1352180 - Crash in RtlProcessFlsData | LdrShutdownThread

Status: RESOLVED INCOMPLETE

(Core :: XPCOM, defect, P3)

Description

[Andrew McCreight (out of office until 8/21) [:mccr8]]

This bug was filed from the Socorro interface and is 
report bp-6416fbf6-b259-4920-b573-60e302170330.
=============================================================

This is not a very common crash (around 100 across all branches in the last week), but it is much more frequent on Nightly (56%, which is 42 crashes across 11 installations), which is concerning.

I looked at a half dozen of these crashes on Nightly, and in all of them, the main thread was doing zlib stuff, called from mozilla::scache::CacheCloseHelper(). I don't know if that's related or if we just happen to shut this thread down late in shutdown.

Comment 1

[Andrew McCreight (out of office until 8/21) [:mccr8]]

Any ideas, Nathan?

Comment 2

[Nathan Froyd [:froydnj]]

I think that's just an artifact of the once you were looking at?  The ones I looked at on Aurora were all doing perfectly normal things, like waiting to receive messages and the like.

https://crash-stats.mozilla.com/report/index/b94f30be-98ad-479e-bd1c-152542170402
https://crash-stats.mozilla.com/report/index/bde3804d-2849-43a0-a80c-c2bce2170402
https://crash-stats.mozilla.com/report/index/17b5108c-53e0-48c4-9bf1-bd86c2170329
https://crash-stats.mozilla.com/report/index/65d67658-b956-4fe4-ac86-bd5672170329

Plugin shutdown seems to show up rather highly as well.

The crashes in the last week or so are *all* EXCEPTION_ACCESS_VIOLATION_EXEC crashes, which worries me...we're trying to execute memory that doesn't have the correct permissions?  That can't be any good.  Marking this s-s.

Comment 3

[Stephanie Ouillon [:arroway] (needinfo me)]

Nathan, who could be assigned to this bug?

Comment 4

[Nathan Froyd [:froydnj]]

(In reply to Stephanie Ouillon [:arroway] from comment #3)
> Nathan, who could be assigned to this bug?

Not sure.  Benjamin, do we have a resident Windows expert who can look at this?

Comment 5

[Benjamin Smedberg]

302 jimm as platform integration manager. I expect this will not be an urgent priority though, compared to other work.

Comment 6

[Jim Mathies [:jimm]]

Low volume, I don't see a lot of main threads in this cache code, in the signatures I've looked at the main thread is idle.

Comment 7

[Andrew McCreight (out of office until 8/21) [:mccr8]]

People have looked at it, so I think there's no reason to keep this unactionable bug open. It is still present in higher volume on Nightly, so maybe there are some Nightly users who have a troublesome machine.