Somil Gupta commented: 
https://gitlab.rtems.org/rtems/programs/gsoc/-/issues/18#note_144149


In continuation of the discussion on RTEMS Users\
https://users.rtems.org/t/project-interest-in-add-filesystem-benchmarking-tools-to-rtems/482

I am pleased to report that I am now able to successfully run an FIO job on a 
QEMU-based target.

During this process, I have identified the following issues:

1. **Command Parser Limitations**\
   The command parser is not functioning correctly. As a workaround, I executed 
FIO by placing the configuration inside a job file instead of passing arguments 
via the command line.
2. **Hardcoded Build Configuration**\
   The Makefile in the GitHub repository referenced in the paper 
(https://gedare.github.io/pdf/agarwal_comparison_2019.pdf) is hardcoded for the 
BeagleBone Black and makes several board specific assumptions.\
   I modified it to work with my QEMU-based target (Xilinx Zynq-A9), but my 
changes are also currently hardcoded. The build system needs to be generalized 
to support multiple BSPs cleanly.
3. **Endianness Configuration Issue**\
   Running `./configure` does not correctly set `bigendian = yes` for 
SPARC/ERC32 target .
4. **Board-Specific Time Implementation**\
   `gettime.c` produces an error on the first run and requires board-specific 
adjustments.
5. **Minor Function Definition Corrections**\
   Some function definitions require slight modifications (e.g., explicitly 
adding `int` in certain places).
6. **NULL Handling in strcmp**\
   The code aborts when `NULL` is passed to `strcmp`, particularly in `init.c`. 
This needs handling to avoid runtime failures.

At this stage, I believe I can begin writing my gsoc proposal and stabilizing 
the initial FIO port by addressing these issues systematically.

Regarding workflow should I fork the GitHub repository referenced in the paper 
and start contributing fixes there, or is there a preferred workflow given that 
the long goal is to create a new home for benchmarking tools? ( 
https://gitlab.rtems.org/groups/rtems/-/epics/17 ) 

I’d appreciate some guidance on how you’d prefer me to move forward with 
development and contributions.  @gedare @c-mauderer 

![image.png](/uploads/472c0dca731c186237bdc3a96544b4cf/image.png){width=721 
height=269}

-- 
View it on GitLab: 
https://gitlab.rtems.org/rtems/programs/gsoc/-/issues/18#note_144149
You're receiving this email because of your account on gitlab.rtems.org.


_______________________________________________
bugs mailing list
[email protected]
http://lists.rtems.org/mailman/listinfo/bugs

Reply via email to