880726 - [Leo] adding a quirk for correctly paring RIL parcel in call waiting scenario

Status: RESOLVED FIXED

(Core :: DOM: Device Interfaces, defect)

Description

[Etienne Segonzac (:etienne)]

Gaia: master
Gecko: b2g18 from yesterday

The call waiting is basically broken.
STR:
- call your voicemail
- from another phone, call the b2g phone

Expected:
- you see the call waiting UI displayed on the call screen

Actual:
- Nothing happens

I added the following logs:
- one in the oncallschanged handler in dialer.js
- one in the oncallschanged handler in oncall.js (attention screen, another window)

When I dial the first call, both oncallschanged get triggered (we open the call screen and display the voicemail call).

When the second call comes in, only the one in dialer.js gets triggered (so the callscreen does nothing).
When I hangup the second call (form the calling device), again, only the one in dialer.js is triggered (we do nothing here neither).

When I hangup the first call, both oncallschanged get triggered (so we close the call screen)

Comment 1

[Etienne Segonzac (:etienne)]

Asking leo? since it breaks the multiple call support but does not happen on older geckos.

Probably related to bu 823958.

Comment 2

[Yoshi Cheng-Hao Huang [:allstars.chh][:allstarschh][:yoshi]]

It works for me with the following version:

Gecko - hg mozilla-b2g18 119559:8e3f39363c54 (Jun 07 09:52:58 2013 -0700), DEBUG and OPT build
Gaia - mozillaorg/v1.0.1 d1d6753bb2018b9b8192e997ae5ee43e9dfb13e2 (Jun 7 03:47:43 2013 -0700) 
       mozillaorg/v1-train 1aed138dfef6279a00475f51ce0c491f32799ab6 (May 22 16:10:28 2013 -0500)

Comment 3

[Hsin-Yi Tsai (she/her) [:hsinyi]]

Hi Etienne, 

I just tested with latest mozilla-b2g18, and gaia master/v1-train/v1.0.1. All work for me as well. 

This issue looks interesting but seems not obviously related with bug 823958 to me, according to your description. I am also wondering if some messages between chrome and content got lost. Anyway, more investigation is needed. Since I cannot reproduce this issue, it would be helpful you provide more log info, with ril debugger enabled.

Comment 4

[Etienne Segonzac (:etienne)]

Attachment: Logcat

Some RIL logs + simple logs from the dialer app (when we get an oncallschanged) prefixed by "+++".

* for the first (outgoing) call we can see the oncallschanged in dialer.js (main app) and oncall.js (attention screen).
* for the second (incoming) call (line 254) we have only the oncallschanged in dialer.js

Comment 5

[Hsin-Yi Tsai (she/her) [:hsinyi]]

(In reply to Etienne Segonzac (:etienne) from comment #4)
> Created attachment 760935 [details]
> Logcat
> 
> Some RIL logs + simple logs from the dialer app (when we get an
> oncallschanged) prefixed by "+++".
> 
> * for the first (outgoing) call we can see the oncallschanged in dialer.js
> (main app) and oncall.js (attention screen).
> * for the second (incoming) call (line 254) we have only the oncallschanged
> in dialer.js

Thanks for the log! Could you also please tell me the gecko and gaia version you are testing with?

Comment 6

[Hsin-Yi Tsai (she/her) [:hsinyi]]

(In reply to Hsin-Yi Tsai [:hsinyi] from comment #5)
> (In reply to Etienne Segonzac (:etienne) from comment #4)
> > Created attachment 760935 [details]
> > Logcat
> > 
> > Some RIL logs + simple logs from the dialer app (when we get an
> > oncallschanged) prefixed by "+++".
> > 
> > * for the first (outgoing) call we can see the oncallschanged in dialer.js
> > (main app) and oncall.js (attention screen).
> > * for the second (incoming) call (line 254) we have only the oncallschanged
> > in dialer.js
> 
> Thanks for the log! Could you also please tell me the gecko and gaia version
> you are testing with?

Hmm... are you using QC RIL? I see some qcril debug messages, such as 'D/QCRIL_RPC(  148): Enter qcril_sms_cfg_event_callback.'

Comment 7

[Etienne Segonzac (:etienne)]

(In reply to Hsin-Yi Tsai [:hsinyi] from comment #6)
> > Thanks for the log! Could you also please tell me the gecko and gaia version
> > you are testing with?
> 
> Hmm... are you using QC RIL? I see some qcril debug messages, such as
> 'D/QCRIL_RPC(  148): Enter qcril_sms_cfg_event_callback.'

I just tested it again on a leo phone (originally saw the issue on my dogfooding buri):

gaia: 641ed2177c4a2ad92ce36c4d7c66227e1aa870aa (master from today)
gecko: b4f8f0a288ca (b2g18 tip from today)

What's the safest way to know for sure which RIL I'm using?

Comment 8

[Hsin-Yi Tsai (she/her) [:hsinyi]]

(In reply to Etienne Segonzac (:etienne) from comment #7)
> (In reply to Hsin-Yi Tsai [:hsinyi] from comment #6)
> > > Thanks for the log! Could you also please tell me the gecko and gaia version
> > > you are testing with?
> > 
> > Hmm... are you using QC RIL? I see some qcril debug messages, such as
> > 'D/QCRIL_RPC(  148): Enter qcril_sms_cfg_event_callback.'
> 
> I just tested it again on a leo phone (originally saw the issue on my
> dogfooding buri):
> 
> gaia: 641ed2177c4a2ad92ce36c4d7c66227e1aa870aa (master from today)
> gecko: b4f8f0a288ca (b2g18 tip from today)
> 

I still cannot reproduce it with latest code. And I'm using unagi.

gaia: master, adac4272383cc292f3c617ce7a482967f3471b24 
gecko: hg-mozilla-b2g18, 119578:bc472d357309

> What's the safest way to know for sure which RIL I'm using?

If QC RIL correctly loads, you can see RIL_QC_B2G in radio log section (logcat -b radio).

Comment 9

[Hsin-Yi Tsai (she/her) [:hsinyi]]

We have a parcel error in Etienne's log (as follows), which may be the root cause of problems with firing a callwaiting event. Seems we have a different parcel length in leo/buri.

I/Gecko   (  144): RIL Worker: We have at least one complete parcel.
I/Gecko   (  144): RIL Worker: Solicited response for request type 9, token 138, error 0
I/Gecko   (  144): RIL Worker: Handling parcel as REQUEST_GET_CURRENT_CALLS
I/Gecko   (  144): RIL Worker: Something's wrong, found string delimiter: 12
I/Gecko   (  144): RIL Worker: Parcel handling threw Error: Trying to read data beyond the parcel end!
I/Gecko   (  144): ensureIncomingAvailable@resource://gre/modules/ril_worker.js:226
I/Gecko   (  144): readUint8@resource://gre/modules/ril_worker.js:272
I/Gecko   (  144): readUint16@resource://gre/modules/ril_worker.js:294
I/Gecko   (  144): readString@resource://gre/modules/ril_worker.js:318
I/Gecko   (  144): REQUEST_GET_CURRENT_CALLS@resource://gre/modules/ril_worker.js:5495
I/Gecko   (  144): handleParcel@resource://gre/modules/ril_worker.js:5388
I/Gecko   (  144): processParcel@resource://gre/modules/ril_worker.js:617
I/Gecko   (  144): processIncoming@resource://gre/modules/ril_worker.js:562
I/Gecko   (  144): onRILMessage@resource://gre/modules/ril_worker.js:9441

Comment 10

[Hsin-Yi Tsai (she/her) [:hsinyi]]

Attachment: PR to device-leo

Adding a quirk on leo device

Comment 11

[Hsin-Yi Tsai (she/her) [:hsinyi]]

The gecko change has been handled in bug 861896. Once the PR to device-leo is merged, we are fine here.

Comment 12

[Michael Wu [:mwu]]

Comment on attachment 761928 [details]
PR to device-leo

Will we also need this for hamachi/buri?

Comment 13

[Michael Wu [:mwu]]

I just merged. Not resolving though, since I think we also need this for hamachi/buri.

Comment 14

[Hsin-Yi Tsai (she/her) [:hsinyi]]

(In reply to Michael Wu [:mwu] from comment #13)
> I just merged. Not resolving though, since I think we also need this for
> hamachi/buri.

I don't have hamachi/buri on hand so not able to test them. That's why the bug description focuses on "Leo." ;-) 

If the root cause is the same on hamachi and buri, then I think we will need them.

Comment 15

[Ben Turner (not reading bugmail, use the needinfo flag!)]

We should probably open a separate bug to make sure that RIL parcel problems are easily recognized and reported to us, maybe with a special crash report?

Comment 16

[Michael Wu [:mwu]]

I will test then.

Comment 17

[Hsin-Yi Tsai (she/her) [:hsinyi]]

(In reply to Michael Wu [:mwu] from comment #16)
> I will test then.

Thanks, Michael.

I am going to close this bug. Let's file separate bugs if you encounter problems in hamachi/buri. Does that sound good?

Comment 18

[Michael Wu [:mwu]]

It's the same fix. No reason to open a new bug. We'll also want to uplift this so this will reduce the number of leo? requests.

Comment 19

[Michael Wu [:mwu]]

I can't seem to get call waiting to work with or without the quirk. Let's go for a new bug.

Comment 20

[bhavana bajaj [:bajaj]]

Blocking as this blocks a blocker 861896.

Comment 21

[bhavana bajaj [:bajaj]]

Marking status-b2g:18 as fixed as nothing else is needed here for v1.1