1012691 - crash in NS_ABORT_OOM(unsigned int) | PL_DHashTableInit(PLDHashTable*, PLDHashTableOps const*, void*, unsigned int, unsigned int) | nsCycleCollector::BeginCollection(ccType, nsICycleCollectorListener*)

Status: RESOLVED INCOMPLETE

(Core :: XPCOM, defect)

Description

[Tracy Walker [:tracy]]

This bug was filed from the Socorro interface and is 
report bp-d60b8812-6da3-414f-986a-1a7f22140512.
=============================================================

Filling this a follow up to bug 1006181 to address this crash on Fx30 and Fx31

OOM Allocation Size: 262144

Comment 1

[Benjamin Smedberg]

I don't understand what this bug is tracking. The big increase in this signature was because of the memory leak/usage from HTTP cache v2, which isn't on fx30/31. Otherwise this is just an OOM crash, which we don't typically file or track.

Comment 2

[Robert Kaiser]

(In reply to Benjamin Smedberg  [:bsmedberg] from comment #1)
> Otherwise this is just an OOM crash, which we don't typically file or track.

We want a bug on file about any signature in the higher ranks on any release. This is in the top 20 on Fx30, so that warrants having a bug to track it. If it's a dummy because this is generic OOM, so be it.

That said, I wonder if it's good that we require to allocate new memory to run CC, which should hopefully free up some memory.

Comment 3

[Robert Kaiser]

Oh, and comment #2 answers the ni? - even more as I asked tracy to file a new bug instead of sticking Fx30 stuff into bug 1006181 which is about cache2, which does not apply to 30 (or 31).

Comment 4

[Benjamin Smedberg]

Is the total volume of OOM crashes worse in FF30 than in previous releases? If so, please file a bug on that specifically.

This bug is not actionable. I think I disagree with your policy of filing a bug for each OOM bug separately, instead of adding them to some meta-bug, but if you do, please mark it INCOMPLETE as you file it.

Comment 5

[Robert Kaiser]

(In reply to Benjamin Smedberg  [:bsmedberg] from comment #4)
> This bug is not actionable. I think I disagree with your policy of filing a
> bug for each OOM bug separately, instead of adding them to some meta-bug

I didn't say we need to have each one filed separately. We just want to have the signature be connected to some bug that actually has to do with what this is. I do not think we should resolve bugs that are valid crashes as INCOMPLETE even if they might not be actionable.

Comment 6

[Sylvestre Ledru [:Sylvestre]]

INCOMPLETE, so, not tracking.

Comment 7

[u279076]

This signature has disappeared completely based on https://crash-stats.mozilla.com/topcrasher_ranks_bybug/?bug_number=1012691

Comment 8

[Andrew McCreight [:mccr8]]

Yeah, this will have morphed into one of the generic OOM signatures.