Assignment: Asynchronous Device Driver
Asynchronous Device Driver
This assignment is a Xinu assignment allowing the student to more firmly understand how an operating system works. This assignment is part of the Student Built Xinu track for professors that are Teaching With Xinu. The entire working directory containing your Xinu operating system will be submission for this assignment.
First, make a fresh copy of your work thus far:
cp -R <old Xinu directory> <new Xinu directory>
Untar the new project files on top of this new directory:
tar xvzf <tar-ball location>
You should now see the new project files in with your old files. Be certain to
make clean before compiling for the first time.
Semaphore Loose Ends
system/signaln.c to finish off the interface for our semaphore subsystem. The ansynchronous driver will require all of these functions to maintain proper syncronization, but none of these semaphore helper functions is more than a few lines of code.
Dr. Dennis Brylow, quoting one of his students:
Question: We are looking at
freesem. When it says" release the process from the semaphore's waiting queue," what exactly are we supposed to do with the processes once they are released from the semaphore? Should we kill them, or put them back into the ready queue?
Answer: Put them back into the ready queue.
The project tarball includes the following files, which comprise the device interface layer for Xinu:
system/write.c, and a revised
system/initialize.c. The device layer allows user programs (like the processes in
main.c) to work with device-specific functions using standard names like
read(), rather than have to explicitly name the underlying device driver.
You do NOT need to make any changes to the device interface layer files. However, you do need to know what they are doing. Trace through a few of the high-level device functions. They are very short, and almost all the same. The file
system/devtable.c contains the master device table declarations.
The Asynchronous TTY Driver
For this project, you will be implementing an asynchronous device driver for an interrupt-driven UART. The difference between the existing synchronous UART driver and the new asynchronous tty driver is that the synchronous driver always waits for the serial port hardware to complete a read or write before returning. A proper asynchronous driver does not wait needlessly for the hardware, but instead only communicates with the UART when a hardware interrupt indicates to the system that it is ready for the next batch of work.
The Universal Asynchronous Receiver/Transmitter (UART) on our hardware seems to be largely compatible with the venerable National Semiconductor 16550 family of UARTs that have been reliably driving serial ports in PC's for decades. The PDF of the device documentation will explain the control and status registers (CSRs) on this device. The base address of its memory-mapped I/O region is noted in the master device table, devtab.
The Upper Half of a driver consists of the functions called by various user-level processes, functions such as
ttyWrite. As much as possible, the upper half functions shield the user from the messy details of the hardware.
The Lower Half of a driver consists of the handler functions run when hardware interrupt requests arrive. Lower half functions must do their work quickly, and deal with data that has been buffered up by the upper half functions.
You will be working on the upper half functions
device/tty/ttyWrite.c and the lower half function
ttyIntr() function is called via a new hook in
system/xtrap.c when a hardware interrupt request from the UART hardware arrives.
Look carefully at the structure of the asynchronous driver suggested by the structures in
include/tty.h. It is critically important that you be able to envision how the upper and lower halves of the driver should interact before you start adding code to the system.
Provide a main program that demonstrates that your asynchronous driver is working properly.
The Marquette University Embedded MIPS backends have two serial port connections. All of the backends in the lab have their second serial port available via the xinu-console system. When running
mips-console normally, note the name of the backend you are allocated. Append a "2" to the name, and run the command
xinu-console name2 to gain access to the second serial port for backend "name." The xinu-console reservation system tracks these two connections independently. Please be courteous to your neighbors and release any consoles as soon as you are done.
The "2" is assigned for ease of use, a Xinu lab may set up individual names for access to the second serial port on a backend and thus the user should understand what standard the lab he or she works with before assuming a naming convention is used.
The command "
xinu-status" will list the users on each backend, and "
xinu-status -c uart" will list the users on each backend's second serial port. Also recall that a user can bump another user off of a specific backend after 10 minutes of activity.