Jump to: navigation, search

About Me

Tom Lazar

  • Incoming Sophmore at Marquette University | Computer Science Major
  • Working On Upgrading Embedded Xinu for the Multi-Core Raspberry Pi 3
  • Interests
    • OSS
    • CS
    • Hackathons
    • Photography
    • Games, both video and board
  • Involvement
    • Marquette Student ACM, President
    • Marquette Girls Who Code, Mentor
    • FRC BEAST Robotics, Mentor

Weekly Logs

Week 1

Wednesday started working on the original pis, understanding how xinu worked and what would need to be done to get it working on the new pi 3 boards
Thursday Started working with a pi3 bare metal kernel to see the differences. found out there are a lot, the pi3's use a new architecture the aarch64, meaning we had to convert all of the assembly files from 32 to 64 bits.
Friday Got all the compilers installed and began converting the code.

Week 2

Monday Got the lite version of xinu working with the LED code, and began work on getting the serial driver to work.
Tuesday Attended RCR training.
Wednesday Got the mini-UART driver working and begun trying to get the more complicated pl011 serial driver to work. Also began to set up the xinu codebase to work with our project.
Thursday Worked on debugging the serial driver using the Oscilloscope, and did some more work setting up the xinu codebase to work with the new RPI3. The xinu-project source code is split up into platform and components, each platform has its own loader and system code, but the components subsystems remain similar for most of the various platforms

Week 3

Monday Got the Full PL011 UART Working, the problem was in the frequency we were using to control the baud rate.
Tuesday Went through the git repo and deleted some branch, edited the README and added all stable code to a dev branch.
Wednesday Helped Priya and Patrik work on getting processes working, we ran into problems with the stack that we were not able to solve. Also experimented with running XSH on the PI3's, since it required the asynchronous serial driver it wouldn't work, so I added a quick workaround that would use the synchronous.
Thursday Started researching process management in other operating systems, including Linux's completely Fair scheduler.
Friday Worked on identifying which core we are currently using and looking at different ways to disable and use them. Also helped fix several errors and problems that arose from improper check-ins and broken code.