Closed
Bug 908123
Opened 11 years ago
Closed 11 years ago
[Buri][Usage]The data usage increase abnormally when data connection and WiFi turned off.
Categories
(Firefox OS Graveyard :: Gaia::Cost Control, defect, P1)
Firefox OS Graveyard
Gaia::Cost Control
Tracking
(blocking-b2g:-)
RESOLVED
INCOMPLETE
blocking-b2g | - |
People
(Reporter: sync-1, Unassigned)
References
Details
(Whiteboard: [com-ril])
Attachments
(6 files)
Firefox os v1.1
Mozilla build ID:20130817041203
DEFECT DESCRIPTION:
The data usage increase abnormally when data connection and WiFi turned off.
REPRODUCING PROCEDURES:
1. forget all AP in WiFi settings, turned off WiFi and 3g data connection.
2. enter 'usage' and reset all the usage data to 0.
3. enter 'here maps' and scroll the map.
4. will find data transfer icon flashes and map downloaded and showed. --KO1
5. enter 'usage', find Mobile usage not 0. -- KO2
6. turn on WiFi but not connect to any AP.
7. enter 'here maps' again and scroll the map.
8. enter 'usage', find WiFi usage not 0. -- KO3
EXPECTED BEHAVIOUR:
The data usage must not increase.
ASSOCIATE SPECIFICATION:
TEST PLAN REFERENCE:
TOOLS AND PLATFORMS USED:
USER IMPACT:
hi
REPRODUCING RATE:
5/5
For FT PR, Please list reference mobile's behavior:
Comment 1•11 years ago
|
||
Ray,
Please attach logs in problem state
Updated•11 years ago
|
QA Contact: sparsons
Comment 4•11 years ago
|
||
This issue was reproduced on the Buri device Build ID: 20130829041204
Environmental Variables
Build ID: 20130829041204
Gecko: http://hg.mozilla.org/releases/mozilla-b2g18/rev/0c3c48cc121d
Gaia: 23eadfd88057dc524f4452e48813503627f77fbe
Platform Version: 18.1
RIL Version: 01.01.00.019.206
However, unable to reproduce this issue on Leo Build ID: 20130828041203
Environmental Variables
Build ID: 20130828041203
Gecko: http://hg.mozilla.org/releases/mozilla-b2g18/rev/c3b4683d1360
Gaia: 518178ca234c98eb1ca0c0997b9d59faea640a85
Platform Version: 18.1
RIL Version: 01.01.00.019.205
Comment 5•11 years ago
|
||
QAWanted - Can we try testing this on the latest leo & buri partner images available? Triage is looking to understand if this happens on a specific partner build or not.
Keywords: qawanted
Updated•11 years ago
|
Component: Gaia → Gaia::Cost Control
Updated•11 years ago
|
QA Contact: sparsons → laliaga
Comment 6•11 years ago
|
||
Got the repro on Unagi 1.1 build. Does not repro on Leo and Buri partner builds. After launching "Here Maps" with Cell & Data and Wifi disabled, data usage remains at 0.
- Buri partner -
Environmental Variables
Build ID: 20130822172034
Gecko: /rev/
Gaia: 743bc0fd1c239632a65f337a7a1e2c9d17ff8190
Platform Version: 18.1
- Leo partner -
Environmental Variables
Build ID: 20130828155538
Gecko: /rev/
Gaia: bcfbffbbb59406b8a07def8ab264792c8d28f23a
Platform Version: 18.1
- Unagi 1.1 -
Environmental Variables
Build ID: 20130828043201
Gecko: http://hg.mozilla.org/releases/mozilla-b2g18_v1_0_1/rev/9c62297d11b0
Gaia: 054cdc27404e2daca91d3065d9783681032b2151
Platform Version: 18.0
Keywords: qawanted
Comment 7•11 years ago
|
||
Sorry, had the wrong environmental details for Unagi 1.1. Here are the proper details.
Environmental Variables
Build ID: 20130830041201
Gecko: http://hg.mozilla.org/releases/mozilla-b2g18/rev/2fb19341d934
Gaia: 577ab5105af418efc8ded00798fd3644b3dac674
Platform Version: 18.1
Comment 8•11 years ago
|
||
Salva,
Please look into this bug as it suggests it is a usage issue.
Assignee: nobody → salva
Comment 9•11 years ago
|
||
(In reply to sync-1 from comment #0)
> Firefox os v1.1
> Mozilla build ID:20130817041203
>
> DEFECT DESCRIPTION:
> The data usage increase abnormally when data connection and WiFi turned off.
>
> REPRODUCING PROCEDURES:
> 1. forget all AP in WiFi settings, turned off WiFi and 3g data connection.
> 2. enter 'usage' and reset all the usage data to 0.
> 3. enter 'here maps' and scroll the map.
> 4. will find data transfer icon flashes and map downloaded and showed. --KO1
This is not Usage application related.
> 5. enter 'usage', find Mobile usage not 0. -- KO2
Cost Control application reflect what is in the Network Statistics database. If you see data transfers in step 4, it is expected the Usage application reflect that.
> 6. turn on WiFi but not connect to any AP.
> 7. enter 'here maps' again and scroll the map.
> 8. enter 'usage', find WiFi usage not 0. -- KO3
Did you see the transfer icon flashing? If not, this is strange but please note the Usage application is still showing what is in the Network Statistics DB. Maybe you should check if the interfaces are actually disconnected when you turn off the connection.
You can see it by connecting via adb shell and using the commands netcfg.
Comment 10•11 years ago
|
||
(In reply to Salvador de la Puente González [:salva] from comment #9)
I check logs when it happened, and find HERE MAPS open data connection. Could you tell me which API can control data connection. Thanks.
Comment 11•11 years ago
|
||
Andrew,
It is suggested to be a gecko issue. Please reassign.
Flags: needinfo?(overholt)
Updated•11 years ago
|
Assignee: salva → nobody
Comment 13•11 years ago
|
||
Any uodates?
Comment 14•11 years ago
|
||
If the network traffic upload/download indicators at the top are blinking, that almost definitely means we're actually sending network traffic (they're pumped by actual network traffic). I suspect we're not properly turning off the network device. I can't remember who the usual suspects are for the B2G network device code (and layerwise this could be as far down as the actual device code, or, more likely, code above it that's not calling TurnOn/TurnOff on the device code).
Comment 15•11 years ago
|
||
Ken, could this be caused by something in the network manager?
Flags: needinfo?(overholt) → needinfo?(kchang)
Comment 16•11 years ago
|
||
(In reply to Andrew Overholt [:overholt] from comment #15)
> Ken, could this be caused by something in the network manager?
Ethan, can you please check if Services.io.offline is in wrong state.
Assignee: nobody → ettseng
Flags: needinfo?(kchang)
Comment 17•11 years ago
|
||
Blocking on this, but Ethan or Jonas - if you feel differently based on your analysis of this bug, feel free to renominate.
blocking-b2g: leo? → leo+
Comment 18•11 years ago
|
||
Hi everyone,
I am working on this bug, but I failed to build Unagi 1.1.
The compile error is related to lack of declarations of INT32_MAX, UINT32_MAX, ..., etc.
Some coworker encountered the same problem as well.
I already sent an email to dev-b2g mailing list to request helps.
In the meantime, I will try to reproduce this bug on Buri 1.1.
Comment 19•11 years ago
|
||
(In reply to ettseng from comment #18)
> Hi everyone,
>
> I am working on this bug, but I failed to build Unagi 1.1.
> The compile error is related to lack of declarations of INT32_MAX,
> UINT32_MAX, ..., etc.
> Some coworker encountered the same problem as well.
> I already sent an email to dev-b2g mailing list to request helps.
> In the meantime, I will try to reproduce this bug on Buri 1.1.
Asking Albert...
Flags: needinfo?(acperez)
Comment 20•11 years ago
|
||
I followed exactly the reproducing procedures but cannot reproduce this bug on the latest Unagi 1.1 code base.
In step 4, the data transfer icon didn't flash at all, and 'here maps' was not downloaded nor showed.
In step 5 and 8, the 'usage' remains at zero.
Device: Unagi
OS version: 1.1.1.0-prerelease
Platform version: 18.1
Build ID: 20130924174142
Lucas, as you reproduced this bug on Unagi, I would like to try the build you used (build ID 20130830041201).
Did you get the image from pvtbuilds.mozilla.org? If yes, could you please provide the link?
Flags: needinfo?(laliaga)
Comment 21•11 years ago
|
||
minus as it is not reproducible. Please re-nom if this is reproducible and turns out serious.
blocking-b2g: leo+ → -
Comment 22•11 years ago
|
||
Comment 23•11 years ago
|
||
Hi ettseng and Preeti,
I test in our side, it is also exist. attachment 809714 [details] is a video when it happended,please help check.
Our version is Firefox os v1.1 Mozilla build ID:20130916041201
Comment 24•11 years ago
|
||
Comment 25•11 years ago
|
||
Comment 26•11 years ago
|
||
Hi Ray,
Thanks for you video. :)
Apparently after you turned off Wifi and 3G data connection, certain connection still exists.
In my case, when I enter "Here Maps" app right after I turned off Wifi and 3G, the app would detect that there's no available connection. Its search bar shows "Connect to internet to search" (https://bugzilla.mozilla.org/attachment.cgi?id=810404).
When either Wifi or 3G is turned on, the search bar would become "Search for places" (https://bugzilla.mozilla.org/attachment.cgi?id=810402).
In your video, the search bar of Here Maps shows "Search for places", which means the application gets an available connection to use.
I think we can analyze this issue further with the information of routing tables.
1) When I turned on 3G, the routing table is as follows:
$ adb shell cat /proc/net/route
Iface Destination Gateway Flags RefCnt Use Metric Mask MTU Window IRTT
rmnet0 00000000 2EC98872 0003 0 0 0 00000000 0 0 0
rmnet0 2CC98872 00000000 0001 0 0 0 FCFFFFFF 0 0 0
rmnet0 01015FA8 2EC98872 0007 0 0 0 FFFFFFFF 0 0 0
rmnet0 01C05FA8 2EC98872 0007 0 0 0 FFFFFFFF 0 0 0
2) When I turned on Wifi, the routing table is as follows:
$ adb shell cat /proc/net/route
Iface Destination Gateway Flags RefCnt Use Metric Mask MTU Window IRTT
wlan0 00000000 0118F70A 0003 0 0 0 00000000 0 0 0
wlan0 0018F70A 00000000 0001 0 0 313 00F8FFFF 0 0 0
3) When I turned off 3G and Wifi, the routing table is empty. There is no any default route for data to fly.
Ray, could you please provide the routing table of the device when this bug occurs?
Flags: needinfo?(ettseng)
Comment 27•11 years ago
|
||
(In reply to Ethan Tseng [:ethan] from comment #18)
> Hi everyone,
>
> I am working on this bug, but I failed to build Unagi 1.1.
> The compile error is related to lack of declarations of INT32_MAX,
> UINT32_MAX, ..., etc.
> Some coworker encountered the same problem as well.
> I already sent an email to dev-b2g mailing list to request helps.
> In the meantime, I will try to reproduce this bug on Buri 1.1.
Same problem here after updating mozilla-central-b2g18. Seems that build is broken. Did you open a bug for this?
Flags: needinfo?(acperez)
Comment 28•11 years ago
|
||
Hi Albert,
We found that the compile error is because we used the wrong b2g branch while building FFOS v1.1 (gecko-18).
We used b2g "v1-train" branch instead of master and built v1.1 successfully.
Try to pull a clean repository on v1-train branch and build again. :)
Comment 29•11 years ago
|
||
Hi Ray,
Here is one more idea.
The bug was generated and reported on a build with commercial RIL. We can tell that from the log (https://bugzilla.mozilla.org/attachment.cgi?id=797011).
One difference between our reproducing environment is that we are based on build with mozRIL, not commercial RIL.
Could you help to verify that, this bug is a commercial-RIL-only bug or not?
Could we reproduce this bug on a build with mozRIL?
Comment 30•11 years ago
|
||
(In reply to Ethan Tseng [:ethan] from comment #26)
> Ray, could you please provide the routing table of the device when this bug
> occurs?
******@aclgcl-ubnt:~$ adb shell cat /proc/net/route
Iface Destination Gateway Flags RefCnt Use Metric Mask MTU Window IRTT
rmnet0 00000000 A1401D0A 0003 0 0 0 00000000 0 0 0
rmnet0 A0401D0A 00000000 0001 0 0 0 F0FFFFFF 0 0 0
rmnet0 034616D2 A1401D0A 0007 0 0 0 FFFFFFFF 0 0 0
rmnet0 035416D2 A1401D0A 0007 0 0 0 FFFFFFFF 0 0 0
Comment 31•11 years ago
|
||
(In reply to Ethan Tseng [:ethan] from comment #29)
> Could you help to verify that, this bug is a commercial-RIL-only bug or not?
> Could we reproduce this bug on a build with mozRIL?
Hi Ethan, I can't get mozRIL in our office, could you provide mozRIL's .so file to me, I will try to replace it? Thanks.
Comment 32•11 years ago
|
||
Need commercial RIL to check why PDP is still alive after DEACTIVATE_DATA_CALL request.
Hi Michael:
Could you please have a RD to check this issue?
Thanks!!
sku
08-28 22:00:33.502 D/RIL_QC_B2G( 140): [0212]> SETUP_DATA_CALL radioTech=1, profile=0, apn=3gnet, user=(null), password=(null), authType=3, protocol=IP, usrObj=44dbf100
08-28 22:00:36.262 D/RIL_QC_B2G( 140): [0212]< SETUP_DATA_CALL { ver=6, num=1, ver=6, status=0, retry=-1, cid=0, active=2, type=IP, ifname=rmnet0, addresses="10.33.173.102/30", dnses=["210.22.70.3","210.22.84.3"], gateways="10.33.173.101}
08-28 22:01:11.562 D/RIL_QC_B2G( 140): [0238]> DEACTIVATE_DATA_CALL cid=0, reason=0, usrObj=0
08-28 22:01:13.182 E/DATA_CALL_QC_B2G( 140): Release - Unexpected release of a permanent DataCall
08-28 22:01:13.182 D/RIL_QC_B2G( 140): [0238]< DEACTIVATE_DATA_CALL
Flags: needinfo?(mvines)
Whiteboard: [com-ril]
Comment 33•11 years ago
|
||
Hi Ethan,
Does mozRIL locate in dir :hardware/ril/mock-ril?
The Android.mk file needs libv8.so, but there is no rule to make it, can you provide it? or can you provide mozRil.so?
Thank you!
Comment 34•11 years ago
|
||
Please create an SR for commercial RIL support, thanks!
Flags: needinfo?(mvines)
Comment 35•11 years ago
|
||
Hi, Ray and yangshjuan,
As Shawn Ku commented above (comment #32), we have to resolve this bug within commercial RIL. I think we can suspend the bug reproduction on mozRIL for now.
The people who wrote the original implementation over in bug 746069 are probably the best bet for getting this moving.
Flags: needinfo?(jonas)
Updated•11 years ago
|
Assignee: ettseng → nobody
Comment 37•11 years ago
|
||
Comment 38•11 years ago
|
||
Hi Ethan,
I add new logs and we think it maybe caused by GPS. Could you have some good ideas to fixed?
Some logs like following:
D/Diag_Lib( 143): LOCK mutex qcril_reqlist_mutex by thread 263 success!
D/Diag_Lib( 143): UNLOCK mutex qcril_reqlist_mutex by thread 263
D/Diag_Lib( 143): Reply to RIL --> Automatic network selection mode
D/Diag_Lib( 143): LOCK mutex qcril_reqlist_mutex by thread 263
D/Diag_Lib( 143): LOCK mutex qcril_reqlist_mutex by thread 263 success!
D/Diag_Lib( 143): [RID 0] ReqList entries :
D/Diag_Lib( 143): RIL_REQUEST_SETUP_DATA_CALL (27), token id 161
D/Diag_Lib( 143): UNLOCK mutex qcril_reqlist_mutex by thread 263
D/Diag_Lib( 143): UI <--- RIL_REQUEST_QUERY_NETWORK_SELECTION_MODE (45) Complete --- RIL [RID 0, Token 169, Success, Len 4 Automatic]
E/EONS_QC_B2G( 140): Eons::UpdateEons
E/EONS_QC_B2G( 140): Eons::UpdateEonsFromOplAndPnn
Comment 39•11 years ago
|
||
Please see comment 34.
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → INCOMPLETE
Comment 40•11 years ago
|
||
SR is expected for commercial RIL to check this issue. (comment 34)
clear ni?
Flags: needinfo?(sku)
Comment 42•11 years ago
|
||
The data connected by GPS(AGPS), and it has been fixed by qualcomm. Maybe you could close it. Thanks.
Comment 43•11 years ago
|
||
Hi Ray,
Could you tell me the SR number or give me the patch?
Thanks
Flags: needinfo?(rll)
Flags: needinfo?(laliaga)
Flags: needinfo?(ettseng)
Flags: needinfo?
Comment 44•11 years ago
|
||
(In reply to lecky from comment #43)
> Hi Ray,
>
> Could you tell me the SR number or give me the patch?
>
> Thanks
The SR number is 01327589 and I will upload the patch.
Flags: needinfo?(rll)
Flags: needinfo?
Comment 45•11 years ago
|
||
Comment 46•11 years ago
|
||
(In reply to Ray REN from comment #45)
> Created attachment 822091 [details] [diff] [review]
> 0001_Only_allow_MMS_data_call_when_data_is_disabled.patch
Hi Ray,
Thanks very much for your help.I have one question:
(1)The patch can fix mobile usage increasement bug ,but it seems can not fix the wifi usage increasement bug(traffic include mobie uase and wifi usage)
Comment 47•11 years ago
|
||
(In reply to lecky from comment #46)
> Hi Ray,
>
> Thanks very much for your help.I have one question:
> (1)The patch can fix mobile usage increasement bug ,but it seems can not
> fix the wifi usage increasement bug(traffic include mobie uase and wifi
> usage)
Now we could not reproduce the wifi usage increasement issue.
You need to log in
before you can comment on or make changes to this bug.
Description
•