Re: [PATCH] selftests/ftrace: Limit number of lines processed in 'trace'

["Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com>]

Masami Hiramatsu wrote:
> On Wed, 19 Oct 2022 15:15:09 +0530
> "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com> wrote:
> 
>> Steven Rostedt wrote:
>> > On Wed, 19 Oct 2022 00:19:49 +0900
>> > Masami Hiramatsu (Google) <mhiramat@kernel.org> wrote:
>> > 
>> >> > You need to make sure that the "pause-on-trace" option is set or tracing_on
>> >> > is set to 0 (disabled). Otherwise, if the tracing is still active, then the
>> >> > reading of the trace file could potentially never end.  
>> >> 
>> >> initialize_ftrace() does this setting. So it must be set.
>> >> If you run the ftracetest on old kernel, this feature is not there and
>> >> it may cause a trouble. Naveen, can you clarify it?
>> 
>> Yes, the change to not pause on opening the trace file looks to be the 
>> problem.
>> 
>> > 
>> > But for old kernels that do not have "pause-on-trace" it should be the
>> > default. The "pause-on-trace" was added when the default was changed to not
>> > pause the trace while reading it.
>> 
>> It looks like the kernel patch was picked up, but Masami's patch for the 
>> selftest wasn't backported. I have requested a test with that applied.
> 
> Good! That should be backported too.
> BTW, which kernel version do you test?

This was reported in internal tests with RHEL kernel.

> 
>> 
>> Separately, before I saw your response, I came up with the below patch 
>> to update the selftests to disable tracing before reading the trace 
>> file. I have also requested this to be tested.
> 
> Yeah, OK. This also looks OK to me.

This patch doesn't seem to be enough - event/subsystem-enable.tc also 
needs updates. In light of that, I'm not sure if the large number of 
disable_tracing statements required makes sense. So, I'm inclined to 
drop this patch in favour of the existing pause-on-trace option.


Thanks,
Naveen