1181797 - [Aries-L][Lockscreen] The screen will flash brighter before the screen deactivates when the display times out

Status: RESOLVED WONTFIX

(Firefox OS Graveyard :: Vendcom, defect)

Description

[Adam Alldredge [:AdamA][Inactive]]

Attachment: logcat

Description:
When the display times out the screen will go dark before it times out. right as the screen turns off the brightness will return to normal before it turns off.

Repro Steps:
1) Update a Aries to 20150707205829
2) Let phone sit until screen goes darker
3) Observe phone as display times out

Actual:
the screen will flash brighter right before the display times out

Expected:
It is expected that the display transitions from being darkened to off without becoming brighter

Environmental Variables:
Device: Aries 2.5 [Full flash]
Build ID: 20150707205829
Gaia: 553e4e281d11500bea702918841f011bfb5405f7
Gecko: 0246d86e619670bf381d45d587e8ab036e7a8db5
Version: 42.0a1 (2.5)
Firmware Version: D5803_23.1.A.1.28_NCB.ftf
User Agent: Mozilla/5.0 (Mobile; rv:42.0) Gecko/42.0 Firefox/42.0

Repro frequency: 10/10
See attached: video clip, logcat

Comment 1

[Adam Alldredge [:AdamA][Inactive]]

This issue DOES NOT occur on the earliest Aries-l build

Environmental Variables:
Device: Aries 2.5
BuildID: 20150616130603
Gaia: 6d941efc270505fcf604a39568248913e256b363
Gecko: f1830a2b6c0e5fbf03e42ce29d13a7ddc01e4d1a
Version: 41.0a1 (2.5) 
Firmware Version: D5803_23.1.A.1.28_NCB.ftf
User Agent: Mozilla/5.0 (Mobile; rv:41.0) Gecko/41.0 Firefox/41.0

Result:
The screen does not flash before it times out

Comment 2

[Adam Farden [:adfad666]]

I notice the reverse is also true on sirius-l when waking from standby; The screen wakes at high brightness then jumps to a reasonable brightness.

Comment 3

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

Adam, Do you suspect that this is a kernel issue then?

QAWanted to see if flashing the boot.img from Alexandre's build and having the recent lollipop build resolves the issue.  That would indicate a kernel issue.

NI: Ken, could you find a developer to look at this bug please?

Comment 4

[Ken Chang[:kenkai]]

Hi Boris, can you please check what the root cause is? thanks.

Comment 5

[Adam Farden [:adfad666]]

My first idea was a kernel problem. However I just confirmed it's not a problem on tianchi-l (msm8226 device) which is built with the same kernel sources (at least on my build from 2015-06-29).

On all devices I checked, on wake the backlight comes on before the lockscreen is displayed, so you wake up briefly to a black screen.

I don't see these issues on AOSP on any device. Could it be that B2G is waking up the backlight and display in the wrong order, and the flashes on Shinano platform are a side-effect of this?

I'll try to do some fresh builds when I can.

Comment 6

[Pi Wei Cheng [:piwei] (inactive)]

(In reply to Naoki Hirata :nhirata (please use needinfo instead of cc) from comment #3)
> QAWanted to see if flashing the boot.img from Alexandre's build and having
> the recent lollipop build resolves the issue.  That would indicate a kernel
> issue.

This issue is still reproducible on the Aries L 2 build that we got from you. Right before the screen turns off it flashes bright.

Device: Aries (L2, full flashed)
BuildID: 20150709132856
Gaia: ad76c159c641c977d9140c5fedea84aea04e0e60
Gecko: cc42883d641b720223054f29f096175e26a0b5f3
Version: 42.0a1 (2.5 Master) 
Firmware Version: D5803_23.1.A.1.28_NCB.ftf
User Agent: Mozilla/5.0 (Mobile; rv:42.0) Gecko/42.0 Firefox/42.0

Comment 7

[Pi Wei Cheng [:piwei] (inactive)]

Comment 6 wasn't tested using lollipop v9 blobs boot image. Redid the test with that boot image and issue does NOT reproduce. Screen timing out with expected brightness transition.

Tested on same environmental variables as comment 6 except boot image was swapped from v9 blobs build.

Comment 8

[Adam Farden [:adfad666]]

I just flashed an AOSP build onto sirius-l and it has the brightness bug there too, I just simply hadn't noticed / ignored it. It's less noticeable because there is an off animation, and the screen seems to come on quicker when waking.

This suggests to me it's a kernel bug.

piwei: It seems unusual that swapping boot images, with the only difference being which blob zip was used during build, would affect this. AFAIK there aren't any blobs used in creating the boot image.

Comment 9

[AngeloGioacchino Del Regno]

This is a well known bug and ... a wanted one.

I didn't want to solve this because it requires NASTY and BAD hacks in MDSS and eventually, if not MDSS-driven, in the backlight driver.
If you REALLY need to solve this in kernel, I can provide some more informations, so that you'll be able to do the job.

The real question is:
Are you willing to make the video driver dirty and to make nasty things? :)

Comment 10

[Boris Chiou [:boris]]

(In reply to AngeloGioacchino Del Regno from comment #9)
> This is a well known bug and ... a wanted one.
> 
> I didn't want to solve this because it requires NASTY and BAD hacks in MDSS
> and eventually, if not MDSS-driven, in the backlight driver.
> If you REALLY need to solve this in kernel, I can provide some more
> informations, so that you'll be able to do the job.
> 
> The real question is:
> Are you willing to make the video driver dirty and to make nasty things? :)

I agree. Looks like we need more discussion about this bug.

Comment 11

[:gerard-majax]

That's reproduced on AOSP builds so upstream issue

Comment 12

[BMO Automation]

Firefox OS is not being worked on