Closed Bug 1416602 Opened 7 years ago Closed 7 years ago

Crash in mozalloc_abort | abort | webrender::resource_cache::ResourceCache::update_resources

Categories

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

x86_64
Linux
defect

Tracking

()

RESOLVED WORKSFORME
Tracking Status
firefox-esr52 --- unaffected
firefox56 --- unaffected
firefox57 --- unaffected
firefox58 --- unaffected

People

(Reporter: jan, Unassigned)

References

(Blocks 1 open bug)

Details

(Keywords: crash, nightly-community, Whiteboard: [wr-reserve] [gfx-noted])

Crash Data

Seen on Socorro. The one and only report so far: bp-a49daf64-91a7-4a3c-a5b0-e0fc90171112 (20171112100435) > MOZ_CRASH Reason: called `Option::unwrap()` on a `None` value (Maybe a regression?)
Whiteboard: [wr-mvp] [triage] → [wr-mvp]
Whiteboard: [wr-mvp] → [wr-mvp] [gfx-noted]
Priority: P2 → P3
Whiteboard: [wr-mvp] [gfx-noted] → [wr-reserve] [gfx-noted]
I just hit this locally: https://crash-stats.mozilla.com/report/index/f8bf3530-aed3-4d07-b3e0-fcde20171206 I had closed a few tabs, realized I needed one of them, so I hit Shift+Ctrl+T several times in quick succession to unclose them and Firefox crashed. Unfortunately the stack isn't super helpful given the Rust compiler's heavy use of inlining + weirdness of debug info for macros, since the frame that panic'ed is just: 8 libxul.so webrender::resource_cache::ResourceCache::update_resources src/libcore/macros.rs:20 (that source line is literally the implementation of `panic!` in the standard library.) It's clearly in here somewhere: https://hg.mozilla.org/mozilla-central/file/79d3e25106f8/gfx/webrender/src/resource_cache.rs#l274 ...but I can't tell exactly where. That file does contain a number of calls to `unwrap()`, which doesn't seem great.
I got objdump to disassemble that whole function but it doesn't help much: https://gist.github.com/luser/3992395fb238238dc4d5fb29711a0d06 The call to panic is here: _ZN4core6option8{{impl}}46unwrap<&webrender::platform::unix::font::Face>E(): /build/mozilla-central/macros.rs:20 3c09aa6: 48 8d 3d c3 a3 b5 01 lea 0x1b5a3c3(%rip),%rdi # 5763e70 <ref.4K.2926> 3c09aad: e8 1e c6 3a 00 callq 3fb60d0 <core::panicking::panic> So I guess it's something dealing with fonts, at least?
This combination of crash signature and crash reason hasn't happened anymore.
Status: NEW → RESOLVED
Closed: 7 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.