907628 - [B2G][Helix][STK][wangchao]Sometimes it do not display STK Menu in Setting.

Status: RESOLVED INCOMPLETE

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

Description

[wangchao]

Attachment: no stk log.txt

User Agent: Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.1; Trident/4.0; SLCC2; .NET CLR 2.0.50727; .NET CLR 3.5.30729; .NET CLR 3.0.30729; Media Center PC 6.0; .NET4.0C; .NET4.0E; Tablet PC 2.0)

Steps to reproduce:

1.Insert a card and boot.
2.Open setting to see STK menu if exists.(log in attachment)

I think this is a existed problem. But it is better for you to fixed it in this ver: HD v1.1.


Actual results:

1.No STK Menu


Expected results:

1.There is a STK Menu.

Comment 1

[lecky]

I caputured another log, it indicates that  init: function icc_init()
//in icc.js 
is init after SetUp Menu Proactive command, so it missed to process the command.

This caused there is no stk menu in setting.

Thanks!

Comment 2

[lecky]

Attachment: show the icc is init after receive setup menu command.

Comment 3

[Evelyn Hung]

It seems a event racing problem.
William, can we reproduce it?

Comment 4

[William Hsu [:whsu]]

Hi, Evelyn,

I think I need more information to reproduce this problem because I don't know which SIM card and build partner used.

So, I guess that Lecky meant "sometime" the STK menu cannot be displayed after installing the "same" SIM card.
As I briefly tested this problem on partner released build(20130814), I cannot reproduce it.
STK menu is always displayed with CHT SIM card.

But, I found a interesting thing. The STK menu is only displayed while installing the specific SIM cards.
I test the following SIM cards. Result as attachment
(1) CHT(中華電信)：Display STK menu
(2) USIM (中国联通)：Display STK menu
(3) AT&T：No STK menu
(4) T-MOBILE：No STK menu

I'm not sure if the behavior relate to this problem.

Comment 5

[William Hsu [:whsu]]

Attachment: STK menu

Comment 6

[William Hsu [:whsu]]

Hi, Wangchao,

Could I know which build and SIM card you tested?
One more question, do you mean if you install the same SIM card on the device, the STK menu will sometimes disappear?
Thank you!

Comment 7

[lecky]

We used build(20130814) and used our lib test usim card.

The test usim card are used for all device: Android, WP, FFOS and ect. It is a normal card.

And just as you said, if I install the same SIM card on the device, the STK menu will sometimes disappear. 

Thanks!

Comment 8

[William Hsu [:whsu]]

Hi,Lecky,

Thanks for the reminder and information.
O~I see! I will do the second test.
Thank you.


--------------------------------------------------
Needinfo myself to remind me to handle this case.

Comment 9

[lecky]

Hi William:

   Is there any progress? This is a High Probability problem. It affact our test for STK.

Thanks!

Comment 10

[lecky]

Next week, our device will be sent to customer.

We afraid this problem do give them a big shock and make them cat not tolerate.

Please pay more attention to this.

If you need Information, please tell me, i will try my best effort to assist you.

Thanks!

Comment 11

[William Hsu [:whsu]]

Hi, Lecky,

After I traced this problem, I found this problem seems to cause by your test build.(20130814).
Because I "sometimes" can reproduce this problem on your test build by manually test.
But I cannot reproduce this case on our (Mozilla) build.(20130905) by manually test and automation test.

Please test this case on the latest build to see whether the problem still happens.
Thank you!

Comment 12

[William Hsu [:whsu]]

Please reopen it if you still can reproduce this case on the latest build.

Comment 13

[lecky]

Hi William:

I use the 2013-09-11 build(gaia gecko and gonk-misk come from mozilla), it still exist.

I try 10 times, reproduce 1.

 
Please see the attachment.

Comment 14

[lecky]

Attachment: screenshot and adb log

Comment 15

[lecky]

 Hi William:

I test it on 2013-9-11 Helix branch, v1.1HD.

Do you used helix branch to test before or used master?

Thanks!

Comment 16

[William Hsu [:whsu]]

Hi, Evelyn,

Do you know who can help on this problem?
I cannot reproduce this case.

Is it possible to trace the log to narrow down the problem?

---------------------------------------------------------------
Hi, Lecky,

Could you please specifically describe your test steps?
I want to double check our test steps.
Thank you.

My test steps are as below,
1. Insert USIM card
2. Power on the device
3. Launching the settings app
4. Check the "Operator Services" whether it has "USIM卡應用"
5. Long press power key
6. Tap "Restart" option to restart the device.
7. Repeat the step3~step6

My SIM card is "WO 128k USIM".

Comment 17

[lecky]

Hi Evelyn:
   My test steps are exactly as yours.

   My build is also a mozilla build, a 2013-9-11 user build with no differences with mozilla git(gaia, gecko, gonk-misc).

   From the adb log I uploaded, you can see:

//it show the QCContentHelper_QC_B2G has send SET_UP_MENU command to content process, but receive no reponse. Expected result is : a ok reponse is received to say SET_UP_MENU is correctly executed.

I/Gecko   (  158): -*- QCContentHelper_QC_B2G: sendMessage to content process: RIL:StkCommand{commandNumber : 1,                         typeOfCommand : 37,                         commandQualifier : 0, options : {title : 'STK命令测试', items:[{identifier : 1,text : 'Proactive SIM',nai : ''},{identifier : 2,text : 'Data download to SIM',nai : ''},{identifier : 3,text : 'Call Control by SIM',nai : ''},{identifier : 4,text : 'Event Download',nai : ''},{identifier : 5,text : 'Timer Expiration',nai : ''}], presentationType : 0}}

I/Gecko   (  158): -*- QCContentHelper_QC_B2G: Enqueue message RIL:StkCommand
   
Thanks!

Comment 18

[William Hsu [:whsu]]

Hi, Lecky,

Sorry! My bad. I forgot your RIL was QC-RIL.
This seems a QC-RIL's problem. I will contact Shawn to address this problem.
I will reply you soon. Thanks!

Comment 19

[shawn ku [:sku]]

Hi William:
  We should focus on sim/radio status on this issue first.
It looks to me that SIM is not in READY state, therefore there is no reportStkServiceIsRunning to cause stk setup menu event.

Hi Wayne:
  SR might be needed for partner to check this issue in parallel.


E/ICCIO_QC_B2G(  158): HandleGetSimStatus received error 1
E/ICCIO_QC_B2G(  158): HandleGetSimStatus received error 1
I/Gecko   (  158): -*- QCContentHelper_QC_B2G: sendMessage to content process: RIL:StkCommand{commandNumber : 1,                         typeOfCommand : 37,                         commandQualifier : 0, options : {title : 'STK命令测试', items:[{identifier : 1,text : 'Proactive SIM',nai : ''},{identifier : 2,text : 'Data download to SIM',nai : ''},{identifier : 3,text : 'Call Control by SIM',nai : ''},{identifier : 4,text : 'Event Download',nai : ''},{identifier : 5,text : 'Timer Expiration',nai : ''}], presentationType : 0}}
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received error 2 for EF 0x6fc7
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received error 2 for EF 0x6fca
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received error 2 for EF 0x6fcb
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received error 2 for EF 0x6f50
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received error 2 for EF 0x6fcd
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received error 2 for EF 0x6fc6
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received error 2 for EF 0x6fc5
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received error 2 for EF 0x6f46

Thanks!!
sku

Comment 20

[lecky]

 (In reply to shawn ku [:sku] from comment #19)
> Hi William:
  We should focus on sim/radio status on this issue first.
It
> looks to me that SIM is not in READY state, therefore there is no
> reportStkServiceIsRunning to cause stk setup menu event.

Hi Wayne:
  SR
> might be needed for partner to check this issue in parallel.

> E/ICCIO_QC_B2G(  158): HandleGetSimStatus received error 1
E/ICCIO_QC_B2G( 
> 158): HandleGetSimStatus received error 1
I/Gecko   (  158): -*-
> QCContentHelper_QC_B2G: sendMessage to content process:
> RIL:StkCommand{commandNumber : 1,                         typeOfCommand :
> 37,                         commandQualifier : 0, options : {title :
> 'STK命令测试', items:[{identifier : 1,text : 'Proactive SIM',nai :
> ''},{identifier : 2,text : 'Data download to SIM',nai : ''},{identifier :
> 3,text : 'Call Control by SIM',nai : ''},{identifier : 4,text : 'Event
> Download',nai : ''},{identifier : 5,text : 'Timer Expiration',nai : ''}],
> presentationType : 0}}
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received
> error 2 for EF 0x6fc7
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received
> error 2 for EF 0x6fca
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received
> error 2 for EF 0x6fcb
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received
> error 2 for EF 0x6f50
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received
> error 2 for EF 0x6fcd
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received
> error 2 for EF 0x6fc6
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received
> error 2 for EF 0x6fc5
E/ICCFILEHANDLER_QC_B2G(  158): HandleSimIO received
> error 2 for EF 0x6f46

Thanks!!
sku

Hi shawn ku:

   I think b2g_ril is ok, the OK log shows the same as you point out. I will update it. You can compare them. 

   For no stk menu log, it says the command has 
I/Gecko   (  158): -*- QCContentHelper_QC_B2G: Enqueue message RIL:StkCommand.

  Maybe it is a 'Enqueue' problem, or The STK SERVICE is not registered, so the message can not deliver to right object.

Thanks!

Comment 21

[lecky]

Attachment: show stk log for the same sim card

Comment 22

[Evelyn Hung]

Looks like a RIL problem that I can't help. :(

Comment 23

[lecky]

Hi Evelyn Hung:

   I have done a test, just let b2g_ril:stkservice sleep 5 seconds. Then test for 20 times in our own build. It did not happened again. I 80% confirm it's a racing problem.

   At current version, b2g did not know whether SETTINGS APP is launghed and send SET_UP_MENU COMMAND to SETTINGS. But sometimes, SETTINGS APP is not launghed and do not receive the COMMAND.
  
   I think in gaia and gecko, this is no mechanism to confirm this. And b2g_ril also can not know whether the APP in gaia is ready. So APP in gaia should make some protection in case the message is lost.

   This is a guess according to present phenomenon and logs. I think you must know more about the gaia and all the possiable reasons.

Thanks!

Comment 24

[shawn ku [:sku]]

Hi Lecky:
 Frankly, I can not see set_up_menu in "screenshot and adb log", maybe just did not capture the valid information within it. Please correct me if anything is wrong.

Hi Wayne:
 Per talk, please follow up this bug with partner. Any help needed, please kindly let me know.


// Log from "show stk log for the same sim card"
I/Gecko   (  158): -*- QCContentHelper_QC_B2G: sendMessage to content process: RIL:StkCommand{commandNumber : 1,                         typeOfCommand : 37,                         commandQualifier : 0, options : {title : 'STK命令测试', items:[{identifier : 1,text : 'Proactive SIM',nai : ''},{identifier : 2,text : 'Data download to SIM',nai : ''},{identifier : 3,text : 'Call Control by SIM',nai : ''},{identifier : 4,text : 'Event Download',nai : ''},{identifier : 5,text : 'Timer Expiration',nai : ''}], presentationType : 0}}

Comment 25

[Wayne Chang [:wchang]]

+ Anshul, M1

Hi Anshul, Michael,

This seems to only occur on commercial ril and Shawn has checked both successful and unsuccessful logs to find the set_up_menu missing in the fail logs.

Comment 26

[Michael Vines [:m1] [:evilmachines]]

Thanks Wayne.  We will certainly support via SR.

Comment 27

[lecky]

Hi sku:
   typeOfCommand : 37 is just the SET_UP_MENU command.

   IN "screenshot and adb log", also 
//-*- QCContentHelper_QC_B2G: sendMessage to content process: RIL:StkCommand{commandNumber : 1,                         typeOfCommand : 37,  

   SO i think commercial ril have send SET_UP_MENU to content process, but content process does not save and handle it or miss the message.

Thanks!

Comment 28

[shawn ku [:sku]]

Hi Lecky:
 Thanks for your information.
sku

Comment 29

[Wayne Chang [:wchang]]

HD+'ing this assume confirmed and helix only.

Nivi, once you have identify the problem can you confirm if this is helix specific?

Comment 30

[Carol Yang [:cyang]]

(In reply to lecky from comment #27)
> Hi sku:
>    typeOfCommand : 37 is just the SET_UP_MENU command.
> 
>    IN "screenshot and adb log", also 
> //-*- QCContentHelper_QC_B2G: sendMessage to content process:
> RIL:StkCommand{commandNumber : 1,                         typeOfCommand :
> 37,  
> 
>    SO i think commercial ril have send SET_UP_MENU to content process, but
> content process does not save and handle it or miss the message.

As Lecky already mentioned here, commercial RIL already sent the SET_UP_MENU to content process.

Fernando, can you please have a look at this? See if this is indeed a race condition?

Comment 31

[lecky]

Hi Wayne:

   As Carol Yang said: he think this is a race condition.

   As you said: FFOS used its own ril can not reproduce this. I think maybe this is a adapt issue.

We did not consider the sequence for different module, so the communication message may sometimes lost.

  This is a high rate issue and easily to reproduce.

  Can you pay some attention to this?

Thanks!

Comment 32

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

(In reply to Carol Yang [:cyang] from comment #30)
> (In reply to lecky from comment #27)
> > Hi sku:
> >    typeOfCommand : 37 is just the SET_UP_MENU command.
> > 
> >    IN "screenshot and adb log", also 
> > //-*- QCContentHelper_QC_B2G: sendMessage to content process:
> > RIL:StkCommand{commandNumber : 1,                         typeOfCommand :
> > 37,  
> > 
> >    SO i think commercial ril have send SET_UP_MENU to content process, but
> > content process does not save and handle it or miss the message.
> 
> As Lecky already mentioned here, commercial RIL already sent the SET_UP_MENU
> to content process.
> 
> Fernando, can you please have a look at this? See if this is indeed a race
> condition?

Hi,

If Gaia receives the message it is stored in mozSettings database which is checked when settings is opened.

It never failed to me, but can be failing here. How can I reproduce it?

Comment 33

[shawn ku [:sku]]

Hi Wangchao:
 To better understand this issue, could you pleae try to re-get the log with time stamp and radio inforamtion for us to check?

adb command:
adb logcat -b radio -b main -v threadtime > /tmp/907628.log

and attach 907628.log to bugzilla.

Note:
Please make sure radio log is available, and timestamp is enabled (to check the timing between set_up_menu and sim ready event).

Thanks!!
sku

Comment 34

[lecky]

Attachment: 907628.log

Comment 35

[lecky]

Hi shawn ku:

   I have uploaded the log. Could you take a look?

   Thanks!

Comment 36

[shawn ku [:sku]]

Hi Lecky:
 SETUP menu comes up at 10-08 15:56:06.279, and sim ready is around 10-08 15:56:06.609.
That means SETUP menu via STK proactive command just come to gecko too early.


Besides, I did not see any command - REQUEST_REPORT_STK_SERVICE_IS_RUNNING was triggered in log. it's weird to me.


From AOSP/FFOS design, there should be a REQUEST_REPORT_STK_SERVICE_IS_RUNNING after SIM ready event due to RILD was brought up very early. RILD should report caching SETUP menu after REQUEST_REPORT_STK_SERVICE_IS_RUNNING request to avoid missing message in RIL.java (for AOSP) and ril_worker.js (for FFOS).


In short, this is a timing issue to me.
Please correct me if anything is wrong.

cheers,
sku



10-08 15:56:06.279   161   286 D RIL_QC_B2G: [UNSL]< UNSOL_STK_PROACTIVE_COMMAND {D081A2810301250082028182850F80005500530049004D53615E9475288F0A01807CBE54C163A883508F0A028059296C14988462A58F0A0380822A73ED67E58BE28F0A048051FA884C630753578F0A05804F53575B5FEB8BAF8F0C068065B095FB65E9665A62A58F0A0780624B673A97F34E508F0A0880624B673A96058BFB8F0A0980624B673A90AE7BB18F0A0A8077ED4FE152A974068F0C0B80624B673A84254E1A5385}
10-08 15:56:06.279   161   286 D RIL_QC_B2G: [0003]< GET_SIM_STATUS fails by RadioNotAvailable(1)
10-08 15:56:06.289   161   286 D RIL_QC_B2G: [0004]< GET_SIM_STATUS fails by RadioNotAvailable(1)
10-08 15:56:06.449   161   286 D RIL_QC_B2G: [0005]< GET_SIM_STATUS fails by RadioNotAvailable(1)
10-08 15:56:06.529   161   161 I Gecko   : -*- QCContentHelper_QC_B2G: sendMessage to content process: RIL:StkCommand{commandNumber : 1,                         typeOfCommand : 37,                         commandQualifier : 0, options : {title : 'USIMå¡åºç¨', items:[{identifier : 1,text : 'ç²¾åæ¨è',nai : ''},{identifier : 2,text : '天æ°é¢æ¥',nai : ''},{identifier : 3,text : 'èªç­æ¥è¯¢',nai : ''},{identifier : 4,text : 'åºè¡æå',nai : ''},{identifier : 5,text : 'ä½å快讯',nai : ''},{identifier : 6,text : 'æ°é»æ©ææ¥',nai : ''},{identifier : 7,text : 'ææºé³ä¹',nai : ''},{identifier : 8,text : 'ææºé读',nai : ''},{identifier : 9,text : 'ææºé®ç®±',nai : ''},{identifier : 10,text : 'ç­ä¿¡å©ç',nai : ''},{identifier : 11,text : 'ææºè¥ä¸å',nai : ''}], presentationType : 0}}

10-08 15:56:06.609   161   286 D RIL_QC_B2G: [0013]< GET_SIM_STATUS {CardState=CARDSTATE_PRESENT,PinState=PINSTATE_UNKNOWN,GsmUmtsSub=0,CdmaSub=-1,mImsSub=-1,num_apps=1,{app_type=APPTYPE_USIM,app_state=APPSTATE_READY,substate=PERSOSUBSTATE_READY,aid=a0000000871002ff86ffff89ffffffff,app_label=556e6976657253494d,pin1_replaced=0,pin1=PINSTATE_DISABLED,pin1_num_retries=0,puk1_num_retries=0,pin2=PINSTATE_ENABLED_NOT_VERIFIED,pin2_num_retries=0,puk2_num_retries=0}}
10-08 15:56:06.709   161   161 I Gecko   : -*- QCContentHelper_QC_B2G: Enqueue message RIL:StkCommand

10-08 15:58:53.050   161   286 D RIL_QC_B2G: [UNSL]< UNSOL_STK_SESSION_END
10-08 15:58:53.050   161   161 I Gecko   : -*- QCContentHelper_QC_B2G: sendMessage to content process: RIL:StkSessionEndnull

Comment 37

[shawn ku [:sku]]

Hi Carol:
 Could you please help check commercial RIL side to see if the flow meet what I said?
And, things about REPORT_STK_SERVICE_IS_RUNNING also.

Thanks!!
sku

Comment 38

[Carol Yang [:cyang]]

A couple of things I'm checking here and will update when I hear back.

Comment 39

[Carol Yang [:cyang]]

Hi Shawn,

Thanks for the analysis! Quick question:

> From AOSP/FFOS design, there should be a REQUEST_REPORT_STK_SERVICE_IS_RUNNING after SIM ready event due to RILD was brought up very early.

There seems to still be a possibility that even after SIM is ready, Gaia is not ready for the STK menu. Instead of relying on SIM being ready, should there be some kind of indication coming from Gaia to say that UI is ready to receive?

Comment 40

[shawn ku [:sku]]

Hi Carol:
 Yes, your suggestion might be a possible solution.
However, there is no such symptom in Moz RIL build be found so far.
Please dig out the real root cause for this case on comm. RIL.

Besides, I don't think a new interface for gaia to get setup menu is an easy way in v1.1.

Thanks!!
sku

Comment 41

[lecky]

Hi shawn ku:

Will this issue be fixed in V1.1HD? It is very important and will affact us TA. 

Thanks!

/////////////////////

To Beatriz:

If this is really can not be fixed in v1.1HD?
Will you accept this?

Thanks!

Comment 42

[Carol Yang [:cyang]]

Closing this bug as we are already supporting this through our SR system.

Comment 43

[Beatriz Rodríguez [:brg]]

Thanks Carol.

Comment 44

[Nivi]

Hi sku,

I analyzed this issue on both Mozilla RIL and commercial RIL. On Mozilla RIL, the fix is to delay sending the STK_SERVICE_IS_RUNNING arbitrarily with the assumption that gaia would be ready at that point of time but there's no guarantee that it will be. We still have risks with this in the case where the message is sufficiently delayed in the ril and we open the settings, the message has still not being sent to Gaia. Then the menu won't show up in Settings either. There is a race condition happening here. 

The correct fix, imho, will be for Gaia to let us know by sending say a -"StkMessageListenerReady". Then, both Mozilla ril and commercial RIL could register for that message and on receiving this, send the STK_SERVICE_IS_RUNNING to the modem. That would trigger the proactive commands to be sent that were queued and settings would behave correctly and show the menu. Please let me know if you have any questions/concerns.

Comment 45

[shawn ku [:sku]]

Hi Nivi:
 Basically I agree with you that might be timing issue for such case to be hit. However, I am not sure if any other conditions might run into that result in gecko cache the set_up menu w/o sending it to gaia forever.

Hi Yoshi:
 Could I have your comment/suggestion?

Thanks!!
sku

Comment 46

[Beatriz Rodríguez [:brg]]

Reopening this bug after comment 44 proposal. Feel free to re-close it if the suggestion is not feasible.

Comment 47

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

CC Ken and Georgia for this.

Comment 48

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

STK_SERVICE_IS_RUNNING is used to notify modem, RIL is ready for receiving STK proactive commands.

When RIL receives any proactive command, it will send system-message to Gaia.
And system-message will be responsible to delieve the message to Gaia, even Gaia hasn't completely started yet.

Comment 49

[Wayne Chang [:wchang]]

How do we progress here?

I think what's suggested in comment 44 may need additional architectural change if i understand it correctly.

For the time being, can we use a delay in com_ril to correct/reduce the problem at hand? Given the same delay seems to reduce it enough to not occur on moz ril.

Comment 50

[shawn ku [:sku]]

Hi Wayne:
 Yoshi already reply this with enough informtion.
systemMessenger will be used to send message to gaia.

if Gaia is not ready, that mechanism will cache the the message, and re-send it after app ready.
Hence, there should be no need to add extra interface for STK setup menu.


Besides, Did commercial RIL update why they send up setup menu before SIM ready?
And, here comes the new question, 
Did commercial RIL use SystemMessenger to notify stk proactive command?


2131   handleStkProactiveCommand: function handleStkProactiveCommand(message) {      
2132     if (DEBUG) this.debug("handleStkProactiveCommand " + JSON.stringify(message));
2133     gSystemMessenger.broadcastMessage("icc-stkcommand", message);               
2134     gMessageManager.sendIccMessage("RIL:StkCommand", this.clientId, message);   
2135   },

Comment 51

[Anshul]

(In reply to shawn ku [:sku] from comment #50)
> Hi Wayne:
>  Yoshi already reply this with enough informtion.
> systemMessenger will be used to send message to gaia.
> 
> if Gaia is not ready, that mechanism will cache the the message, and re-send
> it after app ready.
Are you suggesting that this mechanism is already present or that the mechanism to cached and resend needs to be added?
> Hence, there should be no need to add extra interface for STK setup menu.
> 
> 
> Besides, Did commercial RIL update why they send up setup menu before SIM
> ready?
> And, here comes the new question, 
> Did commercial RIL use SystemMessenger to notify stk proactive command?
Yes we have always been using system message manager to notify of the stk command in addition to sending the iccmessage.

> 
> 
> 2131   handleStkProactiveCommand: function
> handleStkProactiveCommand(message) {      
> 2132     if (DEBUG) this.debug("handleStkProactiveCommand " +
> JSON.stringify(message));
> 2133     gSystemMessenger.broadcastMessage("icc-stkcommand", message);      
> 
> 2134     gMessageManager.sendIccMessage("RIL:StkCommand", this.clientId,
> message);   
> 2135   },

Comment 52

[Anshul]

(In reply to Wayne Chang [:wchang] from comment #49)
> How do we progress here?
> 
> I think what's suggested in comment 44 may need additional architectural
> change if i understand it correctly.
> 
> For the time being, can we use a delay in com_ril to correct/reduce the
> problem at hand? Given the same delay seems to reduce it enough to not occur
> on moz ril.
We have already tried that and notice that it doesn't always work and looking for a better solution than just adding arbitrary delays.

Comment 53

[shawn ku [:sku]]

(In reply to Anshul from comment #51)
> (In reply to shawn ku [:sku] from comment #50)
> > Hi Wayne:
> >  Yoshi already reply this with enough informtion.
> > systemMessenger will be used to send message to gaia.
> > 
> > if Gaia is not ready, that mechanism will cache the the message, and re-send
> > it after app ready.
> Are you suggesting that this mechanism is already present or that the
> mechanism to cached and resend needs to be added?

Already present.

> > Hence, there should be no need to add extra interface for STK setup menu.
> > 
> > 
> > Besides, Did commercial RIL update why they send up setup menu before SIM
> > ready?

How about this one?

> > And, here comes the new question, 
> > Did commercial RIL use SystemMessenger to notify stk proactive command?
> Yes we have always been using system message manager to notify of the stk
> command in addition to sending the iccmessage.

We need more information (enable more log) to see what went wrong then.


> 
> > 
> > 
> > 2131   handleStkProactiveCommand: function
> > handleStkProactiveCommand(message) {      
> > 2132     if (DEBUG) this.debug("handleStkProactiveCommand " +
> > JSON.stringify(message));
> > 2133     gSystemMessenger.broadcastMessage("icc-stkcommand", message);      
> > 
> > 2134     gMessageManager.sendIccMessage("RIL:StkCommand", this.clientId,
> > message);   
> > 2135   },

Comment 54

[Michael Vines [:m1] [:evilmachines]]

(In reply to shawn ku [:sku] from comment #53)
> (In reply to Anshul from comment #51)
> > (In reply to shawn ku [:sku] from comment #50)
> > > Hi Wayne:
> > >  Yoshi already reply this with enough informtion.
> > > systemMessenger will be used to send message to gaia.
> > > 
> > > if Gaia is not ready, that mechanism will cache the the message, and re-send
> > > it after app ready.
> > Are you suggesting that this mechanism is already present or that the
> > mechanism to cached and resend needs to be added?
> 
> Already present.

Can you please point to that part of Gecko.  With that info, we can check here to confirm it's working as expected.

Comment 55

[shawn ku [:sku]]

Hi Michael/Anshul:
 Plesae check broadcastMessage@gecko/dom/messages/SystemMessageInternal.js

    // Buffer system messages until the webapps' registration is ready,
    // so that we can know the correct pages registered to be broadcasted.
    if (!this._webappsRegistryReady) {
      this._bufferedSysMsgs.push({ how: "broadcast",
                                   type: aType,
                                   msg: aMessage });
      return;
    }
    ...

I will also check the flow locally. 
Any finding, please kinbdly share with me.

Thanks!!
sku

Comment 56

[shawn ku [:sku]]

Hi Michael/Anshul:
 I did a local trial, and see re-broadcast stk setup menu after 'webapps-registry-read'.
This cache/forward mechansim should work properly.

Please share the thing you found if any.
Thanks!!
sku

10-31 09:15:28.337   107   107 I Gecko   : before ready - {"commandNumber":1,"typeOfCommand":37,"commandQualifier":0,"rilMessageType":"stkcommand","options":{"title":"中華電信區","items":[{"identifier":1,"text":"emome理財"},{"identifier":2,"text":"emome加值"}],"presentationType":0}}

10-31 09:15:32.411   107   107 I Gecko   : webapps-registry-ready

10-31 09:15:32.411   107   107 I Gecko   : this._bufferedSysMsgs - [{"how":"broadcast","type":"icc-stkcommand","msg":{"commandNumber":1,"typeOfCommand":37,"commandQualifier":0,"rilMessageType":"stkcommand","options":{"title":"中華電信區","items":[{"identifier":1,"text":"emome理財"},{"identifier":2,"text":"emome加值"}],"presentationType":0}}}]

10-31 09:15:32.411   107   107 I Gecko   : re-send broadcast message - {"commandNumber":1,"typeOfCommand":37,"commandQualifier":0,"rilMessageType":"stkcommand","options":{"title":"中華電信區","items":[{"identifier":1,"text":"emome理財"},{"identifier":2,"text":"emome加值"}],"presentationType":0}}

Comment 57

[Michael Vines [:m1] [:evilmachines]]

Thanks so much for checking Shawn!  Looks like we have a fix for this now, so there's no action that I see here and therefore closing the bug.

Comment 58

[lecky]

Using the build with the latest the commerical ril(AU250) and Gaia, i found a issue:
Settings is not listen to the change the 'icc.applications', and should reopen Settings to get the STK Menu 

Steps to reproduce:

Repro step:
1.when device boot, open the Settings, to see if the STK menu is exist.

Expected:
1.there is STK Menu.

Actual:
1.STK Menu is missing and will not show up. 
2.But only if kill the Settings and reopen it, can see the STK menu.

My opinion:
1.the first time i open Settings, 'icc.applications' is not cached yet, it can not get the STK Menu.
2.After the Settings is opened, 'icc.applications' is cached, but Settings is not listen to the change and do nothing, so STK Menu can not show up.
3.Only kill Settings and reopen it, it read the 'icc.applications', then STK Menu shows up. 

I think Settings should listen to the change of 'icc.applications' if Settings is already opened.

Thanks

Comment 59

[wangchao]

Attachment: no_stk_first_but_show_up_after_reopen.txt

Comment 60

[wangchao]

Since sometimes the stk menu is still can not show up. 

Reopen it.

Comment 61

[Michael Vines [:m1] [:evilmachines]]

Please see bug 933835 for how to move this issue forward.  Thanks!