836576 - App perf script intermittently dies with IPC errors

Status: RESOLVED WORKSFORME

(Firefox OS Graveyard :: General, defect)

Description

[Jonathan Griffin (:jgriffin)]

Our app startup performance script launches specific apps 30 times in a row to measure the startup time each time.

Occasionally this dies at a random point.  In the logcat I see:

I/Gecko   (  109): [Parent 109] WARNING: pipe error (151): Connection reset by peer: file /data/jenkins/jobs/build-unagi-eng/workspace/gecko/ipc/chromium/src/chrome/common/ipc_channel_posix.cc, line 432
I/Gecko   (  109): [Parent 109] WARNING: pipe error (153): Connection reset by peer: file /data/jenkins/jobs/build-unagi-eng/workspace/gecko/ipc/chromium/src/chrome/common/ipc_channel_posix.cc, line 432
I/Gecko   (  109): 1359591137229	Marionette	INFO	sendToClient: {"from":"conn2.marionette1","error":{"message":"timed out","status":28,"stacktrace":null}}, undefined, {379192c9-3c8f-4c5e-a2e4-c81d27c9e354}
I/Gecko   (  109): 1359591137246	Marionette	WARN	got a response with no command_id
I/Gecko   (  109): 
I/Gecko   (  109): ###!!! [Parent][RPCChannel] Error: Channel error: cannot send/recv
I/Gecko   (  109): 
I/Gecko   (  109): 1359591137693	Marionette	INFO	sendToClient: {"from":"conn2.marionette1","ok":true}, {dd20a5a5-3161-4484-bcb0-5533a46b721c}, {dd20a5a5-3161-4484-bcb0-5533a46b721c}
E/GeckoConsole(  109): [JavaScript Error: "[Exception... "Component returned failure code: 0x80470002 (NS_BASE_STREAM_CLOSED) [nsIOutputStream.write]"  nsresult: "0x80470002 (NS_BASE_STREAM_CLOSED)"  location: "JS frame :: chrome://global/content/devtools/dbg-transport.js :: DT_onOutputStreamReady :: line 94"  data: no]" {file: "chrome://global/content/devtools/dbg-transport.js" line: 94}]
I/Gecko   ( 4160): ###################################### forms.js loaded
I/Gecko   (  109): 
I/Gecko   (  109): ###!!! [Parent][AsyncChannel] Error: Channel error: cannot send/recv
I/Gecko   (  109): 
I/Gecko   (  109): 
I/Gecko   (  109): ###!!! [Parent][AsyncChannel] Error: Channel error: cannot send/recv

I'm not positive this is the cause of the problem, will have to wait for a few more bad logs.

Comment 1

[Jonathan Griffin (:jgriffin)]

After this occurs, an entry for this app is stuck in the app list even though the app isn't running.  This breaks the app launching code, so you need to reboot the phone before trying another test.

Comment 2

[Jonathan Griffin (:jgriffin)]

This may be resolved by increasing the delay between killing an app and launching one from 1 to 5s, per a note from Vivien.

Comment 3

[Jonathan Griffin (:jgriffin)]

We're this seeing quite frequently with the camera app.  In the logcat, after this happens, I see:

I/Gecko   ( 3740): 
I/Gecko   ( 3740): ###!!! [Child][SyncChannel] Error: Channel error: cannot send/recv
I/Gecko   ( 3740): 
E/AudioHardwareMSM76XXA( 3768): failed to open AUTO_VOLUME_CONTROL /system/etc/AutoVolumeControl.txt: No such file or directory (2)
E/QualcommCamera( 3768): Qint android::get_number_of_cameras(): E
I/QualcommCameraHardware( 3768): getCameraInfo: IN
I/QualcommCameraHardware( 3768): getCameraInfo: loading libqcamera at 0xb000ee28
V/QualcommCameraHardware( 3768):  Storing the current target type as 3 
I/QualcommCameraHardware( 3768): getCameraInfo: numOfCameras = 1
I/QualcommCameraHardware( 3768): Camera sensor 0 info:
I/QualcommCameraHardware( 3768): camera_id: 0
I/QualcommCameraHardware( 3768): modes_supported: 5
I/QualcommCameraHardware( 3768): position: 0
I/QualcommCameraHardware( 3768): sensor_mount_angle: 90
V/QualcommCameraHardware( 3768): getCameraInfo: dlclose(libqcamera)
I/QualcommCameraHardware( 3768): getCameraInfo: OUT
W/AudioFlinger( 3768): Thread AudioOut_1 cannot connect to the power manager service

This is the same output we see when B2G first boots, so I'm guessing that the phone is rebooting here.

I've removed Camera from the list of apps we're running and will see if we can get some data without it.

Comment 4

[Jonathan Griffin (:jgriffin)]

I can reproduce the camera problem locally, it is a crash.  I'll file a separate bug for this.

Comment 5

[Ben Kelly [:bkelly, not reviewing]]

Jonathan, do you know if this is still a real bug or if it has been resolved?

Comment 6

[Jonathan Griffin (:jgriffin)]

This has been working well for a while now.