action)]

(In reply to pcheng from comment #2)
> It seems to be specific to BT headset. 
> The one I used to reproduce this problem is Jabra BT2045. I also tried on
> another one Jabra Talk, it could not reproduce this issue.

Please ignore above comment. I reproduced again with Jabra Talk.

Comment 4

[Joe Cheng [:jcheng] (please needinfo)]

shawn, is this related to the other bug you are looking at? bug 994411?

Comment 5

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Joe Cheng [:jcheng] from comment #4)
> shawn, is this related to the other bug you are looking at? bug 994411?
Yes, for bug 994411 tarako part (v1.3t), but not master branch.
Can we discuss this bug for tarako and leave bug 994411 to master only?

Comment 6

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

It's clear that b2g was been killed due to low memory.

04-14 03:51:46.100 I/log     ( 2950): <4>0[ 2900.705203] lowmem_shrink select 2678 (Homescreen), adj 8, size 5309, to kill 
04-14 03:51:47.800 I/log     ( 2951): <4>0[ 2902.569964] lowmem_shrink select 2609 (b2g), adj 0, size 23166, to kill

Comment 8

[bhavana bajaj [:bajaj]]

reopening per #: https://bugzilla.mozilla.org/show_bug.cgi?id=994411#c20

Comment 9

[Joe Cheng [:jcheng] (please needinfo)]

triage: it seems like there is some problem in the Bluetooth firmware upgrade. to James for now. thanks

Comment 10

[Danny Liang [:dliang]]

This issue is caused by update of BT fw/binary. We cannot reproduce it w/ image on 3/31.

Comment 11

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Danny Liang [:dliang] from comment #10)
> This issue is caused by update of BT fw/binary. We cannot reproduce it w/
> image on 3/31.
Thanks for clarification. I hope we can know the root cause why memory is low after using bluetooth.

Comment 12

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Danny Liang [:dliang] from comment #10)
> This issue is caused by update of BT fw/binary. We cannot reproduce it w/
> image on 3/31.

Hi Danny,
Can you clarify how to do you make the problem

Comment 13

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

(In reply to Danny Liang [:dliang] from comment #10)
> This issue is caused by update of BT fw/binary. We cannot reproduce it w/
> image on 3/31.
Hi Danny,
Can you clarify how to do you have the conclusion that this bug caused by BT fw/binary update?

Comment 14

[Xinhe Yan]

(In reply to Shawn Huang [:shuang] [:shawnjohnjr] from comment #12)
> (In reply to Danny Liang [:dliang] from comment #10)
> > This issue is caused by update of BT fw/binary. We cannot reproduce it w/
> > image on 3/31.
> 
> Hi Danny,
> Can you clarify how to do you make the problem
device/sprd/partner/trout/bluetooth$ git log ./bt_ram_code.bin
commit 18303da7f8898cca861b451a8e8de89760d92d0e
Author: bing.li <[e-mail]>
Date:   Tue Apr 8 19:47:02 2014 +0800

        Bug#295133 merge the xunrui patch.
    
                    [bug number  ]295133
                    [root cause  ]
                    [changes     ]
                    [side effects]
                    [self test   ] smoke test ok.
                    [reviewers   ]
    
    Change-Id: I23c36e1a8cf2592b51da652884c486905956242a

You can revert this commit and push bt_ram_code.bin to system/lib/modules/

Comment 15

[Xinhe Yan]

Just talk with BT guy. Bluetooth only need set up SCO link in this case. 
bt_ram_code.bin is for controller. Controller running in a separate ARM core. 
So We need find out why and who consume memory in this case.

Comment 16

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Xinhe,
I'm not sure why revert bluetooth firmware could make differences. But I dig into bug 994411 (for master branch) and found out the root cause. I'm not sure this is the case for tarako, since revert bt_ram_code.bin indeed make memory consumption problem disappear. Does bt_ram_code.bin affect modem?

Comment 17

[Eric Chou [:ericchou] [:echou]]

(In reply to Shawn Huang [:shuang] [:shawnjohnjr] from comment #16)
> Xinhe,
> I'm not sure why revert bluetooth firmware could make differences. But I dig
> into bug 994411 (for master branch) and found out the root cause. I'm not
> sure this is the case for tarako, since revert bt_ram_code.bin indeed make
> memory consumption problem disappear. Does bt_ram_code.bin affect modem?

Xinhe,

Please try Shawn's patch (which is attached in bug 994411) with the latest bt firmware. Both Shawn and I consider that patch should be able to fix the issue w/o rolling back bt firmware version.

Comment 18

[Xinhe Yan]

Yes, that patch can fix this bug without revert firmware.

Can you rebase that patch to v1.3t branch?

Comment 19

[James Zhang (Spreadtrum)]

(In reply to Xinhe Yan from comment #18)
> Yes, that patch can fix this bug without revert firmware.
> 
> Can you rebase that patch to v1.3t branch?

Please land bug 994411 patch on v1.3t.

Comment 20

[Shawn Huang [:shawnjohnjr][:shuang] (as a happy gecko contributor)]

Attachment: Bug 995860 - [tarako] b2g process restart when make a call with BT headset connected, r=echou, a=1.3t+

Comment 21

[Eric Chou [:ericchou] [:echou]]

(In reply to Shawn Huang [:shuang] [:shawnjohnjr] from comment #20)
> Created attachment 8408200 [details] [diff] [review]
> Bug 995860 - [tarako] b2g process restart when make a call with BT headset
> connected, r=echou, a=1.3t+

Hi Sheriff,

Please notice that this is a 1.3T-specific patch. The patch for m-c has landed in bug 994411.

Thanks.

Comment 22

[Eric Chou [:ericchou] [:echou]]

> Please notice that this is a 1.3T-specific patch. The patch for m-c has

s/notice/note

Comment 23

[[:fabrice] Fabrice Desré]

https://hg.mozilla.org/releases/mozilla-b2g28_v1_3t/rev/9ff031daa73c

Comment 24

[James Zhang (Spreadtrum)]

Xinhe will verify it on my side and close it.

Comment 25

[Xinhe Yan]

Peipei,
this patch already merge to our branch and fix this bug.
Can you close this bug?

Comment 26

[Peipei Cheng (needinfo if you need my action)]

(In reply to Xinhe Yan from comment #25)
> Peipei,
> this patch already merge to our branch and fix this bug.
> Can you close this bug?

I will verify on mozilla build and the close it.

Comment 27

[Peipei Cheng (needinfo if you need my action)]

Verified on following build. MO and MT calls via BT headset could work now.

Gaia      f0872318d46781bb59d0a13a2e1fffb115b8ca2b
Gecko     https://hg.mozilla.org/releases/mozilla-b2g28_v1_3t/rev/9ff031daa73c
BuildID   20140417164004
Version   28.1
ro.build.version.incremental=eng.cltbld.20140417.201933
ro.build.date=Thu Apr 17 20:19:40 EDT 2014