1624250 - Crash in [@ core::option::expect_failed | webrender::scene_building::NodeIdToIndexMapper::get_spatial_node_index]

Status: RESOLVED INACTIVE

(Core :: Graphics: WebRender, defect, P3)

Description

[Gabriele Svelto [:gsvelto]]

This bug is for crash report bp-d55a7d0c-ffa7-4e76-aade-d166f0200322.

Top 10 frames of crashing thread:

0 xul.dll RustMozCrash mozglue/static/rust/wrappers.cpp:16
1 xul.dll mozglue_static::panic_hook mozglue/static/rust/lib.rs:89
2 xul.dll core::ops::function::Fn::call<fn ../f3e1a954d2ead4e2fc197c7da7d71e6c61bad196/src/libcore/ops/function.rs:232
3 xul.dll std::panicking::rust_panic_with_hook ../f3e1a954d2ead4e2fc197c7da7d71e6c61bad196//src/libstd/panicking.rs:475
4 xul.dll std::panicking::begin_panic_handler ../f3e1a954d2ead4e2fc197c7da7d71e6c61bad196//src/libstd/panicking.rs:375
5 xul.dll core::panicking::panic_fmt ../f3e1a954d2ead4e2fc197c7da7d71e6c61bad196//src/libcore/panicking.rs:84
6 xul.dll core::option::expect_failed ../f3e1a954d2ead4e2fc197c7da7d71e6c61bad196//src/libcore/option.rs:1188
7 xul.dll webrender::scene_building::NodeIdToIndexMapper::get_spatial_node_index gfx/wr/webrender/src/scene_building.rs:206
8 xul.dll webrender::scene_building::SceneBuilder::get_clip_and_scroll gfx/wr/webrender/src/scene_building.rs:1026
9 xul.dll webrender::scene_building::SceneBuilder::process_common_properties gfx/wr/webrender/src/scene_building.rs:1043

This is a GPU-process crash happening on all platforms. The crashing line is this:

https://hg.mozilla.org/mozilla-central/annotate/c4d2ca8f78b7680dc0b199a2cb0e2c6f18cd8963/gfx/wr/webrender/src/scene_building.rs#l205

And the reason for the Rust panic is:

no entry found for key

Comment 1

[Darkspirit]

This crash spiked between build 20200319094444 (1 crash/1 install) and 20200319215651 (30 crashes/14 installs):
https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=f985871b3e622041babc71f3770275f045ae7f33&tochange=018964cefa63a14d1e43b31be1b91526a9d708fc
Could have been caused by bug 1616901.

Comment 2

[Rahul Sharma]

my website crashed during the loading- https://www.yolobus.in

Comment 3

[Kris Taeleman (:ktaeleman)]

Crashed while recording a profile of https://www.youtube.com/watch?v=1La4QzGeaaQ.

Profile here in case that can give a clue:
https://bit.ly/2QHGSYU
You can clearly see where it crashes and switched to Advanced Layers.

It seems to have crashed 4 times in a row trying to restore webrender before giving up:
bp-806c7300-6d70-4371-8c49-8fc320200324 3/24/2020, 2:23 PM
View
bp-39f52829-d8a1-44aa-bef8-14cfc0200324 3/24/2020, 2:23 PM
View
bp-68978ac6-e40a-478f-a511-27f1d0200324 3/24/2020, 2:23 PM
View
bp-e5d6c433-5f1a-4985-a5de-481db0200324 3/24/2020, 2:23 PM
View

Comment 4

[Darkspirit]

[Tracking Requested - why for this release]: regression, gpu process topcrash

Comment 5

[BugBot [:suhaib / :marco/ :calixte]]

This bug is tracked by a release manager but with a small severity so change it to major.
For more information, please visit auto_nag documentation.

Comment 6

[Nicolas Silva [:nical]]

The texture cache code modified in bug 1616901 can't affect scene building. It's much more likely that the regression comes from bug 1616412 (unfortunately).

Comment 7

[Jamie Nicol [:jnicol]]

Fairly high number of crashes, so setting as P1

Comment 8

[Miko Mynttinen]

The patch for bug 1624125 might help with this.

Comment 9

[Jessie [:jbonisteel] pls NI]

Assigning to Miko for now, if that patch does not help we will look again

Comment 10

[Miko Mynttinen]

If this crash has the same root cause as bug 1624125, then this should not affect release schedule, as the feature is only enabled for Nightly builds.

Comment 11

[Ryan VanderMeulen [:RyanVM]]

We can mark it as disabled when 76 moves to Beta and the crashes are confirmed to have gone away. In the mean time, this crash still affects 76 where it currently is.

Comment 12

[Jessie [:jbonisteel] pls NI]

Seems like still a few crashes on the 31st - might want to keep an eye on this further.

Comment 13

[Jens Stutte [:jstutte]]

(In reply to Jessie [:jbonisteel] pls NI from comment #12)

Seems like still a few crashes on the 31st - might want to keep an eye on this further.

But should we adjust the priority as the volume has gone down?

Comment 14

[Julien Cristau [:jcristau]]

Volume's gone down, untracking.