[LITMUS^RT] /st_trace can not find sched_trace 0-7
Glenn Elliott
gelliott at cs.unc.edu
Sun Dec 15 17:42:49 CET 2013
On Dec 15, 2013, at 9:13 AM, Luo, Zheng <luozheng at wustl.edu> wrote:
> Hi every,
>
> I try to use the st_trace, however the program can not start correctly. I think when I compile the kernel, I already enable the trace flag (I attached the image). However it still can not find the sched_trace 0-7.
>
> I send the screen shots and the .config too. Can anyone tell me that where I did wrong?
>
>
> <Screenshot 2013-12-15 08:04:15.png>
>
> <Screenshot 2013-12-15 08:03:17.png>
>
>
>
> Zheng Luo
> <config>_______________________________________________
> litmus-dev mailing list
> litmus-dev at lists.litmus-rt.org
> https://lists.litmus-rt.org/listinfo/litmus-dev
Hi Zheng,
Take a look at litmus-rt/litmus/Kconfig. In the file you can see the configurable compile-time options in Litmus and their dependencies. You need to select CONFIG_SCHED_TASK_TRACE to enable the tracing features you want. As you can see in the Kconfig, CONFIG_SCHED_TASK_TRACE depends upon CONFIG_FEATHER_TRACE. Looking at your screenshot, I see that (1) CONFIG_FEATHER_TRACE is not set and (2) you don’t even have the option to select CONFIG_FEATHER_TRACE. This is because CONFIG_FEATHER_TRACE has the dependency “!CONFIG_RELOCATABLE” (again, look at the FEATHER_TRACE entry in the Kconfig). So:
1) Disable a relocatable kernel. I can’t recall where it is exactly, but it is probably somewhere in the general settings. In the menuconfig, you can hit “/“ and search (just like how search works in vim) for the string RELOCATABLE.
2) Enable feather trace.
3) Enable sched task trace
Note that “TRACE() debugging” generates text-based logs that can be read from /dev/litmus/log. It is purely for debugging. If you need to take performance measurements of Litmus, it is best to disable TRACE() debugging.
-Glenn
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.litmus-rt.org/pipermail/litmus-dev/attachments/20131215/51dbbeb6/attachment.html>
More information about the litmus-dev
mailing list