1922707 - Quietly exit content processes on release that can't find an omnijar, on Android

Status: RESOLVED FIXED

(Core :: XPCOM, defect)

Description

[Andrew McCreight [:mccr8]]

In the various forms of bug 1915788, we're seeing a very high volume of crashes in content processes where it can't find the OmniJar. This is the number one crash on release by a large margin. To reduce the noise a bit, we want to try returning to the state before bug 1471720, and killing the content process silently. Nika suggested that we shouldn't go through all sorts of odd untested code when ContentChild fails to initialize (which resulted in the peculiar initial ~ContentChild crash), so instead we can QuickExit. The ~ContentChild crash will remain if we miss some other failure path in the future. Nika also mentioned that a drawback of the current assertion in Omnijar::ChildProcessInit() is that we don't know for sure whether sGREOmni.Get or XRE_GetFileFromPath failed, so I've also split those up.

Comment 1

[Andrew McCreight [:mccr8]]

Attachment: Bug 1922707 - Quietly exit content processes that can't find an omnijar, on Android release.

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/b147c5b60778
Quietly exit content processes that can't find an omnijar, on Android release. r=jstutte

Comment 3

[Cosmin Sabou [:CosminS]]

https://hg.mozilla.org/mozilla-central/rev/b147c5b60778

Comment 4

[BugBot [:suhaib / :marco/ :calixte]]

The patch landed in nightly and beta is affected.
:mccr8, is this bug important enough to require an uplift?

• If yes, please nominate the patch for beta approval.
• If no, please set status-firefox132 to wontfix.

For more information, please visit BugBot documentation.

Comment 5

[Andrew McCreight [:mccr8]]

Attachment: Bug 1922707 - Quietly exit content processes that can't find an omnijar, on Android release.

As seen in bug 1915788, on Android in content processes we are
frequently failing to find the omnijar file. This is the top crash
on Android release, by a large margin.

This patch does two things:

1. On release Android builds, do a QuickExit() in this situation to avoid
   the noise on crash stats. In DIAGNOSTIC builds, we will continue to crash,
   so we can keep an eye on things on Nightly and early beta.

2. This splits the two failure conditions (missing sGREOmni.Get and
   XRE_GetFileFromPath) on Android so we can get a little more information
   about what is going wrong.

Original Revision: https://phabricator.services.mozilla.com/D224562

Comment 6

[Phabricator Automation]

beta Uplift Approval Request

• User impact if declined: no real impact; this should hide the crash signature from bug 1915788 on release Android, which is the #1 crash
• Code covered by automated testing: yes
• Fix verified in Nightly: no
• Needs manual QE test: no
• Steps to reproduce for manual QE testing: NA
• Risk associated with taking this patch: low
• Explanation of risk level: this should only affect behavior in a content process that is already broken
• String changes made/needed: none
• Is Android affected?: yes

Comment 7

[Pulsebot]

https://hg.mozilla.org/releases/mozilla-beta/rev/8efd079afd20