1582029 - Crash in [@ webrender::batch::BatchBuilder::add_prim_to_batch]

Status: RESOLVED WORKSFORME

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

Description

[Pascal Chevrel:pascalc]

This bug is for crash report bp-1ec52bd2-1e25-4145-a327-429e60190917.

Top 10 frames of crashing thread:

0 xul.dll GeckoCrash toolkit/xre/nsAppRunner.cpp:5116
1 xul.dll static void gkrust_shared::panic_hook toolkit/library/rust/shared/lib.rs:250
2 xul.dll static void core::ops::function::Fn::call<fn src/libcore/ops/function.rs:69
3 xul.dll static void std::panicking::rust_panic_with_hook src/libstd/panicking.rs:481
4 xul.dll static void std::panicking::continue_panic_fmt src/libstd/panicking.rs:384
5 xul.dll static void std::panicking::rust_begin_panic src/libstd/panicking.rs:311
6 xul.dll static void core::panicking::panic_fmt src/libcore/panicking.rs:85
7 xul.dll void core::panicking::panic src/libcore/panicking.rs:49
8 xul.dll static void webrender::batch::BatchBuilder::add_prim_to_batch gfx/wr/webrender/src/batch.rs
9 xul.dll static void webrender::frame_builder::build_render_pass gfx/wr/webrender/src/frame_builder.rs:910

Comment 1

[Jamie Nicol [:jnicol]]

This is the pushlog for the build with the first crashes: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=286cbc3b542b&tochange=46ca31fd6a722f37c0b097301396d3a2c710b086

Out of that, bug 1580711 looks suspicious. Though the numbers are low enough that it's possible an early build was affected..

Comment 2

[Glenn Watson [:gw]]

It seems possibly related to that patch, yea. We'll need a repro to solve it though, do we have any reliable URLs and/or repro steps yet?

Comment 3

[Rui Matos]

This happened to me while clicking around in https://drive.google.com/drive/my-drive . Crash report is https://crash-stats.mozilla.org/report/index/16b88d8e-139e-4262-83a6-4f7d10191108#tab-details .

Comment 4

[Darkspirit]

(That's Beta crashing on Google Drive.)

Comment 5

[Nicolas Silva [:nical]]

Attachment: Bug 1582029 - Don't unwrap() clip task addresses during batching. r=gw

It looks like we don't expect to get None there so I left a debug assertion. It's not worth crashing release builds here. If skipping the primitive leads to incorrect rendering we'll hit the assertion when reproducing with a debug build.

Comment 6

[Glenn Watson [:gw]]

Do we have any URLs / repro steps for this, but only for cases that still occur after the main cause of this bug was fixed? (see how the crash data drops dramatically after 30th Oct, which I believe was helped by https://bugzilla.mozilla.org/show_bug.cgi?id=1592121).

Comment 7

[Glenn Watson [:gw]]

(the above comment is in relation to https://bugzilla.mozilla.org/show_bug.cgi?id=1604492#c2)

Comment 8

[Nicolas Silva [:nical]]

No crash in the last month, closing.