1409699 - Crash in mach_msg_trap

Status: RESOLVED FIXED

(Core :: General, defect)

Description

[Calixte Denizet (:calixte)]

This bug was filed from the Socorro interface and is 
report bp-ec3d373c-012c-4386-b903-5b01f0171018.
=============================================================

There are 66 crashes in nightly 58 (vs 20 in 57.0a1 and 19 in 56.0a1).
In 58, 95% of the crashes are with OS X 10.12 and 5% with OS X 10.11.

:spohl, could you investigate please ?

Comment 1

[Stephen A Pohl [:spohl]]

Virtually all of the crash comments are from me when I intentionally triggered SIGABRT crashes while working on bug 1405151. There is no reason to believe that this is a legitimate issue at this time.

Comment 2

[[:philipp]]

there are crash reports from two dozen different installations on 57.0b9 and 58.0a1 as well.
this signature is also the top content process crash for mac users in 57.0b9 now.

Comment 3

[Stephen A Pohl [:spohl]]

Oh, interesting. I just looked at this again and it looks like the crashes that were previously landing in the "Crash in EMPTY: no crashing thread identified; OK" bucket (bug 1405151) are now showing up here with mach_msg_trap as top frame. However, there are at least a couple of different crashes here. The stacks show that there is at least one crash related to cubeb[1]. Another one has to do with shared surfaces[2].

Is there a way to split these out and be more granular based on the crashing thread stack, rather than having just one bug (this one) for all the various crashes?

[1] https://crash-stats.mozilla.com/report/index/a2e5766e-5d9c-4972-967b-13f1e0171019
[2] https://crash-stats.mozilla.com/report/index/2e177d22-8236-40dc-88f7-2e5b10171018

Comment 4

[Calixte Denizet (:calixte)]

Yes it's possible to use "proto signature" field:
 - crashes with SharedSurface in the stack: [1]
 - crashes without SharedSurface and facet on "proto signature" to see other potential causes: [2]

[1] https://crash-stats.mozilla.com/search/?signature=%3Dmach_msg_trap&proto_signature=~SharedSurface&date=%3E%3D2017-10-12T14%3A26%3A00.000Z&date=%3C2017-10-19T14%3A26%3A00.000Z&_sort=-date&_facets=signature&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#crash-reports

[2] https://crash-stats.mozilla.com/search/?signature=%3Dmach_msg_trap&proto_signature=%21~SharedSurface&date=%3E%3D2017-10-12T14%3A26%3A00.000Z&date=%3C2017-10-19T14%3A26%3A00.000Z&_sort=-date&_facets=signature&_facets=proto_signature&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#facet-proto_signature

Comment 5

[Stephen A Pohl [:spohl]]

ESR52 becomes affected if we uplift bug 1405151. That would be a good thing, as we would get better crash reports from ESR52 as well.

Comment 6

[Stephen A Pohl [:spohl]]

This was resolved by improving the way we filter these incoming crash reports (bug 1410580).