By Wenzhe jiang (wj3) & Jun Liu (jl84)

Description of submitted source files:

1. yalnix.c: This is the main file of the kernel. It includes the kernelStart function,
SetKernelBrk function, and all trap handlers.

2. kernelcall.c: This file is the implementation of all kernel calls.

3. load.c: This file is specifically for loading a specified program into memory.

4. utilities.c: This file includes all self-defined functions in order to help yalnix.c, 
kernelcall.c, and load.c.

5. kernel.h: This is the headfile for all source c files.

6. README: A description of the whole project.

7. other test files

Design and algorithm highlights:

1. Page table: In order to efficiently utilize the memory, we write a separate function
for page table allocation. Every time we need a page table, we separately get a free page
from the free page list. If one page is only half occupied (with one page table), we 
remember it and use the other half for next time use. In this way, we guarantee there will
be at most half page of unused memory at any time.

2. Fork implementation: In fork function, we need to copy the r0 region from one process
to another. A straightforward way is to first copy it to r1, then switch page table and 
copy from r1 back to new r0 region. This method requires 2 times memory of one process.
In order to avoid this, we borrow a PTE from the guaranteed unused region in r0 region 
(in our implementation, we use the page start from USER_STACK_LIMIT) to map to
the physical address of new r0 region. Then we copy the contents page by page and 
invalidate the borrowed PTE afterwards.

Testing:

We thoroughly tested our kernel using all given samples in the pub/samples-lab2 folder on
clear without any problem.