1577730 - Mozilla Firefox crashes in [@ js::AtomsTable::tracePinnedAtoms ]

Status: VERIFIED DUPLICATE of bug 1555936

(Core :: JavaScript: GC, defect, P2)

Description

[Virtual_ManPL [:Virtual] 🇵🇱 - (please needinfo? me - so I will see your comment/reply/question/etc.)]

STR:

• no real STR, I was just browsing some random website pages and crash in [@ js::AtomsTable::tracePinnedAtoms ] happened

Crashlog report:
https://crash-stats.mozilla.org/report/index/71a3140f-4964-4723-a2a6-0c2d00190830

Comment 1

[Jon Coppeard (:jonco)]

Null pointer in atoms table, that shouldn't be possible.

Comment 2

[Nicolas B. Pierron [:nbp]]

(In reply to Jon Coppeard (:jonco) from comment #1)

Null pointer in atoms table, that shouldn't be possible.

Paul, do you know if we have assertion to prevent such things to be inserted in the Atoms table?
Do we have an easy way to make such impossible things crash sooner than the GC?

Comment 3

[Virtual_ManPL [:Virtual] 🇵🇱 - (please needinfo? me - so I will see your comment/reply/question/etc.)]

I want to add that I had lately for few days some minor hardware instability which I fixed now, so in end it doesn't have to be Mozilla Firefox issue.
I can close this bug as WORKSFORME.

Comment 4

[Paul Bone [:pbone]]

nbp: I can't see how a null atom could be added, there are assertions in place.

There are 285 crashes here in the last week, so I'm on the fence about closing it as Virtual_ManPL has suggested, it may be their hardware but might not be. I'm not sure it's actionable. Let's leave it open.

Jonco: is this a duplicate of Bug 1555936? (Could you CC me on that bug, I can't see it.)

(Marking as secure because 1555936 is secure.)

Comment 5

[Jon Coppeard (:jonco)]

This could very well be caused by hardware instability and it's probably not worth investigating more in this bug unless it happens again on a known-good system.

The crash is the same as the one in bug 1555936 yes, caused by null entries showing up in the atoms table (I've CCed you).

Comment 6

[Nicolas B. Pierron [:nbp]]

(In reply to Virtual_ManPL [:Virtual] 🇵🇱 - (please needinfo? me - so I will see your comment/reply/question/etc.) from comment #3)

I want to add that I had lately for few days some minor hardware instability which I fixed now, so in end it doesn't have to be Mozilla Firefox issue.
I can close this bug as WORKSFORME.

Thank you very much for this feedback!
This is actually something we have really hard time figuring out from crash reports, which is the list of bugs which are likely to be caused by faulty hardware.

Comment 7

[Paul Bone [:pbone]]

Okay, I'll make this as a duplicate. Jon commented (bug 1555936 comment 13) that this could be a general heap corruption [maybe bad hardware] and the recent increase in the number of crashes here is due to another change.

Thanks.