Issue 1197152 in chromium: Unable to input on the omnibox - Hana/Elm

[evanbenn via monorail]

Updates:
Cc: h...@chromium.org graphi...@chromium.org

Comment #92 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c92

from hob@:

I don't think this Finch experiment has changed since M60 all the way back in 2018- it should always be enabled and is primarily available in Chrome flags so developers can more easily diagnose issues.

For some context, GPU rasterization is how Chrome produces all visible content (e.g. draw a red div in the top left corner). As such, this flag affects a majority of the Chrome graphics codebase and finding the actual root cause will be tricky. If we want a quick fix to paper over the issue, we can disable GPU rasterization specifically on IMG at a huge cost to performance (software-only rasterization is far slower).

That being said, can you try toggling the flag "Skia API for compositing" and see if it changes anything? It's part of an experiment I've been running since M90 beta, but I'm getting some new bug reports now with the M91 beta rollout. I've already sent out a CL to disable my experiment in M91, and it'd be nice to know if it fixes this issue as well.

--
You received this message because:
1. You were specifically CC'd on the issue

You may adjust your notification preferences at:
https://bugs.chromium.org/hosting/settings

Reply to this email to add a comment or make updates.

[evanbenn via monorail]

Comment #93 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c93

hob@

there is no match for any 'skia' flag in chrome://flags, and under chrome://gpu 'skia renderer' is disabled.

[evanbenn via monorail]

Comment #94 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c94

Apologies re #93, I worked out that it was not a chrome flag but a CLI flag. I tried

--enable-skia-renderer
--enable-skia-renderer=blahblah
--enable-skia-renderer=true
--enable-skia-renderer=enabled
--enable-skia-renderer=y
--enable-skia-renderer=n
--enable-skia-renderer=false

But in chrome://gpu, 'skia renderer' is always disabled, so I assume something else is overriding that. (Or I am holding it wrong, but there is no incorrect cli flag error message so I'm not sure)

[evanbenn via monorail]

Comment #95 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c95

I have been informed I actually need to use --enable-features=UseSkiaRenderer, but this also has had no effect, skia still disabled in chrome://gpu.

[nartemiev via monorail]

Comment #96 on issue 1197152 by nart...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c96

I've been looking into how the omnibox actually gets drawn to find what is interfering with it:

The issue looked like it could be caused by text rendering , so I made OmniboxViewViews::OnPaint() fill the canvas with a solid color as follows:
`canvas->FillRect(gfx::Rect(0, 0, 1000, 1000), gfx::kGoogleRed500);` The rectangle drawing code worked as expected, but the suggestions box poping up still caused the red rectangle to disappear. I'd expect the rectangle drawing to work pretty consistently, so text rendering seems to be unrelated.

I also noticed that the omnibox text would only fully disappear 1-2 frames after the suggestion popup appeared. Setting the slow animation flag made it take longer, suggesting the issue is UI-driven, rather than something like a GPU image buffer being cleared somewhere.

Finally, I realized that the suggestions popup (OmniboxPopupContentsView) actually gets drawn over the text. It has a SK_ColorTRANSPARENT background, which I think usually allows the omnibox text to be seen through the popup, but there there seems to be an issue with the layer compositor on hana.

This also explains why the omnibox text disappears slowly: OmniboxPopupContentsView::ShowAnimated() changes the layer opacity 0.0 from to 1.0 in an animated transition. Changing the final opacity to 0.5 left the omnibox text partially visible, confirming that the popup is causing the issue.

It looks like the issue is specifically related with the GPU compositor is failing to handle SK_ColorTRANSPARENT between layers. Adding code to just draw a transparent rectangle in the same layer didn't break anything.

I've dumped the GLSL emitted by the Skia shader backend and will try to see if they contain any hints about what's going wrong.

[marcheu via monorail]

Comment #97 on issue 1197152 by mar...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c97

I have uploaded https://chromium-review.googlesource.com/c/chromium/src/+/2921010 to work around it, it should land soon

[Git Watcher via monorail]

Comment #98 on issue 1197152 by Git Watcher: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c98

The following revision refers to this bug:
https://chromium.googlesource.com/chromium/src/+/abc166ee652e32cfc2b36f88f7da9bea14821cd4

commit abc166ee652e32cfc2b36f88f7da9bea14821cd4
Author: Stéphane Marchesin <mar...@chromium.org>
Date: Wed May 26 23:06:54 2021

Disable GPU rasterization on Chrome OS on imagination

GPU rasterization seems to be causing issues with the UI. Since nothing
changed on the driver side, we aren't sure yet why this is happening.
But since this causes blocker bugs, let's disable GPU rasterization on
this platform until things are figured out.

Bug:1197152
Test:untested

Change-Id: Ia1ee4185d05fd8deae657cc456b2cb9974d5bf95
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2921010
Commit-Queue: Stéphane Marchesin <mar...@chromium.org>
Auto-Submit: Stéphane Marchesin <mar...@chromium.org>
Reviewed-by: Brian Ho <h...@chromium.org>
Reviewed-by: Kenneth Russell <k...@chromium.org>
Cr-Commit-Position: refs/heads/master@{#886945}

[modify] https://crrev.com/abc166ee652e32cfc2b36f88f7da9bea14821cd4/gpu/config/software_rendering_list.json

[evanbenn via monorail]

Comment #99 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c99

Thanks marcheu, I am not aware of how chrome cherry picks work, but I guess we want this in 91.0.4472.81 branch when possible.

I flashed M91 TOT to my hana:

13904.41.0 91.0.4472.81

bug repros.

Flashed home built 91.0.4472.81

bug repros (aside, device does not boot with DCHECK enabled)

Flashed home built 91.0.4472.81 + patch [Disable GPU rasterization on Chrome OS on imagination (Ia1ee4185) · Gerrit Code Review](https://crrev.com/c/2921010/4)

Bug is fixed, but the graphical glitches are present. https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c90

[nartemiev via monorail]

Comment #100 on issue 1197152 by nart...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c100

As noted in comment 99, the patch to disable all GPU compositing introduces significant graphical glitches. I thought we might be able to just disable GPU compositing for the omnibox. I saw this code in OmniboxPopupContentsView::OnBeforeWidgetInit() that forces software compositing as a workaround for a different issue:

#if defined(OS_WIN)
// On Windows use the software compositor to ensure that we don't block
// the UI thread during command buffer creation. We can revert this change
// once http://crbug.com/125248 is fixed.
params.force_software_compositing = true;
#endif

I tried setting `params.force_software_compositing = true;` in OmniboxPopupContentsView::OnBeforeWidgetInit() and RoundedOmniboxResultsFrame::OnBeforeWidgetInit(), but it had no effect.

Maybe the parameter needs to be set somewhere higher up in the view tree? Or it could just be that `force_software_compositing` actually controls something unrelated.

[evanbenn via monorail]

Updates:
Cc: rjkr...@chromium.org

Comment #101 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c101

(No comment was entered for this change.)

[nartemiev via monorail]

Updates:
Cc: csmart...@google.com

Comment #102 on issue 1197152 by nart...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c102

I've posted some discoveries over on b/187364475. To summarize, chrome is now using skia for some drawing operations, including the operation that makes part of the omnibox popup transparent. Skia is failing to fill the transparent region in the popup, causing the omnibox text to be occluded when the popup layer is composited over the omnibox text. The compositor is WAI.

The transparent region is a rounded rectangle, drawn using skia's GrFillRRectOp. This also affects some other UI elements using round rects, such as the login screen password text field.

See comments #12 to #15 on b/187364475 for further details.

Looking at the history of GrFillRRectOp.cpp, I see that there was a recent change to enable analytic MSAA, which was already reverted and relanded due to causing a regression on adreno. I tried reverting the MSAA change but it did not apply cleanly and it didn't improve anything after I resolved the conflicts.

It appears that this specific issue is caused by GrFillRRectOp's shaders triggering a driver bug. However the best way of fixing this for M91 may be to disable GrFillRRectOp entirely; all round rect operations are routed through `GrSurfaceDrawContext::drawRRect()`, which chooses between several different drawing implementations. If we remove GrFillRRectOp as an option, `GrSurfaceDrawContext::drawRRect()` will fall back to a different implementation that works properly on hana.

[nartemiev via monorail]

Comment #103 on issue 1197152 by nart...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c103

The solution I proposed in my last comment won't work, the fallback that's used when we disable GrFillRRectOp doesn't work for some other UI elements.

I realized that the analytic AA reland (https://skia-review.googlesource.com/c/skia/+/380477) was submitted suspiciously close to the commit identified by previous bisects. My previous attempt to just revert it failed, but doing a proper bisect of the skia repo showed that the analytic AA reland is indeed responsible for the breakage.

I think the correct fix is to revert the AA change. I'm not sure how this is normally be handled for chrome releases, but I think it is best to revert on the M91 branch. GrFillRRectOp has changed quite a lot since M91 and reverting in main then cherrypicking to M91 would be significantly more difficult.

[evanbenn via monorail]

Updates:
Owner: csmart...@chromium.org

Comment #104 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c104

(No comment was entered for this change.)

[evanbenn via monorail]

Comment #105 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c105

Issue 1214398 has been merged into this issue.

[bsalomon via monorail]

Updates:
Owner: csmart...@google.com

Comment #106 on issue 1197152 by bsal...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c106

(No comment was entered for this change.)

[csmartdalton via monorail]

Comment #107 on issue 1197152 by csmart...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c107

> I think the correct fix is to revert the AA change.

That will be extremely difficult. Hopefully we can better identify the driver issue.

> The solution I proposed in my last comment won't work, the fallback that's used when we disable GrFillRRectOp doesn't work for some other UI elements.

Am I hearing that disabling GrFillRRect fixes this specific bug, but creates new bugs?

My first guess is standard derivatives. What happens if you return false from can_use_hw_derivatives_with_coverage? https://github.com/google/skia/blob/main/src/gpu/ops/GrFillRRectOp.cpp#L715

[nartemiev via monorail]

Comment #108 on issue 1197152 by nart...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c108

> That will be extremely difficult. Hopefully we can better identify the driver issue.

Agreed, the first time I tried to revert it was on 4403.0 (right after things broke). There were some conflicts but they looked manageable. But after trying the revert on 4472, I think the only way of undoing that commit would be a full rollback.

> My first guess is standard derivatives. What happens if you return false from can_use_hw_derivatives_with_coverage?

That actually fixes it, but I don't think built in derivatives are the problem, it looks like the vertex stage is emitting bad values - b/187364475#comment16.

Disabling hw derivatives seems to work because it writes to arccord.zw and that prevents some compiler optimizations that are causing arccord.xy to get values. Drawing still works if you set `useHWDerivatives = false` while generating the the vertex shader and `useHWDerivatives = true` while generating the fragment shader. I haven't seen anything to indicate that`fwidth()` is returning bad values.

[csmartdalton via monorail]

Comment #109 on issue 1197152 by csmart...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c109

Hm. Very strange. But either way it sounds like we have a workaround.

I'm still a little wary that the derivatives might be causing this issue in some roundabout way though, because I believe the other round rect ops also use derivatives, and they have their own set of bugs.

Regardless, the derivative optimization is pretty minor. Does it sound reasonable to detect this particular hardware and then just disable derivative support?

[marinakz via monorail]

Comment #110 on issue 1197152 by mari...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c110

Hana/Elm is blocked for 2 stable milestones:90 and now 91. If there is a workaround can we merge required changes and update these devices with 91?

[csmartdalton via monorail]

Comment #111 on issue 1197152 by csmart...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c111

What is Hana/Elm? Is there a way to detect this hardware from the GL_VENDOR/GL_RENDERER strings? Skia can disable derivatives on Hana/Elm, but I can't help until I know how to detect this specific hardware.

[chrisw.g7nbp via monorail]

Comment #112 on issue 1197152 by chrisw...@gmail.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c112

I seem to have this or related issue since 91 update in elm beta chan.

When not focussed omnibox has normal contrast text (either light or dark themes) but when focussed and the suggestions dropdown unfolds omnibox text and background have same apparent color.

This particular device not currently in developer mode so I can do little more to diagnose - sorry.

Hope this helps.

Regards, Chris W

My build details:
Google Chrome: Version 91.0.4472.81 (Official Build) beta (32-bit)
Platform: 13904.41.0 (Official Build) beta-channel elm
Channel: beta-channel
Firmware Version: Google_Elm.8438.184.0
ARC enabled: true
ARC: 7360525
Enterprise enrolled: false
Developer mode: false


Google Chrome 91.0.4472.81 (Official Build) beta (32-bit)
Revision 0427ab82b616937c2ee3b4a2cc2e413dfce180f1-refs/branch-heads/4472@{#1282}
Platform 13904.41.0 (Official Build) beta-channel elm
Firmware Version Google_Elm.8438.184.0
Customisation ID elm
ARC 7360525
JavaScript V8 9.1.269.28
User agent Mozilla/5.0 (X11; CrOS aarch64 13904.41.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.81 Safari/537.36
Command Line /opt/google/chrome/chrome --use-gl=egl --enable-native-gpu-memory-buffers --platform-disallows-chromeos-direct-video-decoder --enable-webgl-image-chromium --enable-features=Pepper3DImageChromium,SmartDim,UmaStorageDimensions,Crostini,CrostiniEnableDlc --gpu-sandbox-failures-fatal=yes --enable-logging --log-level=1 --use-cras --enable-wayland-server --user-data-dir=/home/chronos --device-help-content-id=elm --login-profile=user --has-chromeos-keyboard --enable-touchview --guest-wallpaper-large=/usr/share/chromeos-assets/wallpaper/guest_large.jpg --guest-wallpaper-small=/usr/share/chromeos-assets/wallpaper/guest_small.jpg --child-wallpaper-large=/usr/share/chromeos-assets/wallpaper/child_large.jpg --child-wallpaper-small=/usr/share/chromeos-assets/wallpaper/child_small.jpg --default-wallpaper-large=/usr/share/chromeos-assets/wallpaper/elm_large.jpg --default-wallpaper-small=/usr/share/chromeos-assets/wallpaper/elm_small.jpg --default-wallpaper-is-oem --enable-consumer-kiosk --ash-side-volume-button-position={"region":"keyboard","side":"right"} --ozone-nnpalm-properties={} --disable-features=InstantTethering --arc-availability=officially-supported --enterprise-enrollment-initial-modulus=15 --enterprise-enrollment-modulus-limit=19 --enable-crashpad --login-manager --vmodule=*arc/*=1,*night_light*=1,*/chromeos/power/auto_screen_brightness/*=1,*/ash/wm/tablet_mode/*=1,wizard_controller=1,*/webui/chromeos/login/*=1,*/browser/chromeos/login/screens/*=1,enrollment_screen_handler=1,*/browser/chromeos/login/enrollment/*=1,*/ui/ozone/*=1,*/ui/display/manager/chromeos/*=1 --enable-features=Pepper3DImageChromium,SmartDim,UmaStorageDimensions,Crostini,CrostiniEnableDlc --disable-features=InstantTethering --flag-switches-begin --flag-switches-end --origin-trial-disabled-features=SecurePaymentConfirmation --ui-compositor-memory-limit-when-visible-mb=512
Executable Path /opt/google/chrome/chrome
Profile Path /home/chronos/u-f298030a046f71c109ce2e8b6683581c11145233
Variations 9ee6b942-ee63ae02
1d758480-37844028
5213eae1-377be55a
77056158-6fe03e14
e34eb25c-dec40624
85bf39f3-5419e7bd
90a7075b-535a3667
b0f75187-377be55a
1fa5b2f3-142e91a2
59b6f412-d5090a52
60d4b352-ea9f648f
d7561e8c-5419e7bd
b3249ec4-3f4a17df
a9ef513c-f23d1dea
6cb5e962-ee63ae02
da89714-9bc8361a
f5a851ad-17e1b615
3fa8d059-377be55a
d9e26b10-cf84582e
aed6e5d4-c9a37aea
8816d952-377be55a
38eb801c-3f4a17df
65d2bc3-377be55a
646744bf-7682caab
2c3bc653-f23d1dea
7638c831-857014ec
38b9885d-ca7d8d80
edb58ea9-f23d1dea
a2629469-ee63ae02
deb1cb12-a65651e5
d97d6ab5-354cda4a
b9fad6d-71004d07
8e44abde-377be55a
7c2504d0-3354359d
8f83697a-e70e85
dbf7a8af-bf9e86fe
f2cb61f-975892ae
dc2bb3ae-377be55a
c3aebffb-ee63ae02
a2fd384c-5c0c03aa
a6baf2da-ee63ae02
b05c2566-64207fb2
bf4029fe-f2718d9f
65570806-fa67beed
a2e38c9c-19856e2a
dd8d67e-ee63ae02
2cff698a-62d8b965
c64b8fc3-ee63ae02
8ebcf27b-3f4a17df
a582a1b8-ad75ce17
da2d8531-89cbbe4e
8e3b682d-3f4a17df
1d606bb5-f23d1dea
98be3390-d93a0620
e4a357e9-64207fb2
3fd33f16-682bfcfc
6262de83-f23d1dea
7b7adda-3f4b6378
83a11282-cd0c5295
9914b7bf-f23d1dea
e79de56c-20b01172
357a64de-20b01172
3c42bcc5-5419e7bd
fa6aa590-3a447918
c992f345-12ede6a2
9bfe3ed7-17580bb6
28114f9b-10fdb82f
d8692482-ca7d8d80
72ab65ce-ca7d8d80
a120b571-ca7d8d80
291d0672-94a80f74
7a911e9f-ca7d8d80
255dfea8-db94162
7dec9614-377be55a
ca5a2953-81658715
3487aa71-380815a5
12366760-377be55a
f8947f2b-377be55a
5fe247df-ee63ae02
55894b82-410b0ec4
ad46906e-4fd1c8b3
d3566fbd-7736ab62
cb9831f4-377be55a
4ea303a6-2777577e
f48aee36-377be55a
1c00821b-5bd6a972
8c8f9d6c-85023da1
59207d8c-3f4a17df
ef4764d7-22e20947
31f573d2-ca7d8d80
7fcf46d4-ee63ae02
a1c083c6-3f4a17df
f31b480d-f23d1dea
64970026-cf84582e
6a9d5557-377be55a
5d2c15d5-d10c11f7
7760b5b2-b4b26251
d1466cda-ea7b0372
56165f3f-2f8627cd
931c5f72-30394148
494d8760-52325d43
3ac60855-486e2a9c
63dcb6a3-c67cb0df
e706e746-63e17921
f296190c-36c69439
4442aae2-6e3b1976
f690cf64-6e3b1976
ed1d377-e1cc0f14
75f0f0a0-6bdfffe7
e2b18481-d7f6b13c
e7e71889-4ad60575
3a8271ac-12c226
b1ceb06f-6b7b34d
7975f5b5-3d47f4f4
e1368496-f23d1dea
a46926f0-f13a635d
dba92675-2e96769
16d8ab13-64207fb2
dd4eb913-f23d1dea
497bad44-3f4a17df
fb50494f-72cac062
547e761a-3f4a17df
89f843ca-584aa0f
5e31bb48-75513c66
abb39457-139804fd
a9deeaf7-85023da1
a461b170-92b55ebe
dd82d379-fa554100
fc330dc9-377be55a
def27776-f3cd3248
6598898b-48009d14
b53f3ef9-a1f16059
c3e768ab-f9a43703

Attachments:
Screen recording 2021-06-03 18.01.28.webm 427 KB

[chrisw.g7nbp via monorail]

Comment #113 on issue 1197152 by chrisw...@gmail.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c113

Device by the way is Acer R13 CB5-13T if that helps narrow down the search

[chrisw.g7nbp via monorail]

Comment #114 on issue 1197152 by chrisw...@gmail.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c114

Device is Acer R13 CB5-312T if that helps

[csmartdalton via monorail]

Comment #115 on issue 1197152 by csmart...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c115

Unfortunately the device name does not help Skia. Can you please copy the strings returned by glGetString(GL_RENDERER) and glGetString(GL_VENDOR)? These are the strings we need in order to identify a specific device.

[kbr via monorail]

Comment #116 on issue 1197152 by k...@chromium.org: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c116

Actually if you go to chrome's about:gpu page and use the "Copy report to clipboard" button, and paste the output here, that'll contain all the needed information.

[chrisw.g7nbp via monorail]

Comment #117 on issue 1197152 by chrisw...@gmail.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c117

Graphics Feature Status
Canvas: Hardware accelerated
Compositing: Hardware accelerated
Multiple Raster Threads: Enabled
Out-of-process Rasterization: Hardware accelerated
OpenGL: Enabled
Rasterization: Hardware accelerated
Skia Renderer: Disabled
Video Decode: Hardware accelerated
Vulkan: Disabled
WebGL: Hardware accelerated
WebGL2: Hardware accelerated
Driver Bug Workarounds
clear_uniforms_before_first_program_use
exit_on_context_lost
use_virtualized_gl_contexts
disabled_extension_GL_KHR_blend_equation_advanced
disabled_extension_GL_KHR_blend_equation_advanced_coherent
Problems Detected
Share group-related crashes and poor context switching perf on Imagination drivers
Applied Workarounds: use_virtualized_gl_contexts
Clear uniforms before first program use on all platforms: 124764, 349137
Applied Workarounds: clear_uniforms_before_first_program_use
Disable KHR_blend_equation_advanced until cc shaders are updated: 661715
Applied Workarounds: disable(GL_KHR_blend_equation_advanced), disable(GL_KHR_blend_equation_advanced_coherent)
Context lost recovery often fails on PowerVR on CrOS.: 1010121
Applied Workarounds: exit_on_context_lost
Context loss recovery often fails on Chrome OS: 992286, 1177986
Applied Workarounds: exit_on_context_lost
Version Information
Data exported 2021-06-03T21:45:18.742Z
Chrome version Chrome/91.0.4472.81
Operating system Linux 13904.41.0
Software rendering list URL https://chromium.googlesource.com/chromium/src/+/0427ab82b616937c2ee3b4a2cc2e413dfce180f1/gpu/config/software_rendering_list.json
Driver bug list URL https://chromium.googlesource.com/chromium/src/+/0427ab82b616937c2ee3b4a2cc2e413dfce180f1/gpu/config/gpu_driver_bug_list.json
ANGLE commit id 788efd1c721a
2D graphics backend Skia/91 71fcb16fc206466727cb49cbf318b765f8a9d5a7

Command Line /opt/google/chrome/chrome --use-gl=egl --enable-native-gpu-memory-buffers --platform-disallows-chromeos-direct-video-decoder --enable-webgl-image-chromium --enable-features=Pepper3DImageChromium,SmartDim,UmaStorageDimensions,Crostini,CrostiniEnableDlc --gpu-sandbox-failures-fatal=yes --enable-logging --log-level=1 --use-cras --enable-wayland-server --user-data-dir=/home/chronos --device-help-content-id=elm --login-profile=user --has-chromeos-keyboard --enable-touchview --guest-wallpaper-large=/usr/share/chromeos-assets/wallpaper/guest_large.jpg --guest-wallpaper-small=/usr/share/chromeos-assets/wallpaper/guest_small.jpg --child-wallpaper-large=/usr/share/chromeos-assets/wallpaper/child_large.jpg --child-wallpaper-small=/usr/share/chromeos-assets/wallpaper/child_small.jpg --default-wallpaper-large=/usr/share/chromeos-assets/wallpaper/elm_large.jpg --default-wallpaper-small=/usr/share/chromeos-assets/wallpaper/elm_small.jpg --default-wallpaper-is-oem --enable-consumer-kiosk --ash-side-volume-button-position={"region":"keyboard","side":"right"} --ozone-nnpalm-properties={} --disable-features=InstantTethering --arc-availability=officially-supported --enterprise-enrollment-initial-modulus=15 --enterprise-enrollment-modulus-limit=19 --enable-crashpad --login-manager --vmodule=*arc/*=1,*night_light*=1,*/chromeos/power/auto_screen_brightness/*=1,*/ash/wm/tablet_mode/*=1,wizard_controller=1,*/webui/chromeos/login/*=1,*/browser/chromeos/login/screens/*=1,enrollment_screen_handler=1,*/browser/chromeos/login/enrollment/*=1,*/ui/ozone/*=1,*/ui/display/manager/chromeos/*=1 --enable-features=Pepper3DImageChromium,SmartDim,UmaStorageDimensions,Crostini,CrostiniEnableDlc --disable-features=InstantTethering --flag-switches-begin --flag-switches-end --origin-trial-disabled-features=SecurePaymentConfirmation --ui-compositor-memory-limit-when-visible-mb=512

Driver Information
Initialization time 91
In-process GPU false
Passthrough Command Decoder false
Sandboxed true
GPU0 VENDOR= 0x0000 [Imagination Technologies], DEVICE=0x0000 [PowerVR Rogue GX6250] *ACTIVE*
Optimus false
AMD switchable false
Driver vendor
Driver version 1.13
GPU CUDA compute capability major version 0
Pixel shader version 3.20
Vertex shader version 3.20
Max. MSAA samples 8
Machine model name
Machine model version
GL_VENDOR Imagination Technologies
GL_RENDERER PowerVR Rogue GX6250
GL_VERSION OpenGL ES 3.2 build 1.13@5824814
GL_EXTENSIONS GL_ANDROID_extension_pack_es31a GL_EXT_blend_minmax GL_EXT_buffer_storage GL_EXT_clip_control GL_EXT_color_buffer_float GL_EXT_color_buffer_half_float GL_EXT_conservative_depth GL_EXT_copy_image GL_EXT_discard_framebuffer GL_EXT_draw_buffers GL_EXT_draw_buffers_indexed GL_EXT_draw_elements_base_vertex GL_EXT_EGL_image_array GL_EXT_float_blend GL_EXT_geometry_point_size GL_EXT_geometry_shader GL_EXT_gpu_shader5 GL_EXT_memory_object GL_EXT_memory_object_fd GL_EXT_multi_draw_arrays GL_EXT_multisampled_render_to_texture GL_EXT_multisampled_render_to_texture2 GL_EXT_occlusion_query_boolean GL_EXT_polygon_offset_clamp GL_EXT_primitive_bounding_box GL_EXT_pvrtc_sRGB GL_EXT_read_format_bgra GL_EXT_robustness GL_EXT_separate_shader_objects GL_EXT_shader_framebuffer_fetch GL_EXT_shader_group_vote GL_EXT_shader_implicit_conversions GL_EXT_shader_io_blocks GL_EXT_shader_non_constant_global_initializers GL_EXT_shader_pixel_local_storage GL_EXT_shader_pixel_local_storage2 GL_EXT_shader_texture_lod GL_EXT_shadow_samplers GL_EXT_sparse_texture GL_EXT_sRGB_write_control GL_EXT_tessellation_point_size GL_EXT_tessellation_shader GL_EXT_texture_border_clamp GL_EXT_texture_buffer GL_EXT_texture_cube_map_array GL_EXT_texture_filter_anisotropic GL_EXT_texture_format_BGRA8888 GL_EXT_texture_format_sRGB_override GL_EXT_texture_rg GL_EXT_texture_shadow_lod GL_EXT_texture_sRGB_decode GL_EXT_texture_sRGB_R8 GL_EXT_texture_sRGB_RG8 GL_EXT_YUV_target GL_IMG_framebuffer_downsample GL_IMG_multisampled_render_to_texture GL_IMG_program_binary GL_IMG_read_format GL_IMG_shader_binary GL_IMG_texture_compression_pvrtc GL_IMG_texture_compression_pvrtc2 GL_IMG_texture_filter_cubic GL_IMG_texture_format_BGRA8888 GL_IMG_texture_npot GL_KHR_blend_equation_advanced GL_KHR_blend_equation_advanced_coherent GL_KHR_debug GL_KHR_robustness GL_KHR_texture_compression_astc_ldr GL_OES_compressed_ETC1_RGB8_texture GL_OES_depth24 GL_OES_depth_texture GL_OES_draw_buffers_indexed GL_OES_draw_elements_base_vertex GL_OES_EGL_image GL_OES_EGL_image_external GL_OES_EGL_image_external_essl3 GL_OES_EGL_sync GL_OES_element_index_uint GL_OES_fragment_precision_high GL_OES_geometry_point_size GL_OES_geometry_shader GL_OES_get_program_binary GL_OES_gpu_shader5 GL_OES_mapbuffer GL_OES_packed_depth_stencil GL_OES_required_internalformat GL_OES_rgb8_rgba8 GL_OES_sample_shading GL_OES_sample_variables GL_OES_shader_image_atomic GL_OES_shader_io_blocks GL_OES_shader_multisample_interpolation GL_OES_standard_derivatives GL_OES_surfaceless_context GL_OES_tessellation_point_size GL_OES_tessellation_shader GL_OES_texture_border_clamp GL_OES_texture_buffer GL_OES_texture_cube_map_array GL_OES_texture_float GL_OES_texture_half_float GL_OES_texture_npot GL_OES_texture_stencil8 GL_OES_texture_storage_multisample_2d_array GL_OES_vertex_array_object GL_OES_vertex_half_float GL_OVR_multiview GL_OVR_multiview2 GL_OVR_multiview_multisampled_render_to_texture
Disabled Extensions GL_KHR_blend_equation_advanced GL_KHR_blend_equation_advanced_coherent
Disabled WebGL Extensions
Window system binding vendor Mesa Project
Window system binding version 1.4
Window system binding extensions EGL_ANDROID_native_fence_sync EGL_EXT_create_context_robustness EGL_EXT_image_dma_buf_import EGL_EXT_image_dma_buf_import_modifiers EGL_IMG_context_priority EGL_KHR_config_attribs EGL_KHR_create_context EGL_KHR_fence_sync EGL_KHR_get_all_proc_addresses EGL_KHR_gl_renderbuffer_image EGL_KHR_gl_texture_2D_image EGL_KHR_gl_texture_cubemap_image EGL_KHR_image_base EGL_KHR_no_config_context EGL_KHR_reusable_sync EGL_KHR_surfaceless_context EGL_EXT_pixel_format_float EGL_KHR_wait_sync EGL_MESA_configless_context EGL_MESA_drm_image EGL_MESA_image_dma_buf_export
Direct rendering version unknown
Reset notification strategy 0x8252
GPU process crash count 0
gfx::BufferFormats supported for allocation and texturing R_8: not supported, R_16: not supported, RG_88: not supported, BGR_565: not supported, RGBA_4444: not supported, RGBX_8888: not supported, RGBA_8888: not supported, BGRX_8888: not supported, BGRA_1010102: not supported, RGBA_1010102: not supported, BGRA_8888: not supported, RGBA_F16: not supported, YVU_420: not supported, YUV_420_BIPLANAR: not supported, P010: not supported
Compositor Information
Tile Update Mode One-copy
Partial Raster Enabled
GpuMemoryBuffers Status
R_8 CAMERA_AND_CPU_READ_WRITE
R_16 Software only
RG_88 Software only
BGR_565 GPU_READ
RGBA_4444 Software only
RGBX_8888 GPU_READ, SCANOUT, SCANOUT_CPU_READ_WRITE, SCANOUT_FRONT_RENDERING
RGBA_8888 GPU_READ, SCANOUT, SCANOUT_CPU_READ_WRITE, SCANOUT_FRONT_RENDERING
BGRX_8888 GPU_READ, SCANOUT, SCANOUT_CPU_READ_WRITE, SCANOUT_FRONT_RENDERING
BGRA_1010102 SCANOUT
RGBA_1010102 SCANOUT
BGRA_8888 GPU_READ, SCANOUT, SCANOUT_CPU_READ_WRITE, GPU_READ_CPU_READ_WRITE, SCANOUT_FRONT_RENDERING
RGBA_F16 Software only
YVU_420 GPU_READ, SCANOUT_VEA_CPU_READ, VEA_READ_CAMERA_AND_CPU_READ_WRITE
YUV_420_BIPLANAR SCANOUT_CAMERA_READ_WRITE, SCANOUT_CPU_READ_WRITE, GPU_READ_CPU_READ_WRITE, SCANOUT_VEA_CPU_READ, SCANOUT_FRONT_RENDERING, VEA_READ_CAMERA_AND_CPU_READ_WRITE
P010 Software only
Display(s) Information
Info Display[2785062953156608] bounds=[0,0 2194x1234], workarea=[0,0 2194x1186], scale=0.875, rotation=0, panel_rotation=0 internal.
Color space (all) {primaries:BT709, transfer:IEC61966_2_1, matrix:RGB, range:FULL}
Buffer format (all) BGRA_8888
SDR white level in nits 100
Bits per color component 8
Bits per pixel 24
Refresh Rate in Hz 60
Info Display[24967369140605953] bounds=[-1920,0 1920x1080], workarea=[-1920,0 1920x1032], scale=1, rotation=0, panel_rotation=0 external.
Color space (all) {primaries:BT709, transfer:IEC61966_2_1, matrix:RGB, range:FULL}
Buffer format (all) BGRA_8888
SDR white level in nits 100
Bits per color component 8
Bits per pixel 24
Refresh Rate in Hz 60
Video Acceleration Information
Decode h264 baseline 64x64 to 4096x2304 pixels
Decode h264 main 64x64 to 4096x2304 pixels
Decode h264 high 64x64 to 4096x2304 pixels
Decode vp8 64x64 to 4096x2304 pixels
Decode vp9 profile0 64x64 to 4096x2304 pixels
Encode h264 baseline 0x0 to 1920x1088 pixels, and/or 30.000 fps
Encode h264 main 0x0 to 1920x1088 pixels, and/or 30.000 fps
Encode h264 extended 0x0 to 1920x1088 pixels, and/or 30.000 fps
Encode h264 high 0x0 to 1920x1088 pixels, and/or 30.000 fps
Encode vp8 0x0 to 1920x1088 pixels, and/or 30.000 fps
Vulkan Information
Device Performance Information
Log Messages
2021-06-03T17:16:13.538489Z ERROR chrome[22619:22619]: [gbm_surface_factory.cc(199)] : Couldn't create Gbm Device at /dev/dri/renderD128
2021-06-03T17:16:13.541034Z WARNING chrome[22619:22619]: [gbm_surface_factory.cc(118)] : Platform device extensions not found.
2021-06-03T17:16:13.668767Z VERBOSE1 chrome[22619:22632]: [drm_device_manager.cc(60)] : Primary DRM device added: /sys/devices/platform/soc/14000000.clock-controller/drm/card2
2021-06-03T17:16:14.198265Z ERROR chrome[22619:22632]: [drm_gpu_display_manager.cc(237)] : There is no display with ID 24967369140605953
2021-06-03T17:16:14.568989Z ERROR chrome[22619:22632]: [drm_thread_proxy.cc(45)] : Unable to initialize GBM for /sys/devices/platform/soc/13000000.mfgsys-gpu/drm/card1: Operation not permitted (1)
2021-06-03T17:16:14.569263Z VERBOSE1 chrome[22619:22632]: [drm_device_manager.cc(55)] : Could not initialize DRM device for /sys/devices/platform/soc/13000000.mfgsys-gpu/drm/card1
2021-06-03T17:16:14.675682Z ERROR chrome[22619:22632]: [drm_thread_proxy.cc(45)] : Unable to initialize GBM for /sys/devices/platform/vgem/drm/card0: Operation not permitted (1)
2021-06-03T17:16:14.675934Z VERBOSE1 chrome[22619:22632]: [drm_device_manager.cc(55)] : Could not initialize DRM device for /sys/devices/platform/vgem/drm/card0
2021-06-03T17:16:15.938454Z VERBOSE1 chrome[22619:22632]: [drm_gpu_display_manager.cc(192)] : DRM configuring: device=/sys/devices/platform/soc/14000000.clock-controller/drm/card2 crtc=40 connector=29 origin=0,0 size=1920x1080 refresh_rate=60Hz
2021-06-03T17:16:15.938678Z VERBOSE1 chrome[22619:22632]: [drm_gpu_display_manager.cc(192)] : DRM configuring: device=/sys/devices/platform/soc/14000000.clock-controller/drm/card2 crtc=49 connector=31 origin=0,1140 size=1920x1080 refresh_rate=60Hz
2021-06-03T20:31:01.770823Z WARNING chrome[22619:22619]: [ipc_message_attachment_set.cc(49)] : MessageAttachmentSet destroyed with unconsumed attachments: 0/1

[domikakeszei via monorail]

Comment #118 on issue 1197152 by domika...@gmail.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c118

The report is attached below as the output didn't fit in a comment. This is for for 91.0.4472.81 / 13904.41.0 (Official Build) beta-channel hana

Attachments:
hana_GPU_report.txt 51.2 KB

[csmartdalton via monorail]

Comment #119 on issue 1197152 by csmart...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c119

Thanks for the reports!

Can someone try https://skia-review.googlesource.com/c/skia/+/415516 and see if it fixes the bug?

[evanbenn via monorail]

Comment #120 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c120

Hi csmartdalton, the file ../src/gpu/gl/GrGLCaps.cpp does not exist on R91, so the cherry pick does not apply. I will try to manually apply this. Do you mind creating an M91 change?

[evanbenn via monorail]

Comment #121 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c121

No thats my error, apparently there are 2 skia folders in chromium and I was in the wrong one. disregart #120

[evanbenn via monorail]

Comment #122 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c122

Ok with a small change I cherry picked to M91 and the bug is gone on hana: https://skia-review.googlesource.com/c/skia/+/415536

[evanbenn via monorail]

Updates:
Labels: Merge-Request-91

Comment #123 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c123

(No comment was entered for this change.)

[evanbenn via monorail]

Comment #124 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c124

Requesting merge for this patch: https://skia-review.googlesource.com/c/skia/+/415536

This is a stable blocker for hana/elm chromeos devices, see also : https://b.corp.google.com/issues/187364475.

The fix is not landed on canary / dev / TOT. (csmartdalton can we get that ball rolling please?)

[Git Watcher via monorail]

Updates:
Labels: merge-merged-m91

Comment #125 on issue 1197152 by Git Watcher: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c125

The following revision refers to this bug:
https://skia.googlesource.com/skia/+/1a01201b77a9b4e2f9fadd8d863832f1840bacd2

commit 1a01201b77a9b4e2f9fadd8d863832f1840bacd2
Author: Chris Dalton <csmart...@google.com>
Date: Thu Jun 03 22:30:18 2021

Disable standard derivatives on PowerVR Rogue

Bug: chromium:1197152
Change-Id: I542b141217bfcb7590c80bb8144714479535f984
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/415536
Commit-Queue: Evan Benn <evan...@google.com>
Commit-Queue: Chris Dalton <csmart...@google.com>
Reviewed-by: John Stiles <johns...@google.com>

[modify] https://crrev.com/1a01201b77a9b4e2f9fadd8d863832f1840bacd2/src/gpu/gl/GrGLCaps.cpp

[chrisw.g7nbp via monorail]

Comment #126 on issue 1197152 by chrisw...@gmail.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c126

Has the patch reached the dev channel yet??

Because my device has just updated to Version 92.0.4515.36 (Official Build) beta (32-bit) and problem remains

[chrisw.g7nbp via monorail]

Comment #127 on issue 1197152 by chrisw...@gmail.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c127

Has the patch reached the beta channel yet??

[domikakeszei via monorail]

Comment #128 on issue 1197152 by domika...@gmail.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c128

I read the changelogs and it doesn't seem like it

[evanbenn via monorail]

Updates:
Labels: -M-91
Owner: mar...@chromium.org

Comment #129 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c129

I flashed R91-13904.52.0 onto my hana and this is mitigated, although it was not obvious from the commit that it had all merged in (bug filed for that ui issue).

R92 did not land yet. (can someone with chrome priveledges CQ+2 that please: https://skia-review.googlesource.com/c/skia/+/415537

R93-14017.0.0 is also good, but I think the mitigation from stephane earlier that landed is still causing mass graphical glitches.

We can continue using the internal bug to track the real fix with the driver. After R93 and R92 are fixed we can close this.

[evanbenn via monorail]

Comment #130 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c130

Revert here: https://chromium-review.googlesource.com/c/chromium/src/+/2942525

[dgagnon via monorail]

Updates:
Labels: -Merge-Request-91 Merge-Approved-91

Comment #131 on issue 1197152 by dga...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c131

Has this issue been confirmed to fix the issue on M91? I'm not clear if the merge request for M91 is ready for not?

[songsuk via monorail]

Comment #132 on issue 1197152 by son...@chromium.org: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c132

Verified the fix on 93.0.4534.0 / 14019.0.0 - Hana. I'm able to input on the omnibox.

[evanbenn via monorail]

Comment #133 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c133

@dgagnon I verified the fix on R91: https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c129

I requested the merge in https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c124 but it was submitted without merge approval: https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c125

On M91 the fix is landed: https://skia-review.googlesource.com/c/skia/+/415537

On M92 the fix is NOT landed, I do not have ability to CQ+2: https://skia-review.googlesource.com/c/skia/+/415537 I requested someone to do that: https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c129

On M93 the same applies, waiting for +2: https://skia-review.googlesource.com/c/skia/+/415516

M93 TOT is working because of this patch that also causes mass graphical issues for hana: https://chromium-review.googlesource.com/c/chromium/src/+/2921010
I created a revert and CQ+2 it just now: https://chromium-review.googlesource.com/c/chromium/src/+/2942525

[Git Watcher via monorail]

Comment #134 on issue 1197152 by Git Watcher: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c134

The following revision refers to this bug:

https://chromium.googlesource.com/chromium/src/+/c9557bbd7a29fbac8ab9bfd30f48bce1c3325c6b

commit c9557bbd7a29fbac8ab9bfd30f48bce1c3325c6b
Author: Evan Benn <evan...@chromium.org>
Date: Mon Jun 07 23:59:57 2021

Revert "Disable GPU rasterization on Chrome OS on imagination"

This reverts commit abc166ee652e32cfc2b36f88f7da9bea14821cd4.

Reason for revert: this causes graphical glitching on hana, and a better fix has been landed in skia: https://skia-review.googlesource.com/c/skia/+/415516

Original change's description:
> Disable GPU rasterization on Chrome OS on imagination
>
> GPU rasterization seems to be causing issues with the UI. Since nothing
> changed on the driver side, we aren't sure yet why this is happening.
> But since this causes blocker bugs, let's disable GPU rasterization on
> this platform until things are figured out.
>
> Bug:1197152
> Test:untested
>
> Change-Id: Ia1ee4185d05fd8deae657cc456b2cb9974d5bf95
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2921010
> Commit-Queue: Stéphane Marchesin <mar...@chromium.org>
> Auto-Submit: Stéphane Marchesin <mar...@chromium.org>
> Reviewed-by: Brian Ho <h...@chromium.org>
> Reviewed-by: Kenneth Russell <k...@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#886945}

Bug: 1197152
Change-Id: I85556bdc872b9b1ec7ba16fea6ab542b1b832ac9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2942525
Bot-Commit: Rubber Stamper <rubber-...@appspot.gserviceaccount.com>
Reviewed-by: Brian Ho <h...@chromium.org>
Reviewed-by: Kenneth Russell <k...@chromium.org>
Commit-Queue: Evan Benn <evan...@chromium.org>
Cr-Commit-Position: refs/heads/master@{#890008}

[modify] https://crrev.com/c9557bbd7a29fbac8ab9bfd30f48bce1c3325c6b/gpu/config/software_rendering_list.json

[Git Watcher via monorail]

Updates:
Labels: merge-merged-m92

Comment #135 on issue 1197152 by Git Watcher: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c135

The following revision refers to this bug:

https://skia.googlesource.com/skia/+/81ce29695f6e75137e3ff4ecdaa595267cf9c34b

commit 81ce29695f6e75137e3ff4ecdaa595267cf9c34b

Author: Chris Dalton <csmart...@google.com>
Date: Thu Jun 03 22:30:18 2021

Disable standard derivatives on PowerVR Rogue

Bug: chromium:1197152
Change-Id: I542b141217bfcb7590c80bb8144714479535f984

Reviewed-on: https://skia-review.googlesource.com/c/skia/+/415537
Reviewed-by: John Stiles <johns...@google.com>
Commit-Queue: Brian Osman <brian...@google.com>

[modify] https://crrev.com/81ce29695f6e75137e3ff4ecdaa595267cf9c34b/src/gpu/gl/GrGLCaps.cpp

[Git Watcher via monorail]

Comment #136 on issue 1197152 by Git Watcher: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c136

The following revision refers to this bug:

https://skia.googlesource.com/skia/+/3a82b7dee6c46ca4f282f7309e68a1e23e89408c

commit 3a82b7dee6c46ca4f282f7309e68a1e23e89408c

Author: Chris Dalton <csmart...@google.com>
Date: Thu Jun 03 22:30:18 2021

Disable standard derivatives on PowerVR Rogue

Bug: chromium:1197152
Change-Id: I542b141217bfcb7590c80bb8144714479535f984

Reviewed-on: https://skia-review.googlesource.com/c/skia/+/415516
Commit-Queue: Brian Osman <brian...@google.com>
Reviewed-by: John Stiles <johns...@google.com>

[modify] https://crrev.com/3a82b7dee6c46ca4f282f7309e68a1e23e89408c/src/gpu/gl/GrGLCaps.cpp

[Git Watcher via monorail]

Comment #137 on issue 1197152 by Git Watcher: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c137

The following revision refers to this bug:

https://chromium.googlesource.com/chromium/src/+/9342b766f614badaebea758ff81446f7f159d720

commit 9342b766f614badaebea758ff81446f7f159d720
Author: chromium-autoroll <chromium...@skia-public.iam.gserviceaccount.com>
Date: Tue Jun 08 03:14:39 2021

Roll Skia from 2c9a6ec3a0d0 to 3a82b7dee6c4 (1 revision)

https://skia.googlesource.com/skia.git/+log/2c9a6ec3a0d0..3a82b7dee6c4

2021-06-08 csmart...@google.com Disable standard derivatives on PowerVR Rogue

If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/skia-autoroll
Please CC robertp...@google.com on the revert to ensure that a human
is aware of the problem.

To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/master/autoroll/README.md

Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux-blink-rel;luci.chromium.try:linux-chromeos-compile-dbg;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win_optional_gpu_tests_rel
Cq-Do-Not-Cancel-Tryjobs: true
Bug: chromium:1197152
Tbr: robertp...@google.com
Change-Id: I12737437300d36640a3ef035347bccb050f6b6f8
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2946520
Commit-Queue: chromium-autoroll <chromium...@skia-public.iam.gserviceaccount.com>
Bot-Commit: chromium-autoroll <chromium...@skia-public.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#890075}

[modify] https://crrev.com/9342b766f614badaebea758ff81446f7f159d720/DEPS

[evanbenn via monorail]

Comment #138 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c138

re dgagnon@ https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c131

Looks like all patches are now in, I can test all branches tomorrow once the builds are done. Thanks everyone!

For the longer term fix the internal bug is active: https://b.corp.google.com/issues/187364475

[songsuk via monorail]

Comment #139 on issue 1197152 by son...@chromium.org: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c139

Verified the fix on 13904.54.0/91.0.4472.100 - Hana, Elm

[evanbenn via monorail]

Comment #140 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c140

R92-13982.18.0 92.0.4515.47 is still broken, wait for next build.

R93 / TOT changes also not percolated yet.

[evanbenn via monorail]

Updates:
Cc: kyle...@chromium.org mar...@chromium.org osh...@chromium.org

Comment #141 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c141

Issue 1216176 has been merged into this issue.

[chrisw.g7nbp via monorail]

Comment #142 on issue 1197152 by chrisw...@gmail.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c142

Guys, I have 26 ELM ARM based evices out of total of around 40 initial test devices out in the field globally that are hit by this problem, spread around the world for our charity. Most crucially the CEOs own device. Any prospect of a full migration to gsuite (potentially around 5000 devices!) is getting to be unlikely because after all the positive things Ive said and demonstrated in the test group with chromebooks and gsuite, this bug is the single most visible problem not being fixed to the person who holds the purse strings.

I know in the grand scheme of things this is just an annoyance as far as bugs go, but for the migration project Ive spent the last year working on, this is a dealbreaker when the CEO says "I cant type in the box". Prettey much all of the initial test devices I pushed out are on 92.0.4515.36 (Official Build) beta (32-bit) and after nearly a month now there is still no fix landing on auto updating devices.

This is making me, the IT department, and chrome OS look really bad.

[evanbenn via monorail]

Comment #143 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c143

Hi Chris. this bug never hit the stable channel in my understanding. Can you confirm what version your devices are hitting the issue on?

[chrisw.g7nbp via monorail]

Comment #144 on issue 1197152 by chrisw...@gmail.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c144

As commented above they are now all currently on 92.0.4515.36 Beta channel. (Historically they needed to be on beta for particular functionality from a flag as part of the test scenario roll out)

Google Chrome 92.0.4515.36 (Official Build) beta (32-bit)
Revision 565ef04a3927b12aca6ad6ad8f01f28dd0fa41ef-refs/branch-heads/4515@{#142}
Platform 13982.7.0 (Official Build) beta-channel elm
Firmware Version Google_Elm.8438.184.0

I can't easily take the devices out of the beta channel into stable, as they have a custom linux container setup which would be a real pain to set up again for the non-tech users when the chan change forces a power wash. The test devices are spread physically half way around the globe, and are not currently enterprise enrolled - so its going to be a case of getting the devices sent back to me for updating, or replacing those devices with something intel based,

A huge selling point I have been pitching for the proposed migration to gsuite is that these sort of issues just dont happen with chromebooks, but right now Im having a hard time convincing anyone of this and the proof of concept project is going ro sink.

[songsuk via monorail]

Comment #145 on issue 1197152 by son...@chromium.org: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c145

Unable to reproduce the issue on 92.0.4515.56/13982.23.0 - Hana, Elm.

[chrisw.g7nbp via monorail]

Comment #146 on issue 1197152 by chrisw...@gmail.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c146

What can I provide to move forward?

[imasogie via monorail]

Comment #147 on issue 1197152 by imas...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c147

Issue 1218671 has been merged into this issue.

[evanbenn via monorail]

Updates:
Owner: evan...@chromium.org
Status: Verified

Comment #148 on issue 1197152 by evan...@google.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c148

The fix is applied at these versions:

91.0.4472.101 | 13904.55.0
92.0.4515.48 | 13982.19.0
93.0.4537.0 | 14026.0.0

[newcomer via monorail]

Comment #149 on issue 1197152 by newc...@chromium.org: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c149

Do we know why the bug was being triggered by the chrome/VERSION int changing? We should probably try to understand this bug trigger. See comment #85 for more info.

[chrisw.g7nbp via monorail]

Comment #150 on issue 1197152 by chrisw...@gmail.com: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c150

Version 92.0.4515.56 update has landed here this AM.

Issue seems fixed with both light and dark themes.

[Git Watcher via monorail]

Comment #151 on issue 1197152 by Git Watcher: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c151

The following revision refers to this bug:

https://skia.googlesource.com/skia/+/2422ca10c26c25b64612878015a6128b16581911

commit 2422ca10c26c25b64612878015a6128b16581911
Author: Brian Salomon <bsal...@google.com>
Date: Thu Jul 08 13:59:58 2021

Limit PowerVRRougue shader derivative workaround by driver version

The shader optimization bug has been fixed in newer drivers.

Bug: b/187364475
Bug: chromium:1197152
Change-Id: I013d022cc0d8a4b4739160c3e651b3ce06ee5f19
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/425838
Reviewed-by: Brian Osman <brian...@google.com>
Reviewed-by: Chris Dalton <csmart...@google.com>
Commit-Queue: Brian Salomon <bsal...@google.com>

[modify] https://crrev.com/2422ca10c26c25b64612878015a6128b16581911/src/gpu/gl/GrGLCaps.cpp
[modify] https://crrev.com/2422ca10c26c25b64612878015a6128b16581911/src/gpu/gl/GrGLUtil.cpp
[modify] https://crrev.com/2422ca10c26c25b64612878015a6128b16581911/src/gpu/gl/GrGLUtil.h

[Git Watcher via monorail]

Comment #152 on issue 1197152 by Git Watcher: Unable to input on the omnibox - Hana/Elm
https://bugs.chromium.org/p/chromium/issues/detail?id=1197152#c152

The following revision refers to this bug:

https://chromium.googlesource.com/chromium/src/+/8fb5341046985f9f1ac98236364a155eff138419

commit 8fb5341046985f9f1ac98236364a155eff138419
Author: chromium-autoroll <chromium...@skia-public.iam.gserviceaccount.com>
Date: Wed Jul 14 15:11:40 2021

Roll Skia from d421a286df31 to 2422ca10c26c (1 revision)

https://skia.googlesource.com/skia.git/+log/d421a286df31..2422ca10c26c

2021-07-14 bsal...@google.com Limit PowerVRRougue shader derivative workaround by driver version

If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/skia-autoroll

Please CC johns...@google.com on the revert to ensure that a human

is aware of the problem.

To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug

Documentation for the AutoRoller is here:

https://skia.googlesource.com/buildbot/+doc/main/autoroll/README.md

Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux-blink-rel;luci.chromium.try:linux-chromeos-compile-dbg;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win_optional_gpu_tests_rel
Cq-Do-Not-Cancel-Tryjobs: true
Bug: chromium:1197152

Tbr: johns...@google.com
Change-Id: Iee520430a125ad457ed23288562c3ae0f2ce68fb
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3025797

Commit-Queue: chromium-autoroll <chromium...@skia-public.iam.gserviceaccount.com>
Bot-Commit: chromium-autoroll <chromium...@skia-public.iam.gserviceaccount.com>

Cr-Commit-Position: refs/heads/master@{#901499}

[modify] https://crrev.com/8fb5341046985f9f1ac98236364a155eff138419/DEPS