1754350 - Visible seams on https://codepen.io/amit_sheen/pen/YzxbwOx

Status: VERIFIED FIXED

(Core :: Graphics: WebRender, defect)

Description

[Mayank Bansal]

Create new profile
go to https://codepen.io/amit_sheen/pen/YzxbwOx

AR: Visible seams on the page

Reression:
2022-02-09T08:43:53.998000: INFO : application_version: 99.0a1
2022-02-09T08:43:53.998000: INFO : platform_buildid: 20220207222633
2022-02-09T08:43:53.998000: INFO : platform_changeset: b7f88e5c537bb7d64ba9d4cb58b052650d510638
2022-02-09T08:43:53.998000: INFO : platform_repository: https://hg.mozilla.org/integration/autoland
2022-02-09T08:43:53.998000: INFO : platform_version: 99.0a1
2022-02-09T08:44:06.493000: INFO : Narrowed integration regression window from [8e9c1ffd, 3f6deedb] (3 builds) to [8e9c1ffd, b7f88e5c] (2 builds) (~1 steps left)
2022-02-09T08:44:06.493000: DEBUG : Starting merge handling...
2022-02-09T08:44:06.493000: DEBUG : Using url: https://hg.mozilla.org/integration/autoland/json-pushes?changeset=b7f88e5c537bb7d64ba9d4cb58b052650d510638&full=1
2022-02-09T08:44:06.493000: DEBUG : redo: attempt 1/3
2022-02-09T08:44:06.493000: DEBUG : redo: retry: calling _default_get with args: ('https://hg.mozilla.org/integration/autoland/json-pushes?changeset=b7f88e5c537bb7d64ba9d4cb58b052650d510638&full=1',), kwargs: {}, attempt #1
2022-02-09T08:44:06.493000: DEBUG : urllib3.connectionpool: Resetting dropped connection: hg.mozilla.org
2022-02-09T08:44:08.909000: DEBUG : urllib3.connectionpool: https://hg.mozilla.org:443 "GET /integration/autoland/json-pushes?changeset=b7f88e5c537bb7d64ba9d4cb58b052650d510638&full=1 HTTP/1.1" 200 None
2022-02-09T08:44:08.941000: DEBUG : Found commit message:
Bug 1749380 - Improve how WR handles bounding rects for off-screen surfaces r=gfx-reviewers,kvark

This patch introduces a number of subtle but important changes
to how we deal with off-screen surfaces. The overall goals are:

• Improve rendering correctness in a number of edge cases.
• Begin reducing complexity related to surfaces, scaling
  factors, surface size adjustments and clipping.
• Improve CPU performance by removing some per-primitive work.
• Simplify implementation of future SVG and CSS filters by
  having explicit support for picture rects + inflation regions.
• Lay the groundwork for caching child picture surfaces,
  reduction of per-primitive work during visibility pass,
  simplifying picture code.

Unfortunately, the nature of the changes make it impossible to
split up in to small isolated patches. Details below:

• Introduce LocalRectKind concept. This allows us to separate
  out the bounding rect of the surface (a group of primitives
  backed by a texture) from the bounding rect of the picture
  compositing that surface (e.g. a drop-shadow which draws the
  surface once at the local origin and once at a specific offset

  • blur-radius). This fixes a number of correctness bugs we have
    related to culling, clipping, invalidation regions of complex
    primitives such as drop-shadows and blur filters. Importantly,
    it makes it simpler to implement (or fix) SVG filter chains,
    backdrop-filter implementations.

• Establish raster roots for all off-screen surfaces. Every off-screen
  surface uses the spatial node of the enclosing stacking context as
  a coordinate system root, ensuring that each off-screen surface is
  drawn in a 2D coordinate system, with appropriate scaling factors
  applied to ensure high quality rendering. The primary goal is to make
  it possible to correctly inflate and clip off-screen surfaces, removing
  some correctness issues we currently have with complex filters interacting
  with transforms. The initial work here doesn't reduce complexity a huge
  amount, but will allow us to simplify large parts of the picture/surface
  handling code in future, as well as simplify a number of shaders that
  currently must handle arbitrarily complex transform matrices. This will
  also allow us to simplify the implementation of features such as
  mix-blend-mode and backdrop-filter, which rely on readback and UV mapping
  from the parent surface.

• Remove concepts of estimated and precise local rects for pictures. This
  is both a performance optimization and a code simplification. Instead, we
  only determine the estimated local rect during bounding rect propagation,
  and rely on the clipping regions from the tile dirty regions to reduce which
  parts of the picture we allocate if drawing to an off-screen surface. This
  removes some per-primitive work during the visibility pass, and also means
  we can rely on the final picture bounding rect from the start of the visibility
  pass. This also removes much of the complexity in take_context where we
  previously determined surface scale factors and device pixel ratio - instead
  these can be determined earlier during propagate_bounding_rects.

• Remove some complexity in update_prim_visibility. This is still recursive,
  but follow up patches will aim to remove this recursion and integrate this
  pass with the picture graph (similar to how propagate_bounding_rects works).

• Remove PictureOptions struct. Instead, store inflate_if_required with
  the Blur filter enum, which is the only place that uses it.

• Remove root_scaling_factor from text runs - this is handled implicitly
  by the surface device-pixel scale.

• Skip calling update_clip_task for pass-through pictures (since they have
  no defined local rect).

• Improve scaling factors used for determining the render task cache size for
  complex line decorations.

Differential Revision: https://phabricator.services.mozilla.com/D137569

2022-02-09T08:44:08.941000: DEBUG : Did not find a branch, checking all integration branches
2022-02-09T08:44:08.956000: INFO : The bisection is done.
2022-02-09T08:44:08.956000: INFO : Stopped

Comment 1

[Mayank Bansal]

Attachment: about:support

Comment 2

[Mayank Bansal]

Attachment: Reduced testcase.html

Comment 3

[Glenn Watson [:gw]]

Looks like some kind of UV sampling / accuracy issue along tile boundaries when the child surface has a fractional device offset (which is why it appears then disappears as the animation occurs). Will continue looking into this in the morning.

Comment 4

[Glenn Watson [:gw]]

I think the bug is in the usage of calculate_uv_rect_kind for mix-blend pictures - the rects are rounded in device space, not correct accounting for fractional device space rects.

Comment 5

[Glenn Watson [:gw]]

Attachment: Bug 1754350 - Fix mix-blend mode readbacks on fractional device pixel boundaries

Comment 6

[Glenn Watson [:gw]]

Put up a quick patch that I think is probably the correct fix. It fixes the test case for me locally, and all the wrench tests pass. I just kicked off a try run to see if it breaks any of the gecko tests:

https://treeherder.mozilla.org/jobs?repo=try&revision=d48ab849b5b8eb2b7eba94085e66afba4a7648a2

That should have build artifacts shortly for each platform, if you wanted to test locally.

Comment 7

[BugBot [:suhaib / :marco]]

Set release status flags based on info from the regressing bug 1749380

Comment 8

[Mayank Bansal]

(In reply to Glenn Watson [:gw] from comment #6)

Put up a quick patch that I think is probably the correct fix. It fixes the test case for me locally, and all the wrench tests pass. I just kicked off a try run to see if it breaks any of the gecko tests:

https://treeherder.mozilla.org/jobs?repo=try&revision=d48ab849b5b8eb2b7eba94085e66afba4a7648a2

That should have build artifacts shortly for each platform, if you wanted to test locally.

This bug doesnt repro in the test build. I didnt do any other testing

Comment 9

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/c60f4861cfa5
Fix mix-blend mode readbacks on fractional device pixel boundaries r=gfx-reviewers,bradwerth

Comment 10

[Cristian Tuns]

https://hg.mozilla.org/mozilla-central/rev/c60f4861cfa5

Comment 11

[Cristian Tuns]

Backed out as requested by gwatson

• Backout link

Comment 12

[Glenn Watson [:gw]]

See https://bugzilla.mozilla.org/show_bug.cgi?id=1749380#c8 for more info about why this is backed out.

Comment 13

[Glenn Watson [:gw]]

Was opened due to backing out the original patch - however, when I re-land this work next week I'll include this fix. Please re-open if you see it occur again.

Comment 14

[Monica Chiorean]

I tried to reproduce the issue on Win10 /Ubuntu 20.4 on builds without the fix 99.0a1(20220207215603) and on build with the fix 99.0b8, but I see no difference when loading the links attached.
Can you please conform on latest Beta 99.0b8 (https://archive.mozilla.org/pub/firefox/candidates/99.0b8-candidates/) (since you verified it on try build), that the issue is not reproducing on your side? Thank you.

Comment 15

[Mayank Bansal]

I dont use beta, so cant comment on that.
On nightly, this bug has been fixed.

Comment 16

[Monica Chiorean]

Mark issue as verified based on comment 15.