I believe this is operator specific. I'm running Firefox OS 2.0 on ZTE Open C. Operator is Life :) Ukraine. It is not possible to check balance or run any other USSD codes. [Reproducing] 1. Open dialer 2. Dial *111# or any other code like *121# (those are valid codes) [Actual results] Session expired message is displayed. [Desired results] Balance information should be displayed (in case of *111#). [Comment] I also noticed that the balance actually shows up (there's a screen with a bigger amount of text appearing) for a fraction of a second but then the Session Expired screen appears. I know this needs more info, but could you please provide the steps I need to perform to pull the info out. Thank you.

This could be a regression caused by bug 1016885 or bug 1002327 which did quite a few changes to the USSD code. Could you test if this worked correctly in v1.3?

This could also be related to bug 1039850.

I can confirm it didn't work before fixes for bug 1016885 and bug 1002327 landed on v1.4. I received my phone on Jul 3 and I instantly flashed the v1.4 and it didn't work from day 1. I updated v1.4 to v2.0 yesterday with hopes this was fixed in 2.0 but no luck. Can it still be affected by those? Also, I can confirm it is probably related to bug 1039850, because when I send the USSD code and then instantly go back to homescreen I receive both null notification and a notification with an actual balance.

(In reply to Anton Borodyuk from comment #3) > Created attachment 8462459 [details] > 2014-07-25-12-53-25.png > > I can confirm it didn't work before fixes for bug 1016885 and bug 1002327 > landed on v1.4. I received my phone on Jul 3 and I instantly flashed the > v1.4 and it didn't work from day 1. I updated v1.4 to v2.0 yesterday with > hopes this was fixed in 2.0 but no luck. > Can it still be affected by those? No, it's unlikely then. > Also, I can confirm it is probably related to bug 1039850, because when I > send the USSD code and then instantly go back to homescreen I receive both > null notification and a notification with an actual balance. This is very important info actually, it seems that we're getting more than one system message as a reply to your request from the carrier so it's either some quirkiness in how the carrier answer or us missing a corner case. I'll attach a patch that you can use for testing so we'll be able to figure out what's going on.

Okay, waiting for patch. I'll patch up and report back.

Here's a patch that can help me diagnose the issue, it dumps out everything the dialer receives so I'll be able to figure out what messages are we getting. Build and flash gaia with this patch applied then try reproducing the issue. Afterwards capture the output of adb logcat and attach it to the bug. Thanks!

I've done adb logcat -c just before dialing the number to make the log as clean as possible. Here it is. Dialed *111#

This is somewhat bizarre, the ussdreceived event is being triggered twice, once with the correct message and then with a second empty one. The second event is causing the "Session expired" screen to appear. I'm unsure why you're getting two events so we need to dig further. First I'd like to see a RIL log. That is going to be a little more complicated than before, first follow the instructions here: https://wiki.mozilla.org/B2G/QA/Tips_And_Tricks#RIL_Debugging Then reproduce the issue and attach the entire logcat output. Then if you have another Firefox OS phone it would be interesting to see if you hit the same problem or not. My gut feeling is that there's something wront with the RIL but I'm not sure what ships on the Open C. BTW does all the USSD messages present this problem or do other codes work correctly (e.g. *#30#)?

This is the RIL log created. Unfortunately I don't have another FFOS phone to test this. *#30# works correctly. Tells me "Service was enabled". Looks like none of the operator-specific codes that are ending do not work. Menu work. For example, there's a USSD menu with code *124# that works fine, but there's no termination point anywhere. Every menu has a back option so I can't test if this problem appears there.

Any news on this?

(In reply to Anton Borodyuk from comment #10) > Any news on this? Sorry, I've been really busy with blockers. I'll get back to this ASAP.

I checked the log and the root cause is the same as bug 1056535. Close as duplicate.

(In reply to Hsin-Yi Tsai [:hsinyi] from comment #12) > I checked the log and the root cause is the same as bug 1056535. Close as > duplicate. > > *** This bug has been marked as a duplicate of bug 1056535 *** Oops, I reviewed the ril log again and I realized the symptom seems different from bug 1056535. In this case, RIL parsed the mmi code correctly, and did receive ussd message with balance information. Agree with Gabriele that we need to dig into more further to understand the real problem. REOPEN it.