On 11.03.21 10:12, Jan Kiszka wrote:
> On 08.03.21 13:46, Jan Kiszka via Xenomai wrote:
>> On 08.03.21 13:42, Bezdeka, Florian (T RDA IOT SES-DE) wrote:
>>> On Mon, 2021-03-08 at 11:42 +0100, Jan Kiszka wrote:
>>>> On 04.03.21 19:33, florian.bezdeka--- via Xenomai wrote:
>>>>> On Thu, 2021-03-04 at 14:18 +0000, florian.bezdeka--- via Xenomai
>>>>> wrote:
>>>>>> Hi,
>>>>>>
>>>>>> the attached kernel log was taken form a local qemu machine with
>>>>>> Xenomai build from the wip/dovetail branch. There is a failing selftest
>>>>>> (mixed read-lock/lock-write ABBA).
>>>>>>
>>>>>
>>>>> Correction: I can reproduce that with `next` as well. So it's not
>>>>> related to the wip/dovetail branch.
>>>>>
>>>>
>>>> Situation still the same with the latest release?
>>>
>>> Which release? I tried to build the "testing" branch of xenomai-images
>>> (which holds the COW patches) but that doesn't build at all, patches
>>> can not be applied.
>>>
>>
>> Yeah, that's because they are applied to latest x86 I-pipe releases. You
>> don't need that branch for that arch anymore.
>>
>>> used cmdline:
>>> kas-container build
>>> kas.yml:opt-linux-latest-5.4.yml:board-qemu-amd64.yml:opt-xenomai-next.yml
>>>
>>> Building and running the "next" branch of xenomai-images still reports
>>> the same failing self-test.
>>
>> Hmm, too bad...
>>
>> Jan
>>
>
> FWIW, I'm still not getting this one, but I face random SEGV of
> applications while latency is running. I'm not seeing this with latest
> 4.19 so far. Seems like 5.4-x86 is not yet stable...
$ cd xenomai-images
$ git fetch
$ git checkout origin/next
$ kas-container build
kas.yml:opt-linux-latest-5.4.yml:opt-xenomai-next.yml:board-qemu-amd64.yml
$ ./start-qemu amd64
$ login
$ dmesg
[ 0.270682] ------------------------
[ 0.271303] | Locking API testsuite:
[ 0.271953]
----------------------------------------------------------------------------
[ 0.273260] | spin |wlock |rlock
|mutex | wsem | rsem |
[ 0.274480]
--------------------------------------------------------------------------
[ 0.276748] A-A deadlock: ok | ok | ok |
ok | ok | ok | ok |
[ 0.293257] A-B-B-A deadlock: ok | ok | ok |
ok | ok | ok | ok |
[ 0.301759] A-B-B-C-C-A deadlock: ok | ok | ok |
ok | ok | ok | ok |
[ 0.308446] A-B-C-A-B-C deadlock: ok | ok | ok |
ok | ok | ok | ok |
[ 0.314094] A-B-B-C-C-D-D-A deadlock: ok | ok | ok |
ok | ok | ok | ok |
[ 0.319191] A-B-C-D-B-D-D-A deadlock: ok | ok | ok |
ok | ok | ok | ok |
[ 0.323848] A-B-C-D-B-C-D-A deadlock: ok | ok | ok |
ok | ok | ok | ok |
[ 0.328360] double unlock: ok | ok | ok |
ok | ok | ok | ok |
[ 0.332703] initialize held: ok | ok | ok |
ok | ok | ok | ok |
[ 0.336563]
--------------------------------------------------------------------------
[ 0.337730] recursive read-lock: | ok |
| ok |
[ 0.339691] recursive read-lock #2: | ok |
| ok |
[ 0.341601] mixed read-write-lock: | ok |
| ok |
[ 0.343578] mixed write-read-lock: | ok |
| ok |
[ 0.347917] mixed read-lock/lock-write ABBA: |FAILED|
| ok |
[ 0.348703] mixed read-lock/lock-read ABBA: | ok |
| ok |
[ 0.350734] mixed write-lock/lock-write ABBA: | ok |
| ok |
[ 0.353076]
--------------------------------------------------------------------------
Btw: Adding "debug" to the kernel cmdline makes the selftest output
visible while booting up (so it would be part of the LAVA CI log as well).
>
> Jan
>
--
Siemens AG, T RDA IOT
Corporate Competence Center Embedded Linux
