1425471 - Marionette tests fail on Android 7: /system/bin/sh: /system/bin/netcfg: not found

Status: RESOLVED FIXED

(Firefox for Android Graveyard :: Testing, defect, P1)

Description

[Geoff Brown [:gbrown]]

It looks like netcfg was replaced with ifconfig in Android 6: 

https://github.com/bbc/device_api-android/issues/92
https://stackoverflow.com/questions/36294085/how-to-debug-android-6-0-over-wi-fi-netcfg-issue


Only Marionette seems to rely on this:

https://treeherder.mozilla.org/logviewer.html#?job_id=151817064&repo=try

[task 2017-12-15T16:15:20.832Z] 16:15:20     INFO -  /system/bin/sh: /system/bin/netcfg: not found
[task 2017-12-15T16:15:21.975Z] 16:15:21     INFO -  /system/bin/sh: /system/bin/netcfg: not found
[task 2017-12-15T16:15:23.111Z] 16:15:23     INFO -  /system/bin/sh: /system/bin/netcfg: not found
[task 2017-12-15T16:15:24.241Z] 16:15:24     INFO -  /system/bin/sh: /system/bin/netcfg: not found
[task 2017-12-15T16:15:25.257Z] 16:15:25     INFO -  {"thread": "MainThread", "level": "ERROR", "pid": 1103, "source": "Marionette test runner", "time": 1513354525256, "action": "log", "message": "Failure during harness execution", "stack": "Traceback (most recent call last):\n\n  File \"/builds/worker/workspace/build/tests/marionette/harness/marionette_harness/runtests.py\", line 92, in cli\n    failed = harness_instance.run()\n\n  File \"/builds/worker/workspace/build/tests/marionette/harness/marionette_harness/runtests.py\", line 72, in run\n    runner.run_tests(tests)\n\n  File \"/builds/worker/workspace/build/venv/lib/python2.7/site-packages/marionette_harness/runner/base.py\", line 852, in run_tests\n    self.marionette = self.driverclass(**self._build_kwargs())\n\n  File \"/builds/worker/workspace/build/venv/lib/python2.7/site-packages/marionette_driver/marionette.py\", line 621, in __init__\n    self.start_binary(self.startup_timeout)\n\n  File \"/builds/worker/workspace/build/venv/lib/python2.7/site-packages/marionette_driver/marionette.py\", line 632, in start_binary\n    self.instance.start()\n\n  File \"/builds/worker/workspace/build/venv/lib/python2.7/site-packages/marionette_driver/geckoinstance.py\", line 351, in start\n    self.runner.start()\n\n  File \"/builds/worker/workspace/build/venv/lib/python2.7/site-packages/mozrunner/base/device.py\", line 80, in start\n    raise Exception(\"Network did not come up when starting device\")\n\nException: Error possibly due to runner or device args: Network did not come up when starting device\n"}


I wonder if this is even necessary now.

Comment 1

[Geoff Brown [:gbrown]]

$ ./mach android-emulator --version 4.3
...
$ adb shell /system/bin/netcfg
sit0     DOWN                                   0.0.0.0/0   0x00000080 00:00:00:00:00:00
eth0     UP                                   10.0.2.15/24  0x00001043 52:54:00:12:34:56
lo       UP                                   127.0.0.1/8   0x00000049 00:00:00:00:00:00



$ ./mach android-emulator --version 7.0
...
$ adb shell /system/bin/netcfg
/system/bin/sh: /system/bin/netcfg: not found
$ adb shell /system/bin/ifconfig
eth0      Link encap:UNSPEC  
          inet addr:10.0.2.15  Bcast:10.0.2.255  Mask:255.255.255.0 
          inet6 addr: fe80::5054:ff:fe12:3456/64 Scope: Link
          inet6 addr: fec0::5054:ff:fe12:3456/64 Scope: Site
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:32 errors:0 dropped:0 overruns:0 frame:0 
          TX packets:31 errors:0 dropped:0 overruns:0 carrier:0 
          collisions:0 txqueuelen:1000 
          RX bytes:7522 TX bytes:2808 

lo        Link encap:UNSPEC  
          inet addr:127.0.0.1  Mask:255.0.0.0 
          inet6 addr: ::1/128 Scope: Host
          UP LOOPBACK RUNNING  MTU:65536  Metric:1
          RX packets:4 errors:0 dropped:0 overruns:0 frame:0 
          TX packets:4 errors:0 dropped:0 overruns:0 carrier:0 
          collisions:0 txqueuelen:0 
          RX bytes:248 TX bytes:248

Comment 2

[Geoff Brown [:gbrown]]

Also seen on packet.net:

https://treeherder.mozilla.org/logviewer.html#?job_id=184791257&repo=try&lineNumber=1340

Comment 3

[Geoff Brown [:gbrown]]

Attachment: use ADBDevice's get_ip_address() instead

I wonder if wait_for_net() is really needed at all, but I assume it was added for a good reason.

Here I substitute "can we determine the device's ip address?" for "is the network up?"; I think that's reasonable.


This allows Mn to run on Android 7 and doesn't break the Android 4.3 tests:

https://treeherder.mozilla.org/#/jobs?repo=try&revision=6b9aa141779a0ce82dca91ee863fb8217eb30348

Comment 4

[Bob Clary [:bc] (inactive)]

Comment on attachment 8987933 [details] [diff] [review]
use ADBDevice's get_ip_address() instead

Review of attachment 8987933 [details] [diff] [review]:
-----------------------------------------------------------------

r+ lgtm.

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/b077d4c0a0d8
Avoid use of netcfg in Android marionette tests; r=bc

Comment 6

[Dorel Luca [:dluca]]

https://hg.mozilla.org/mozilla-central/rev/b077d4c0a0d8