802418 - [settings][wifi] Wifi disables and disconnects itself

Status: RESOLVED WORKSFORME

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

Description

[Dietrich Ayala (:dietrich)]

Saw wifi was disconnected (again), and went to re-connect. Oddly, it was disabled.

I'm 100% confident I did nothing in Settings to disable it. I was connected then was testing other stuff, set the phone down for a while, picked it up, unlocked, and saw disconnected. Went to Settings and the label said "disabled". I checked the toggle, and certainly was disabled!

Comment 1

[Dietrich Ayala (:dietrich)]

Otoro, 10/16 nightly build. Continually disconnecting itself, while I haven't moved from this couch. Sometimes it reconnects right when I open Settings app. Sometimes it requires a manual reconnection.

Comment 2

[Dietrich Ayala (:dietrich)]

More: Just saw disconnection again. Sitting in same spot. Connected to wifi. Closed Settings app, phone goes to sleep. Unsleep phone, notice on lockscreen that wifi indicator now shows disconnected. Unlock phone, watch wifi connect immediately.

Comment 3

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

I think it's by design that when you are in sleep, wifi turns off to conserve battery power and reconnects.  Is that the case?

I saw a condition when if an e.me app or browser goes to a page that can't be displayed, the wifi turns off.  I can't seem to repro right now.  Still trying to find steps.

Comment 4

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

Still trying to reproduce this issue.  It may take awhile?  Yesterday's build was easiser to reproduce the issue than today's...

Comment 5

[David Scravaglieri [:scravag]]

Can't reproduce too

Comment 6

[Ben Francis [:benfrancis]]

I have seen this. The phone went to sleep, when I woke it up I had no Internet connection. I opened Settings and WiFi was set to Disabled.

Is this more likely to be a platform bug?

Comment 7

[Dietrich Ayala (:dietrich)]

Saw it disable itself again last night.

+ mrbkap & kaze

Comment 9

[John Shih]

I also met this situation during I am using the device
There is no specific steps but I'm sure it did happen which is so weird..
I'll try to reproduce

build info
2012-10-23(TIP)
gaia master: 608ba8a6a931322c96ac1cea7e02f4c4bf9d70fd
gecko: 167d5a73c6aa564d1abcae4654467aa18aba9a3e

Comment 10

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

## Environment :
Otoro phone, build 2012-10-23
Taken from default.xml in b2g-distro: 
* "platform_build" revision= db539a3bd139c93c09b0cd1c3f9396b74d68717c
* "gaia" revision= 0b8ec9b8c16429dc35453dbb7b9342fab3dd18fb 
* "releases-mozilla-aurora" revision= f58edfde05cb708f8a2c440d338f2e429aaf372b
* "gonk-misc" revision= db0c751715f4696515735eb1e0dc5df7a40eb81d

  
## Repro :
1. set alarm for 10 mins
2. let the device sleep during that time
3. look at the status bar when the alarm sounds

## Expected :
1. the wifi should be connected

## Actual :
1. the wifi is scanning

## Note :
- I think basically since the alarm woke up the device and not human interaction, it caused the wifi not to be ready?  Not sure.

Comment 11

[Fabien Cazenave [:kaze]]

Possible explanation here: https://bugzilla.mozilla.org/show_bug.cgi?id=804815

Comment 14

[Blake Kaplan (:mrbkap) (inactive)]

Oops, https://hg.mozilla.org/integration/mozilla-inbound/rev/17534a39a5a5 landed pointing at this bug. It should have pointed to bug 796640 instead.

Comment 15

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/17534a39a5a5

Comment 16

[David Scravaglieri [:scravag]]

+1

gecko-aurora: 46a6feeaf9c5
gaia: 2b1ee6f8389212c46634959a3922288c286bbfc4

Comment 17

[Blake Kaplan (:mrbkap) (inactive)]

To save battery, Gaia turns off wifi when the screen goes to sleep (unless there's a wake lock on wifi). I wonder if this bug is actually just about that behavior. When it was filed, and up to about the beginning of this week, we had two bugs that would:

  * Cause wifi to not come back up when asked (bug 799825).
  * Make us not scan properly (bug 807148).

those bugs should be mostly fixed. Does that mean this bug is fixed?

Comment 18

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

Forgot about this bug... I created Bug 811833.  I think the wifi should auto connect after being woken up.

Comment 19

[Alex Keybl [:akeybl]]

Marking for C2, given this meets the criteria of known P1/P2 blocking-basecamp+ bugs at the end of C1.

Comment 20

[Dietrich Ayala (:dietrich)]

I am now testing on Unagi only, and do not see this problem anymore.

Comment 21

[Blake Kaplan (:mrbkap) (inactive)]

I'm going to mark this WFM then. If people still experience this, we can either reopen this bug or file new ones.

Comment 22

[Henrik Skupin [:whimboo][⌚️UTC+1]]

I can see this behavior constantly with an Unagi device. Once it falls into sleep mode Wifi gets disconnected and the mobile data connection is used instead. Once you press the power button and the lockscreen is displayed I can see that WIFI gets reconnected.

We shouldn't disconnect from Wifi in sleep mode given that it adds traffic for background apps to the mobile plan. If a known WIFI network is available we should always use it except the user disconnected on his own.

I'm using the nightly build from 2012-12-17.

Comment 23

[Johnny Stenback (:jst)]

Henrik, Wifi turning off when the device goes to sleep is per design, we need to do so in order to preserve battery. This bug is about wifi not reconnecting when the device is woken up again after having gone to sleep, and it sounds like your device does reconnect, therefore this bug is fixed.

Closing again. If you feel we should not disable wifi while the device goes to sleep, that's another bug and another product decision that would need to be made, but it's not this bug.

Comment 24

[Henrik Skupin [:whimboo][⌚️UTC+1]]

Well, I had several issues with wifi. The before mentioned issue was one of it and I will make sure to get a new bug filed. Because I don't think the battery consumption will be lesser with 3G/2G enabled.

But I also hit the issue that wifi was disabled as Dietrich mentioned initially. Right now I can't reproduce it but once it happens again I assume *this* is the right bug for, right?