[OE-core] [v2 4/4] valgrind: disable 256 ptests for aarch64

[Randy MacLeod]

On 9/11/19 10:00 AM, Adrian Bunk wrote:
> On Wed, Sep 11, 2019 at 09:38:53AM -0400, Randy MacLeod wrote:
>> On 9/11/19 8:32 AM, Adrian Bunk wrote:
>>> On Tue, Sep 10, 2019 at 05:41:36PM -0400, Randy MacLeod wrote:
>>>> On qemuarm64 systems, about half of the valgrind tests fail and
>>>> often result in defunct processes, eg:
>>>>      root       819   818 25 00:12 pts/0    00:01:52 /usr/bin/valgrind --command-line-only=yes --memcheck:leak-check=no --tool=memcheck -q --track-origins=
>>>>      root       861   420  0 00:13 pts/0    00:00:00 [sh] <defunct>
>>>> Eventually these processes use so much memory that the
>>>> out of memory killer runs.
>>>>
>>>> Hide the tests that fail as a horrid work-around until the
>>>> root cause has been resolved since this at least allows the
>>>> ptest run to complete.
>>>> ...
>>>
>>> I can reproduce this on hardware,
>>
>> Ah good, I was wondering if it was just happening on qemu.
>>
>>> and I have the bad feeling that no one
>>> will work on resolving the root cause if this is work-arounded in ptest.
>>>
>>> IMHO this should instead be listed as M3 blocker,
>>> similar to "strace ptest timeout".
>>
>> Well, I've enabled 100s of tests that were not being run before
>> and most of them pass for qemux86-64 so I don't think it
>> should be an M3 blocker like strace since it isn't a regression.
> 
> I just tried to reproduce in warrior and it did,
> so it doesn't seem to be a regression.
> 
>> I did start to debug valgrind but it's a bit of a complicated
>> program and will take time so this work-around makes sense to
>> me. If no one else gets to it before me, I'll come back to
>> debugging this issue but likely not before early 3.1.
> 
> It would be good if you could also open a bug, so that people can see
> that there is a problem with this package on an important architecture.

Done:

many valgrind tests fail for arm64
https://bugzilla.yoctoproject.org/show_bug.cgi?id=13520

../Randy

> 
>> Thanks,
>> ../Randy
> 
> Thanks
> Adrian
> 


-- 
# Randy MacLeod
# Wind River Linux