1174499 - Loading a DuckDuckGo search webpage could randomly put the bottom of the screen at the top

Status: RESOLVED DUPLICATE of bug 1176301

(Core :: Graphics: Layers, defect)

Description

[Johan Lorenzo [:jlorenzo]]

Prerequisites
Have DuckDuckGo as the default search provider.

Steps to reproduce (happened once, so far)
1. From the homescreen, tap the rocketbar and execute a search

Results
After the page being loaded, the bottom of the screen (including the virtual home button) is located at the top and the top at the bottom. The touch event weren't modified. That is to say, if I wanted to press the home button, I needed to press the bottom of the screen, where the button should have been.

The screen remained like this until I pressed the power button, and then I unlocked the device.

At the same time, I saw these lines in the logcat:
> 06-14 10:44:39.386   316   959 I Gecko   : 0xa271d630 replacing unconfirmed target 0xac9e7800 with real target 0xaca90800

And then, when I turned off the screen:
> 06-14 10:46:04.996   316   316 D qdhwcomposer: hwc_blank: Blanking display: 0
> 06-14 10:46:05.196   316   968 I qdhwcomposer: handle_blank_event: dpy:0 panel power state: 0
> 06-14 10:46:05.196   316   316 D qdhwcomposer: hwc_blank: Done blanking display: 0
> 06-14 10:46:05.286   316   316 E qdhwcomposer: hwc_vsync_control: vsync control failed. Dpy=0, enable=1 : Operation not supported on transport endpoint
> 06-14 10:46:05.336   316   316 D qdhwcomposer: hwc_blank: Unblanking display: 0
> 06-14 10:46:05.676   316   968 I qdhwcomposer: handle_blank_event: dpy:0 panel power state: 1
> 06-14 10:46:05.676   316   316 D qdhwcomposer: hwc_blank: Done unblanking display: 0
> 06-14 10:46:05.996   316   316 I Gecko   : SharedSurface_Gralloc::Create -------
> 06-14 10:46:05.996   316   316 I Gecko   : SharedSurface_Gralloc::Create: success -- surface 0xa625e580, GraphicBuffer 0xac921500.
> 06-14 10:46:06.406   316   316 I Gecko   : SharedSurface_Gralloc::Create -------
> 06-14 10:46:06.406   316   316 I Gecko   : SharedSurface_Gralloc::Create: success -- surface 0xa626ddc0, GraphicBuffer 0xa92d3080.
> 06-14 10:46:06.846   316   316 I Gecko   : SharedSurface_Gralloc::Create -------
> 06-14 10:46:06.846   316   316 I Gecko   : SharedSurface_Gralloc::Create: success -- surface 0xa65ab910, GraphicBuffer 0xad1e9f00.
> 06-14 10:46:07.096   316   316 I Gecko   : [SharedSurface_Gralloc 0xa65ab910] destroyed
> 06-14 10:46:07.096   316   316 I Gecko   : [SharedSurface_Gralloc 0xa625e580] destroyed
> 06-14 10:46:07.096   316   316 I Gecko   : [SharedSurface_Gralloc 0xa626ddc0] destroyed
> 06-14 10:46:08.616   316   316 D qdhwcomposer: hwc_blank: Blanking display: 0
> 06-14 10:46:08.796   316   968 I qdhwcomposer: handle_blank_event: dpy:0 panel power state: 0
> 06-14 10:46:08.796   316   316 D qdhwcomposer: hwc_blank: Done blanking display: 0

Comment 1

[Johan Lorenzo [:jlorenzo]]

I am not able to repro this issue easily. Kats, do you have some hints to trigger the error at [1]?

[1] https://dxr.mozilla.org/mozilla-central/source/gfx/layers/apz/src/InputBlockState.cpp#50

Comment 2

[Johan Lorenzo [:jlorenzo]]

Side note: The screenshot I took, when that happened, doesn't show the discrepancy.

Comment 3

[Kartikaya Gupta (email:[e-mail])]

That message is not an error, it is just an informational message. This doesn't sound like an APZ bug. If you can reproduce it again please take a photo of the device with another device since screenshots don't show the problem.

Comment 4

[Johan Lorenzo [:jlorenzo]]

Attachment: Photoshop'd picture of what I saw (not an actual screenshot or photo)

Oh right, I didn't see the comment saying that printf_stderr was just temporary.

I haven't seen the issue again, I can remember what I saw, though. Here's a photoshop'd picture of the issue I got.

Comment 5

[Kartikaya Gupta (email:[e-mail])]

Thanks; it looks like it might be a tiling issue - there have been some changes in that part of the code recently.

Comment 6

[Kartikaya Gupta (email:[e-mail])]

(Likely a dupe of bug 1176301)

Comment 7

[Sotaro Ikeda [:sotaro]]

(In reply to Kartikaya Gupta (email:[e-mail]) from comment #6)
> (Likely a dupe of bug 1176301)

Yea, I also think so. It seems to related to MDP hw composer.