Not reproducible in Chrome 50.0.2661.90/CrOS 7978.64.0 - veyron_minnie

Deleted: m50.png 764 KB

	m50.png 764 KB View Download

Can you take a look Haxia? 

This seems pretty bad on Veyron. 

This is only happening on Minnie?

Sounds like a mali driver bug, maybe a regression from r8p0. Can you repro on daisy too?

re:#5 not sure, I'll get a daisy updated to branch 51 and check

p.s. I do see this on minnie. It doesn't matter if I'm in street map mode or satellite view mode; it is the driving directions (text bubbles) that are missing

re:#5, #6

No unfortunately this does not repro on a daisy.

Testing with 51.0.2704.26 (Official Build) 8172.12.0 dev-channel daisy
From about://gpu:

GL_VENDOR ARM
GL_RENDER Mali-T604
GL_VERSION OpenGL ES 3.1 v1.r8p0-03dev0.ac3e1b96ba4174174444f958226e1495

Are there multiple mali builds labeled r8p0?

On my minnie (which has the repro) about://gpu says

GL_RENDERER Mali-T760
GL_VERSION OpenGL ES 3.1 v1.r8p0-03dev0.05c3530bcfb345785322afd34e089218

Tested again at TOT daisy build (my local build after repo sync yesterday), still no repro on daisy.

GL_VERSION OpenGL ES 3.1 v1.r8p0-03dev0.895a0c493c31b3f745f751fde0e4693b

Ok I can confirm this is caused by the Mali r8p0 uprev

commit 5c31f911c09afc44aa9d95fd35d44f37d3d6bcb6
Author: Salva Climent <salva.climent@arm.com>
Date:   Fri Mar 11 11:02:21 2016 +0100

    MALI: r8p0-03dev0 TX011 TX004X06X DDK
    
    DEV release of the r8p0 versions of:
    TX011
    TX004X06X
    
    NB: This commit will only work once the LLVM and CLANG
    source has been downloaded with the script provided in
    "compiler/download_and_patch.sh".
    
    BUG=chome-os-partner:51084
    TEST=Checked login page, aquarium, glbench.
    Tested on veyron_jaq and Daisy, with USE set to
    each of: "", "debug"
    
    Change-Id: I0e838e17e01be59cc8d42d52c64918603f94e836
    Reviewed-on: https://chrome-internal-review.googlesource.com/250987
    Commit-Ready: Stéphane Marchesin <marcheu@google.com>
    Tested-by: Stéphane Marchesin <marcheu@google.com>
    Reviewed-by: Ilja Friedel <ihf@google.com>
    Reviewed-by: Stéphane Marchesin <marcheu@google.com>

The following steps were taken:

(1) On veyron_minnie canary 8000.0.0 (mali r5p0): no repro
(2) Emerge and deploy mali-drivers at this commit (5c31f911): repros
(3) Rewind and deploy mali-drivers to the earlier commit (9b307af5): no repro

Here's an even easier way to reproduce the bug. Just open Google Maps (maps.google.com) and zoom out a bit and observe the freeway markers (e.g. "US-101"). At some zoom levels the text will disappear, leaving only empty rectangles.

See attached screenshot.

Deleted: Screenshot 2016-04-29 at 4.27.05 PM.png 292 KB

	Screenshot 2016-04-29 at 4.27.05 PM.png 292 KB View Download

I wonder where we shall look in the mali change specifically regarding text rendering.

Here's another screen capture I just made while zooming and scrolling maps. It is clear that we have trouble rendering text (it is flickering).

https://goo.gl/photos/8QNqicLy2nQyjUU68

Reproducible in Chrome 51.0.2704.30/ CrOS 8172.18.0 - veyron-jerry, veyron-mickey

While loading the maps with search addresses, able to see all text on maps (whileloading.png).  
After loading the page, the text disappear (afterloading.png)

Deleted: whileloading.png 704 KB

	whileloading.png 704 KB View Download

Deleted: afterloading.png 702 KB

	afterloading.png 702 KB View Download

Not reproducible in Chrome 51.0.2704.30/ CrOS 8172.18.0 - Daisy, Blaze

ARM, can you take a look at this regression?

Partner bug tracker filed at https://code.google.com/p/chrome-os-partner/issues/detail?id=52904

Can we revert https://chrome-internal-review.googlesource.com/#/c/250987/ to get this working again?

not really, unless you want to cancel gru.

We could revert it on R51, so gru is not impacted, unless there are concerns reverting it will make things worse for systems that are shipping on R51.

We can go on without a fix for now as well, but that means we miss the second beta for RK3288 based systems on R51. 

@22: it will reopen other bugs though. I am going to ask to postpone 51 instead.

Is this considered fixed now? We have merged a fix from the related internal bug on this into R51 already.

Verified fixed on 51.0.2704.42 (Official Build) 8172.28.0 dev-channel veyron_minnie

[Auto-generated comment by a script] We noticed that this issue is targeted for M-51; it appears the fix may have landed after branch point, meaning a merge might be required. Please confirm if a merge is required here - if so add Merge-Request-51 label, otherwise remove Merge-TBD label. Thanks.

[Bulk edit]

Our blockerbot script was offline; it was recently brought back online, and thus labeled many old issues (including this one) erroneously.  Removing Merge-TBD label since all milestones for this issue are already completed; no further work should be done.