On Thu, Jul 27, 2017 at 10:14 AM, Kevin Wang <kevin.w...@linaro.org> wrote:

> Hi Dmitry,
> This issue is caused by the IPsec test suite trys to call odp_queue_deq()
> directly for a SCHED type queue which is not allowed from ODP API doc.
> For a SCHED type queue, the API doc indicates it must call odp_schedule()
> rather than odp_queue_deq().
> So we may need to identify the queue type before calling the
> odp_queue_deq() in ipsec_sa_destroy() in the new test case.
> I have posted some comments in your patch as well.
>
>
Thanks, Kevin. Sounds like the scalable scheduler found a bug in the test
suite rather than the other way around. A win either way. :)


> Thanks
> Kevin
>
> 2017-07-25 19:49 GMT+08:00 Bill Fischofer <bill.fischo...@linaro.org>:
>
>> Please open a bug to track this. Great to find this.
>>
>> On Tue, Jul 25, 2017 at 4:45 AM, Dmitry Eremin-Solenikov <
>> dmitry.ereminsoleni...@linaro.org> wrote:
>>
>>> On 25/07/17 12:16, Kevin Wang wrote:
>>> > I have located the root cause. We will give a proposal ​soon.
>>>
>>> This might help: https://travis-ci.org/lumag/odp/jobs/257188246
>>>
>>> Hang happens during testsuite shutdown.
>>>
>>> >
>>> > Thanks
>>> > Kevin
>>> >
>>> > 2017-07-25 16:26 GMT+08:00 Maxim Uvarov <maxim.uva...@linaro.org
>>> > <mailto:maxim.uva...@linaro.org>>:
>>> >
>>> >     test after init_main_log test hangs forever.
>>> >     also a lot of compilation issues in:
>>> >     ipsec_test_in.c:52:2: warning: missing initializer for field
>>> ‘type’ of
>>> >     ‘odp_ipsec_tunnel_param_t’
>>> >
>>> >     Is not reproducible locally?
>>> >
>>> >
>>> >     On 25 July 2017 at 10:01, Dmitry Eremin-Solenikov <
>>> >     dmitry.ereminsoleni...@linaro.org
>>> >     <mailto:dmitry.ereminsoleni...@linaro.org>> wrote:
>>> >
>>> >     > Hello colleagues,
>>> >     >
>>> >     > While working on the IPsec testsuite, I detected strange issue.
>>> My
>>> >     > testsuite works w/o issues using old queue/scheduler
>>> >     implementations but
>>> >     > will hang when used with '--enable-schedule-scalable'. I'm trying
>>> >     to get
>>> >     > more data from the system, for now you can see the situation at
>>> >     > https://travis-ci.org/Linaro/odp/builds/256914030
>>> >     <https://travis-ci.org/Linaro/odp/builds/256914030>
>>> >     >
>>> >     > --
>>> >     > With best wishes
>>> >     > Dmitry
>>> >     >
>>> >
>>> >
>>> >
>>> >
>>> > --
>>> > Thanks,
>>> > Kevin
>>>
>>>
>>> --
>>> With best wishes
>>> Dmitry
>>>
>>
>>
>
>
> --
> Thanks,
> Kevin
>

Reply via email to