Open Bug 1803236 (layout-crash-memory-corruption) Opened 3 years ago Updated 2 years ago

[meta] crashes in layout code that seem likely to really be memory corruption / bad RAM

Categories

(Core :: Layout, task)

task

Tracking

()

People

(Reporter: dholbert, Unassigned)

References

(Depends on 3 open bugs)

Details

(Keywords: meta)

Filing this metabug to aggregate layout crashes / crash signatures that we suspect to be partially or entirely explained by memory corruption.

Keywords: meta
Depends on: 707699

In particular, I'm intending for this bug to covers crashes where we seem to be doing a (largely) read-only walk over the frame tree, and we crash when we traverse some bad data. So: the actual "bad stuff" happened earlier, and we're just now tripping over it when we traverse the frame tree.

("bad stuff" includes bad RAM experiencing a random bit-flip; we have evidence of this being the case in e.g. bug 1300342 comment 5.)

Summary: [meta] crashes in layout code that seem likely to really be memory corruption → [meta] crashes in layout code that seem likely to really be memory corruption / bad RAM

Bug 1738651 might be of interest to this effort. Ideally we'd like to be able to identify the majority of these crashes and flag them as suspicious on crash-stats.

Duplicate of this bug: 1821482
Duplicate of this bug: 1826240
No longer duplicate of this bug: 1826240
No longer duplicate of this bug: 1821482

(disregard the intermittent failures robot comments here; those were because I mixed up bug 1440376 and this bug when marking a duplicate.)

You need to log in before you can comment on or make changes to this bug.