[ASM Tutorial] Tarek701's MIPS Tutorial
Users browsing this thread: 1 Guest(s)

Hello, people!
This is the latest revision of my MIPS Tutorial (from end 2015) and should cover the most basics of MIPS ASM. Floating-point operations and other stuff was not finished, but isn't required if you're still a beginner in MIPS ASM.

Have fun, coding.
(This post was last modified: 14-10-2016, 05:33 PM by Tarek701.)

Attached Files
Size: 347.05 KB / Downloads: 550 .zip   mips3.zip

R.I.P Tarek701. 2005-2016

Seems quite helpful, I'll look at this and try my hand at coding ^^
Shining Stars 3: Sanctuary of the Star Comet (SM64 hack) coming 2016!

Seems like it was written for kids. Lets see if I can't find a few errors:

1) The N64's General Purpose Registers are 64 bits wide. You can confirm this by checking the exception handler code, since when swapping threads it saves each GPR with a SD command (unless the N64 OS was updated for Ocarina of Time).
2) In your explanation of the difference between ADDI and ADDIU, you state that "ADDIU will not error when our value goes over 0x7FFF", which is wrong. Overflow (if I'm understanding things right) should occur when the sum of two values exceeds 2^63-1. Don't know for sure though since I never used ADD/ADDI/etc.
3) "This instruction stores the current return address (in RA) to the stack, on SP + 0x14 aka 20-bits." Bits should by bytes.
4) "I'm not angry or anything if you didn't understand some parts": This is a really odd thing to write. I don't think the reader would ever assume that. Also in the same paragraph, "rogue" is the wrong word to use there.
5) "J 0x8033D666"/"BEQ T0, T1, 0x80038A4D" - These examples are invalid, since jumps and branches must always be 4 byte aligned, and it seems poor form to have invalid examples.
6) "A normal J instruction is a "direct" jump" - Technically no, they are not. Only JR/JALR instructions allow you to perform a direct jump. J/JAL are limited to I think a 256 MB address "page", where the full jump is dependent on the state of the PC register. Only hardware level hackers need to worry about this though.

Unrelated, # for decimal? Invented a pseudo-opcode just for the parser? Why

(11-10-2016, 10:04 PM)mzxrules Wrote: Seems like it was written for kids. Lets see if I can't find a few errors:

It's supposed to be for "beginners", therefore it is kept very rough in some parts without giving out too much detail, and once I had the idea of making some "tests" for my tutorial, but I knew that was pointless, so I never continued that path. If it is in your eyes written in the style of a kids' novel, so be it. However, most people seem to have been profiting from this revision more than from my older ones. So, I think this style is completely fine.

1. Yes, they are 64-bit wide. However, as far as I remember the registers were always used in 32-bit mode (at least in N64), not 64-bit mode, thus me purposefully leaving out the fact of the MIPS registers actually being 64-bit wide. Also, most N64 debuggers I know of only display GPU registers in a length of 32- bits. Still, I added this small detail of the registers actually being 64-bit wide.

2. There existed (at least on StackOverflow) a lot of confusion before whether 16-bit immediate values in ADDI and ADDIU are both signed or unsigned (as the latter instruction in its name does imply). In fact, both instructions are sign-extended, with the difference that ADDI throws an overflow exception while ADDIU does not. In that connection the question came up what would happen if the 16-bit immediate value in the ADDIU instruction exceeds 0x7FFF, where I wanted to clarify that in this case no overflow exception is given. Of course, I was wrong on the length though because I mistakenly was referring to the 16-bit immediate value and not the sum, from which (as you said correctly) the overflow exception would occur from, if it exceeds the max integer value in the register. Also, if I'm right the max value should (in the case of N64, running in 32-bit mode) in fact be 2^31-1 and not 2^63-1.

3. Typo, fixed.

4. Odd phrase, fixed to a more objective phrase. "rogue" should be "rough" or "superficial", fixed too.

5. I was kinda optimistic that people would get that themselves, which is why I didn't use it in an actual instruction but in the explanation in the instruction formats. Fixed it anyway.

6. That detail isn't really necessary for beginners though, which is why I roughly summarized it, and marked a hint with quoting "direct" because it actually isn't a direct addressing, more like pseudo-direct addressing as you pointed out. I still add it though in brackets, so "experts" (which I doubt most beginners will be, but whatever...) or curious people get more detail on this.

Thanks for counting up some minor mistakes.
I stopped working on that tutorial for a while and some errors listed by you were supposed to be clarified and fixed (I remember kaze telling me about them long ago) except for the typos (thanks btw), but when I lost interest in SM64 Hacking a while ago, I stopped continuing on this and just released it now in the state as it was before. Sorry 'bout that.

Possibly in the near future I might get enough time to completely rewrite this and walk a middle path (giving additional details (footnotes), but still focusing on the basics instead of just focusing on the basics and completely leaving out the details).

(11-10-2016, 10:04 PM)mzxrules Wrote:
Unrelated, # for decimal? Invented a pseudo-opcode just for the parser? Why

Because it was my decision to do so. Originally, I planned to have decimal values have not any symbol (as this is the case for most assemblers), however I noticed that this didn't seem to work out very well, so I did leave it in the way it was since 1.0 release.

As no one (from the communities I know of that regularly use my assembler) complained about this so far, I don't see it as a big deal anyway and thus there's no need to do anything about it.

Thanks again, btw.
(This post was last modified: 14-10-2016, 05:34 PM by Tarek701.)
R.I.P Tarek701. 2005-2016

I have to say, your tutorial has improved a lot since the last time i checked (it was like 2 years btw). back then, i wasn't able to make a working hello world, but i decided to give MIPS ASM coding a shot again.

Well, this time i was able to do my first "complex object" first-try without any error or problem!


I applied this to a solid platform, and it worked as intended. It basically explodes the floor when i groundpound pressing A at the same time. I was wondering, the original plan i had was to make a platform that allows you to do a "Groundpound Jump" (maybe using the 3rd jump animation) but i'm still very beginner to this and i have no idea on how to continue... any help?

Last, thanks for the tutorial, it's extremely easy to understand and a huge improvement since last I checked it, excellent work tarek c: hope this is my start at asm coding!
(This post was last modified: 23-12-2016, 02:17 PM by Mariohacker14.)
[Image: Mastery_sketchup.png] [Image: Mastery_paintdotnet.png]

[ASM Tutorial] Tarek701's MIPS Tutorial
Users browsing this thread: 1 Guest(s)

  EN ・日本語