rt-tester testcase t4-l2-pi-deboost.tst failed

rt-tester testcase t4-l2-pi-deboost.tst failed

[Yao Zhao]

Hi,

I tested rt-tester with mainline 3.8.0 kernel and
the kernel scripts/rt-tester/t4-l2-pi-deboost.tst failed at

# Verify prio of T0 (still pending owner,
# but T2 is enqueued due to the previous boost by T3
T: prioeq:              0:      82

The current T0's prio is -21 so it failed.

Reading source code, it should be 82 as T2 is at prio 82 and
T0 owns the L1 and T2 is waiting for the L1.

Can I ask
1.is the rt-tester test only for preempt-rt kernel?
2.any idea why it failed?

thanks,
yao

Re: rt-tester testcase t4-l2-pi-deboost.tst failed

[Thomas Gleixner]

On Wed, 27 Feb 2013, Yao Zhao wrote:

> Hi,
> 
> I tested rt-tester with mainline 3.8.0 kernel and
> the kernel scripts/rt-tester/t4-l2-pi-deboost.tst failed at
> 
> # Verify prio of T0 (still pending owner,
> # but T2 is enqueued due to the previous boost by T3
> T: prioeq:              0:      82
> 
> The current T0's prio is -21 so it failed.
> 
> Reading source code, it should be 82 as T2 is at prio 82 and
> T0 owns the L1 and T2 is waiting for the L1.
> 
> Can I ask
> 1.is the rt-tester test only for preempt-rt kernel?

No

> 2.any idea why it failed?

No

It would be nice if you could figure out whether that's a failure of
the test case or a failure in the kernel.

Thanks

	tglx

Re: rt-tester testcase t4-l2-pi-deboost.tst failed

[Yao Zhao]

On 13-02-27 05:16 PM, Thomas Gleixner wrote:
> On Wed, 27 Feb 2013, Yao Zhao wrote:
>
>> Hi,
>>
>> I tested rt-tester with mainline 3.8.0 kernel and
>> the kernel scripts/rt-tester/t4-l2-pi-deboost.tst failed at
>>
>> # Verify prio of T0 (still pending owner,
>> # but T2 is enqueued due to the previous boost by T3
>> T: prioeq:              0:      82
>>
>> The current T0's prio is -21 so it failed.
>>
>> Reading source code, it should be 82 as T2 is at prio 82 and
>> T0 owns the L1 and T2 is waiting for the L1.
>>
>> Can I ask
>> 1.is the rt-tester test only for preempt-rt kernel?
> No
>
>> 2.any idea why it failed?
> No
>
> It would be nice if you could figure out whether that's a failure of
> the test case or a failure in the kernel.
For the whole check-all.sh t4-l2-pi-deboost.tst is the first test case
which failed and in t4-l2-pi-deboost.tst: that line is the first "test"
failed.
I didn't see any failures in the kernel.

But I will spend some time to debug around and get familiar with
the rt-mutex code.

thanks,
yao
> Thanks
>
> 	tglx