962456 - [DSDS] When change non-primary data SIM Data APN setting, established data call will be disconnected first then connected which shouldn't be affected.

Status: RESOLVED WONTFIX

(Firefox OS Graveyard :: Gaia::Settings, defect)

Description

[Enpei]

Attachment: time stamp: 16:20, 1/22

* Build Number  
Fugu
Gaia      ee25b0e45649d2955f340ce4f71ad55712dd0fab
Gecko     913cf2b92845441c9578787362ddad6f2ac15df6
BuildID   20140121095108
Version   28.0a2

* Reproduce Steps
1. SIM 1 data call is established.
2. Go to SIM 2 Data settings, change it's APN setting and press ok.

* Expected Result
SIM 1 data call will not be affected.

* Actual Result
SIM 1 data call will be disconnected first then connected, the way like SIM 1 APN is changed.

* Occurrence rate
100%

Comment 1

[José Antonio Olivera Ortega [:jaoo]]

At a first glance it might not be related to a Gaia issue, see [1] please.

[1] https://github.com/mozilla-b2g/gaia/blob/master/apps/settings/js/carrier.js#L1089-L1107

Comment 2

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

We use only one key "ril.data.apnSettings" pointing to apn settings for 2 SIMs. Both radioInterface instances receive a notification of changes even only one SIM's apn is updated by gaia. In gecko, we don't compare if the apn content really becomes different after getting notified. We simply disconnect the data call if any, and update apn cache in gecko.

To resolve the issue, gecko should be able to tell which SIM the notification is sent for.

Possible solutions:
1) Use individual keys for each SIM
2) Still use one key, but gecko does more check to know which SIM's apn is really updated. The comparison looks expensive I'd say.

Comment 3

[José Antonio Olivera Ortega [:jaoo]]

(In reply to Hsin-Yi Tsai  [:hsinyi] from comment #2)
> Possible solutions:
> 1) Use individual keys for each SIM
> 2) Still use one key, but gecko does more check to know which SIM's apn is
> really updated. The comparison looks expensive I'd say.

Second solution seems so expensive so IMHO we should go for the first one. On the other hand I wonder if we could avoid call `updateRILNetworkInterface()` at [1] and only call it for 'ril.data.enabled' setting changes.

[1] http://mxr.mozilla.org/mozilla-central/source/dom/system/gonk/RadioInterfaceLayer.js#2884

Comment 4

[Jessica Jong [:jessica] (inactive)]

(In reply to José Antonio Olivera Ortega [:jaoo] from comment #3)
> (In reply to Hsin-Yi Tsai  [:hsinyi] from comment #2)
> > Possible solutions:
> > 1) Use individual keys for each SIM
> > 2) Still use one key, but gecko does more check to know which SIM's apn is
> > really updated. The comparison looks expensive I'd say.
> 
> Second solution seems so expensive so IMHO we should go for the first one.
> On the other hand I wonder if we could avoid call
> `updateRILNetworkInterface()` at [1] and only call it for 'ril.data.enabled'
> setting changes.
> 
> [1]
> http://mxr.mozilla.org/mozilla-central/source/dom/system/gonk/
> RadioInterfaceLayer.js#2884

Do you mean 'updateRILNetworkInterface()' after receiving 'ril.data.apnSettings' change? We do this because in 'updateApnSettings()' we will disconnect any connected data, so after apns are updated we need to call 'updateRILNetworkInterface()' to restore default data call if needed. But I noticed that gaia will also trigger 'ril.data.enabled' off and on. Maybe we should do this only in one side...

Comment 5

[José Antonio Olivera Ortega [:jaoo]]

(In reply to Jessica Jong [:jjong] [:jessica] from comment #4)

> > On the other hand I wonder if we could avoid call
> > `updateRILNetworkInterface()` at [1] and only call it for 'ril.data.enabled'
> > setting changes.
> > 
> > [1]
> > http://mxr.mozilla.org/mozilla-central/source/dom/system/gonk/
> > RadioInterfaceLayer.js#2884
> 
> Do you mean 'updateRILNetworkInterface()' after receiving
> 'ril.data.apnSettings' change? We do this because in 'updateApnSettings()'
> we will disconnect any connected data, so after apns are updated we need to
> call 'updateRILNetworkInterface()' to restore default data call if needed.

Sorry for asking this...I used to know the code for data calls in the RIL plumbing better. It's okay if you need to do it.

> But I noticed that gaia will also trigger 'ril.data.enabled' off and on.

I don't think so, as I reviewed Gaia doesn't disable/enable the data call if the APNs have been changed for the ICC not active for data calls. Am I missing something? Please see comment #1 please

Comment 6

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

(In reply to José Antonio Olivera Ortega [:jaoo] from comment #3)
> (In reply to Hsin-Yi Tsai  [:hsinyi] from comment #2)
> > Possible solutions:
> > 1) Use individual keys for each SIM
> > 2) Still use one key, but gecko does more check to know which SIM's apn is
> > really updated. The comparison looks expensive I'd say.
> 
> Second solution seems so expensive so IMHO we should go for the first one.
> On the other hand I wonder if we could avoid call
> `updateRILNetworkInterface()` at [1] and only call it for 'ril.data.enabled'
> setting changes.
> 
> [1]
> http://mxr.mozilla.org/mozilla-central/source/dom/system/gonk/
> RadioInterfaceLayer.js#2884

However, Option2 benefits a single sim case as well, see Bug 931348.

Comment 7

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

triage: user impact is quite low, put to backlog

Comment 8

[BMO Automation]

Firefox OS is not being worked on