Re: [PATCH 0/7] selftests/ftrace: Some improvements of ftracetest

From: Shuah Khan
Date: Fri Jun 30 2017 - 10:18:42 EST


On 06/14/2017 08:17 AM, Shuah Khan wrote:
> Hi Masami,
>
> On 06/13/2017 10:53 PM, Masami Hiramatsu wrote:
>> On Wed, 14 Jun 2017 11:12:17 +0900
>> Masami Hiramatsu <mhiramat@xxxxxxxxxx> wrote:
>>
>>> Hi Shuah and Steve,
>>>
>>> Sorry for replying so late...
>>>
>>> On Thu, 8 Jun 2017 17:51:44 -0400
>>> Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
>>>
>>>> On Thu, 8 Jun 2017 15:44:08 -0600
>>>> Shuah Khan <shuahkh@xxxxxxxxxxxxxxx> wrote:
>>>>
>>>>>> Also, did you see my reply about having the right config? You have
>>>>>> HIST_TRIGGERS undefined but yet the histogram trigger tests all pass in
>>>>>> your report.
>>>>>
>>>>> I did see that one. Didn't get a chance to look at that yet. Do you expect
>>>>> the tests to be skipped in this case? Could it be that tests aren't run?
>>>>> I will look at this and let you know what I find on my test system.
>>>>
>>>> Your first email showed that they passed, which would be hard to do
>>>> since the feature is not even enabled. ftracetest is very much set up
>>>> to look at the existing pseudo files to determine if a feature exists
>>>> or not, and if it does not, it returns "UNSUPPORTED". If it returned
>>>> PASSED when the feature is not enabled, that would be a big time bug in
>>>> the test.
>>>
>>> As far as I can see, what Steve you saw was my result on the 4.9 stable
>>> kernel, which had been compiled with fully ftrace enabled :).
>>>
>>>>>
>>>>>>
>>>>>> One more thing, is this 100% reproducible. That is, does it always fail?
>>>>>>
>>>>> Failed the two times I tried.
>>>
>>> I've also tried to test the latest linus tree with this series with Shuah's
>>> config on qemu, and could not reproduced it. (I attached the log)
>>> As you can see, with Shuah's config, histogram tests are marked UNSUPPORTED.
>>>
>>> Of course, that is 4.12.0-rc5+ and kconfig is a bit tweaked(attached).
>>> I'll try kselftest tree too.
>>
>> I've tried it on linux-kselftest/next, but not able to reproduce it...
>>
>> Thank you,
>
> Hmm. Not sure what's happening on my test system. I will try it one more
> time with a clean build and see what happens.
>
> Thanks for debugging this.
>

Re-tested this on my test system on 4.12-rc7. No failures. Not sure why I was
seeing problems. It works now.

thanks,
-- Shuah