Feedback on Bash script to review BSP File Organization

Vipul Nayyar nayyar_vipul at
Thu Sep 19 00:46:36 UTC 2013


I also encountered the first error of command not found. Actually the real name of  the function defined is test_its_NOT_there. I corrected this mistake but maybe didn't submit it in the patch. Will correct it.

BTW, the older code is written for evaluating a single bsp & mine for multiple but I've integrated the previous code in my file in the function which evaluates only a single bsp. I don't evaluate all bsps for a single issue. Each bsp is evaluated completely in its own order for different issues. So, the previous code should work properly. And I think it is.

Can you please tell me for which bsp you're getting the errors of RTEMS_BSP_CLEANUP_OPTIONS is NOT in ./ & grep ?
I checked the script without giving any extra options. And I'm not getting any problems in arm bsps.

The results that I mentioned were found in 
	* arm/gumstix : bspreset.c not compiled 
	* arm/gumstix : bsp_reset() present in file startup/bspstart.c
	* powerpc/beatnik : bspreset.c not compiled 
	* powerpc/beatnik : bsp_reset() present in file include/bsp.h startup/reboot.c
	* powerpc/mpc55xxevb : start.S not present in correct path 
	* powerpc/mpc55xxevb : bspreset.c not compiled 
	* powerpc/mpc55xxevb : bsp_reset() present in file startup/reset.c 
	* powerpc/mpc55xxevb : bspgetworkarea.c not compiled 
	* powerpc/mpc55xxevb : bsp_work_area_initialize() present in file startup/bspworkareainit.c
Either you can check them manually or the script is also giving me the same output with these warnings.

Vipul Nayyar 

 From: Joel Sherrill <joel.sherrill at>
To: Vipul Nayyar <nayyar_vipul at>; "rtems-devel at" <rtems-devel at> 
Sent: Wednesday, 18 September 2013 11:37 PM
Subject: Feedback on Bash script to review BSP File Organization


I wanted to run the script and see how I could break it. :)

I ran it from libbsp/arm and got a few oddities and error

line 232: test_its_not_there: command not found

The above is clearly an integration issue from merging the


The above appears to be from the older code being setup
to be inside a single BSP directory while the newer code
scans all BSPs for a single issue.

=== Checking for not using -O2
grep: make/custom/*.cfg: No such file or directory
=== Checking for turning on extra GCC warning checks
grep: make/custom/*.cfg: No such file or directory

Same issue as previous but manifested differently.

=== Checking for timing information
Please run the timing tests and include the results.


I think the older tests need to be adjusted. Each may
need to be wrapped in a loop to cycle over each BSP
and invoke the test in the right subdirectory. Or the
test itself modified.

Also please point me to a case where a required method
is in the wrong file. I recall one in your previous output
but am missing it now.

A little clean up or a check that you are in a single BSP
directory is needed. I am not sure which is the right
solution. I was leaning to it working on a single BSP directory
since that is likely how someone will use it.

But supporting both modes correctly is useful.

Joel Sherrill, Ph.D.             Director of Research & Development
joel.sherrill at        On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
Support Available                (256) 722-9985
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the devel mailing list