rilmodem: Do not always use sim's app in status->gsm_umts_index

[Ratchanan Srirattanamet]

In driver/rilmodem/sim.c, function sim_status_cb, there is a line that set variable search_index to status->gsm_utms_index.

The code then checks if the card_state returned by the GET_SIM_STATUS reply is PRESENT and search_index is less than num_apps. If both are true, search_index is used to index the returned reply_sim_app array ( which leads to undefined behavior if the index is negative ) and then passed to configure_active_app().

The code should check search index, and if negative and num_apps >=1, set search_index to 0.

ofono version 1.12.bzr6888+15.04.20150224 in Ubuntu 15.04
Device: LG L90 Dual (I'm doing a port.)

[Tony Espy (awe)]

Proposed fix in:

https://github.com/rilmodem/ofono/pull/166

[Alfonso Sanchez-Beato (alfonsosanchezbeato)]

Looking at the definition of the reply to RIL_REQUEST_GET_SIM_STATUS in

http://androidxref.com/4.4.2_r1/xref/hardware/ril/include/telephony/ril.h

I see that if RIL_CardStatus_v6::gsm_umts_subscription_app_index value in the parcel is -1, it *should* mean that there is no GSM/UMTS application in the SIM card. However, this looks broken in many implementations (I get, for instance, index 8 for "cdma_subscription_app_index" with one phone, although "num_applications" is just one. The phone is not CDMA, so I guess the card is from a GSM operator and really has a SIM/USIM application.

Maybe the right approach here is to not use the indexes in RIL_CardStatus_v6 and look directly at the array RIL_CardStatus_v6::applications, as each of them has an application type.

@Ratchanan, would it be possible to obtain the full hex traces from you phone? To get them please do as root in the phone:

# stop ofono
# OFONO_RIL_TRACE="" OFONO_RIL_HEX_TRACE="" OFONO_RIL_DEVICE=ril ofonod -n -d -P stktest,provision,sap,udev,dun,smart,hfp >& /tmp/ofono.txt

let it run for a few seconds and please attach the resulting file to the bug.

[Ratchanan Srirattanamet (peat-new)]

@Alfonso, I've attached the log. But I think I know why gsm_umts_index is -1 now. After reading the log from Cyanogenmod, I've found that I need to make another request to rild to select app we'll use. I actually have a patch (hack!) that make my phone register with the network and talks USSD, but I think it's out of scope of this bug now.

[Ratchanan Srirattanamet (peat-new)]

Sorry, attach wrong file.

[Alfonso Sanchez-Beato (alfonsosanchezbeato)]

@Ratchanan, thanks for the log. This implies that we should not really use the index if set to -1 for your modem, but send the request for selecting an application before. Great to know that you are already able to register, maybe getting ofonod to work will just need a couple of minor adjustments. If that is the case, you could define a new vendor in

drivers/rilmodem/vendor.h

and use it as necessary. The repo we use for our development is

https://github.com/rilmodem/ofono

Patches are always welcome ;)

[Tony Espy (awe)]

As this patch was merged into our upstream git repository, I'm changing the Status to FixCommitted. See the following pull-request for details:

https://github.com/rilmodem/ofono/pull/174

[Launchpad Janitor (janitor)]

This bug was fixed in the package ofono - 1.12.bzr6896+15.10.20150521-0ubuntu1

---------------
ofono (1.12.bzr6896+15.10.20150521-0ubuntu1) wily; urgency=medium

  [ Ratchanan Srirattanamet ]
  * build, gril, rilmodem, qcommsimmodem, plugins/qcom-msim.c: add
    support for LG L90 Dual which uses a Qualcomm dual-SIM modem (LP: #1427788)

  [ Alfonso Sanchez-Beato ]
  * src/gprs.c: support for 'ResetContexts' (LP: #1338758)
    Adds a new ConnectionManager 'ResetContexts' DBus method
    which deletes all existing gprs contexts and forces
    re-provisioning to occur.
  * plugins/ubuntu-apndb.c: ignore 'ipv6' protocol
    This change causes APNs with an 'ipv6' protocol value
    to instead be provisioned with 'ip', forcing IPv4 to
    be used.

  [ Jonas Drange ]
  * doc/connman-api.txt: add metion of 'ia' gprs context type

  [ Tony Espy ]
  * test/set-context-property: add 'Preferred' support (LP: #1454756, #1454751)
    This change adds support for the 'Preferred' ( and 'Active')
    properties which both need to be DBus variant wrapped booleans.
    Also added support for multi-SIM which was lacking from this script.
  * test/list-modems: add privacy support (LP: #1438715)
    Obfuscate any properties which can be tied to the end-user
    ( eg. SubscriberNumbers, SubscriberIdentity, ... ).

 -- CI Train Bot <email address hidden> Thu, 21 May 2015 08:44:39 +0000