Closed Bug 1038136 Opened 10 years ago Closed 10 years ago

Gaia perf test cause crash with settings.

Categories

(Firefox OS Graveyard :: Gaia::Settings, defect, P1)

ARM
Gonk (Firefox OS)
defect

Tracking

(Not tracked)

RESOLVED DUPLICATE of bug 1042648
2.1 S1 (1aug)

People

(Reporter: hub, Unassigned)

Details

(Keywords: perf, Whiteboard: [c=automation p= s=2014.08.01 u=])

Attachments

(3 files)

|APP=settings make test-perf|

The test will fail on "overfill" as b2g crashes. (you see it reboot, the marionetter connection gets reset)
Mason, would you have an idea? Since the "overfill" involve code in Gecko.

This is a regression.
Flags: needinfo?(mchang)
This block bug 981009 as I can't do it for the overfill test.
Blocks: 981009
Keywords: perf
Priority: -- → P1
Whiteboard: [c=automation p= s= u=]
I tried running locally and this is the log I got. I see that the overfill tests are working locally. This is based on the master of gecko / gaia today:

Gecko: 193791:340b19c14d3d
Gaia: 075f32fc0b411eadbd91bbc65e06dc5983a09ac9
Flags: needinfo?(mchang)
Here is a log of APP=settings instead of App=Settings, which was the first log. I see it crashing at a different place.
Attachment #8455418 - Attachment description: Log of make test-perf → Log of APP=Settings make test-perf
Attachment #8455421 - Attachment description: Log of make test-perf with lower case settings → Log of APP=settings make test-perf (lower case settings)
It definitely show a crash. But this is starting up the settings app. Gotta need to dig more.
Is there a crash report or crash dump that we can take a look at?
No because it is a self build.

And I haven't even been able to run it properly under gdb. Don't know why.
Flags: needinfo?(nhirata.bugzilla)
Which device is this on? Maybe we can try to repro on an releng engineering build.
It is a flame.
(In reply to Mason Chang [:mchang] from comment #4)

> Here is a log of APP=settings instead of App=Settings, which was the first
> log. I see it crashing at a different place.


"App=Settings" should have never worked anyway....

Still b2g crashes.
Attached file logcat
Oh wierd.  it gets stuck on the launching of settings for me.

I didn't see any crash reports in the /data/b2g/mozilla/Crash Reports directory.

root@flame:/data/b2g # b2g-info
                          |      megabytes      |
           NAME  PID PPID CPU(s) NICE   USS   PSS   RSS VSIZE OOM_ADJ USER    
            b2g  306    1  212.5    0 112.9 119.1 136.8 273.2       0 root    
         (Nuwa)  909  306    3.4    0   1.5   4.9  19.8  54.1       0 root    
OperatorVariant  970  909    0.9   18   5.0   7.3  19.4  61.1      11 u0_a970 
Built-in Keyboa 1102  306    1.7    1  11.0  14.2  28.8  69.7       3 u0_a1102
Smart Collectio 1172  909    1.0   18   5.0   7.4  19.7  61.1      11 u0_a1172
    Home screen 1173  306   25.6   18  23.1  28.6  45.5 112.5       8 u0_a1173
(Preallocated a 3047  909    0.2    1   2.4   4.3  13.0  58.2       2 u0_a3047
(Preallocated a 3144  909    0.6   18   5.3   7.5  18.9  59.1       1 u0_a3144

System memory info:

            Total 404.6 MB
     Used - cache 249.6 MB
  B2G procs (PSS) 193.3 MB
    Non-B2G procs  56.3 MB
     Free + cache 155.0 MB
             Free  42.1 MB
            Cache 112.9 MB

Low-memory killer parameters:

  notify_trigger 14336 KB

  oom_adj min_free
        0  4096 KB
       58  5120 KB
      117  6144 KB
      352  7168 KB
      470  8192 KB
      588 20480 KB
Flags: needinfo?(nhirata.bugzilla)
If I disable the overfill test, it doesn't crash.

If I use bug 1027232 to restart b2g each test, it does not crash either.
Apparently solving bug 1042648 fixes this....
No longer blocks: 981009
Tempted to close this as dupe for bug 1042648.

Mason, what do you think?
Flags: needinfo?(mchang)
Sounds good as long as settings isn't crashing.
Flags: needinfo?(mchang)
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → DUPLICATE
Whiteboard: [c=automation p= s= u=] → [c=automation p= s=2014.08.01 u=]
Target Milestone: --- → 2.1 S1 (1aug)
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: