You should utilize the provided skeleton code along with the code from the first picture $($conv$).$ Here's the task: You'll be creating a function to perform $2$D convolution by invoking the original function, following the same assumptions $($a fixed $3$x$3$ filter, with a stride of $1$ and no padding$).$ This time, you'll need to write the output array on the stack and print the entire array within the function. Your $2$D convolution function will take $4$ parameters: input height, input width, a pointer to the start of the input array, and a pointer to the filter array. Assumed maximum size for input and output matrices is $25$ by $25.$

You'll need to manage several iteration variables across calls to the original conv function. Utilize the $s registers for this purpose, ensuring to save and restore them in the conv function using the stack. Additionally, the conv$2$D function should appropriately handle the $s registers, as it will be tested to ensure their unaffected state.

The first two pics are the conv code. The next two are the skeleton code that you need to use.