1532756 - Unable to load OpenH264 plugin on Android

Status: RESOLVED FIXED

(Core :: Audio/Video: GMP, defect)

Description

[Dan Minor [:dminor]]

I see the following error in logcat:

03-05 14:51:59.416 794 3845 W ActivityManager: Unable to start service Intent { cmp=org.mozilla.fennec_aurora/org.mozilla.gecko.process.GeckoServiceChildProcess$gmplugin } U=0: not found

Mozregression points to:

https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=51c86d025ecb710b8d37d3454a1d0adcfff8f89d&tochange=de7c04ff60015787b787416797252148902e4bfb

de7c04ff6001 Nathan Froyd — Bug 1514044 - part 3 - update nsICrashService process constants and add checks for them; r=gsvelto
65f142863701 Nathan Froyd — Bug 1514044 - part 2 - change CrashReporterHost to use GeckoProcessTypes.h; r=gsvelto
1abd03bcbfa3 Nathan Froyd — Bug 1514044 - part 1 - macro-ify process enums, strings, and XRE functions; r=mccr8
6d7562d60590 Nathan Froyd — Bug 1514043 - change gecko media plugin process name; r=mccr8,bobowen,haik,chutten

I tested this using https://mozilla.github.io/webrtc-landing/pc_test.html, select "Require H.264 video" and de-select "Enable Identity Provider". On affected versions of Firefox, you will only see the local video in each video element. On working versions of Firefox, you will see both local and remote videos in each video element.

Due to Bug 1532578, I was not able to test versions of Firefox after Bug 1527811 landed, so I guess it's possible this has been fixed since then.

Comment 1

[Dan Minor [:dminor]]

[Tracking Requested - why for this release]:

This is a recent regression that would prevent us from using OpenH264 on Android.

Comment 2

[Nathan Froyd [:froydnj]]

Ugh. Do we have to fix 66, given that bug 1532578 isn't going to be fixed for 66?

snorp, do I just have to update https://searchfox.org/mozilla-central/source/mobile/android/geckoview/src/main/java/org/mozilla/gecko/process/GeckoServiceChildProcess.java#144 for the new name, and then magic will just make things work?

Comment 3

[Dan Minor [:dminor]]

Bug 1532578 only affects Firefox 67, so it would be good to have a fix for 66.

Comment 4

[Dan Minor [:dminor]]

It is also necessary to update a few manifest files.

Comment 5

[Dan Minor [:dminor]]

Attachment: Bug 1532756 - Rename geckomediaplugin to gmplugin on Android; r=droeh!

In Bug 1514043 we renamed the geckomediaplugin to gmplugin, but missed renaming
it on Android. This prevents the plugin from loading, which breaks OpenH264 in
WebRTC.

Comment 6

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/e038e06397ad
Rename geckomediaplugin to gmplugin on Android; r=droeh

Comment 7

[Arthur Iakab [arthur_iakab]]

https://hg.mozilla.org/mozilla-central/rev/e038e06397ad

Comment 8

[Dan Minor [:dminor]]

Comment on attachment 9049180 [details]
Bug 1532756 - Rename geckomediaplugin to gmplugin on Android; r=droeh!

Beta/Release Uplift Approval Request

• Feature/Bug causing the regression: Bug 1514043
• User impact if declined: Unable to use openh264 on Android, which would break webrtc services which only support H.264.
• Is this code covered by automated tests?: No
• Has the fix been verified in Nightly?: No
• Needs manual test from QE?: No
• If yes, steps to reproduce:
• List of other uplifts needed: None
• Risk to taking this patch: Low
• Why is the change risky/not risky? (and alternatives if risky): Low risk, only code changes are to rename the plugin.
• String changes made/needed: None.

Comment 9

[Liz Henry (:lizzard) (relman/hg->git project)]

Comment on attachment 9049180 [details]
Bug 1532756 - Rename geckomediaplugin to gmplugin on Android; r=droeh!

We probably shouldn't break WebRTC on mobile.
OK to uplift for the RC build.

Comment 10

[Liz Henry (:lizzard) (relman/hg->git project)]

Is there anything we can add to automated tests to catch this in future?

Comment 11

[Bogdan Tara[:bogdan_tara | bogdant]]

https://hg.mozilla.org/releases/mozilla-beta/rev/77dcf9bd89d0

Comment 12

[Dan Minor [:dminor]]

(In reply to Liz Henry (:lizzard) (use needinfo) from comment #10)

Is there anything we can add to automated tests to catch this in future?

At the moment we don't run any openh264 tests on Android at all. We have Bug 1355786 on file for that. We also have Bug 1462786 to use the real OpenH264 plugin which might help things, as I don't think we're building the "fake" h264 plugin for Android.