ASM Workshop

[avantimaestro]

There you go, FBI-o-nacci!

Task 1:

Spoiler:

.text
.align 2
.thumb
.thumb_func

main:
    push {lr}
    ldr r0, =(0x020370B8)
        ldrh r1, [r0]
    mov r2, #0x1
    mov r3, #0x1
    cmp r1, #0x0
    beq storeval_r2greater
    
loop:
    cmp r1, #0x1
    beq storeval_r3greater

    add r2, r2, r3
    sub r1, r1, #0x1
    cmp r1, #0x1
    beq storeval_r2greater

    add r3, r2, r3
    sub r1, r1, #0x1

    b loop

storeval_r2greater:
    strh r2, [r0]
    b end

storeval_r3greater:
    strh r3, [r0]
    b end

end:
    pop {pc}

.align 2

for some reason it only calculates up to the 23rd (or so, dont remember) number, any numbers over 23 make it return strange stuff. probaply because registers cant hold bigger values?

Task 2:

Spoiler:

.text
.align 2
.thumb
.thumb_func

main:
    push {lr}
    ldr r1, =(0x8044EC9)
    bl linker
    ldr r1, =(0x20370B8)
    lsl r0, r0, #0x1B
    lsr r0, r0, #0x1B
    cmp r0, #0x18
    bhs higher

end:
    strh r0, [r1]
    pop {pc}

higher:
    sub r0, r0, #0x8
    b end

linker:
    bx r1

.align 2

i adjusted it so it only returns values up to 23

Task 3:

Spoiler:

.text
.align 2
.thumb
.thumb_func

main:
    push {lr}
    ldr r0, =(0x8A71F81)
    bl  linker
    ldr r0, =(0x8A71E81)
    bl  linker
    pop {pc}

linker:
    bx r0

.align 2

this one combines both of them, nothing special

Challenge:

Spoiler:

.text
.align 2
.thumb
.thumb_func

main:
    push {r4-r5, lr}
    ldr r0, =(0x20370B8)
        ldrh r5, [r0]
    mov r4, #0x0

loop:
    add r4, r4, #0x1
        strh r4, [r0]
    ldr r0, =(0x8A71E81)
    bl  linker
    ldr r0, =(0x20370B8)
    ldrh r2, [r0]
    cmp r5, r2
    beq end
    cmp r5, r2
    blo notfibo
    b loop

notfibo:
    mov r4, #0x0

end:
    strh r4, [r0]
    pop {r4-r5, pc}

linker:
    bx r0

.align 2

takes the value of var 0x8000 (can be randomized via script) and checks if it is a fibonacci number, stores the n at 0x8000 if it is, stores 0 if its not

i have tested all of them and they worked fine for me, except for that issue with task 1.
Is there any way to store bigger values into registers?

 

[DizzyEgg]

Is there any way to store bigger values into registers?

https://www.pokecommunity.com/posts/8625292/

This should answer your question.

 

[daniilS]

Nah, the register isn't actually the problem. A register is able to hold values up to 0xFFFFFFFF. However, since the var you're storing the result in is only a halfword (up to 0xFFFF), the highest Fibonacci number it can hold is the 24th: 46368, because the next one would be 75025, or 0x12511, which is more than a halfword.

 

[avantimaestro]

https://www.pokecommunity.com/posts/8625292/

This should answer your question.

the link is explaining how to load values bigger than 0xff to a register.
i want to make a register hold values like the 36th fibonacci number (0xE3D1B0)

Can variables ony hold halfwords or can we store word to them as well?

 

[daniilS]

the link is explaining how to load values bigger than 0xff to a register.
i want to make a register hold values like the 36th fibonacci number (0xE3D1B0)

Can variables ony hold halfwords or can we store word to them as well?

They only hold halfwords. However, you could use two vars to store a word, it'd just require a small rewrite of commands like buffernumber and compare.

 

[avantimaestro]

Stuff like this would be nice to cover in a future lesson. Lets keep it simple now and accept the limitation.