why is in most assembler languages MOV not CPY?

flat assembler > Heap > why is in most assembler languages MOV not CPY?

Goto page 1, 2 Next

Author

Thread

MSrobots2

Joined: 21 Oct 2017
Posts: 12

I think the MOV in most assembler languages is a complete miss naming.

It never moves a value it always creates a copy and leaves the old value where it is.

it should be named CPY not MOV

Mike

22 Oct 2017, 01:07

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 16508
Location: M87*

ARM have defined CPY in their mnemonics.

Z80 uses LD.

AT&T uses MOV<size>

But to actually answer your question you'd need to ask the original designers of the CPU opcodes. If they are still alive?

22 Oct 2017, 02:02

sleepsleep

Joined: 05 Oct 2006
Posts: 8116
Location: ˛　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣ Posts: 6699

i guess mov is easier to type than cpy, and mov looks cleaner than cpy Wink

22 Oct 2017, 03:11

Furs

Joined: 04 Mar 2016
Posts: 1400

Actually, I agree. Or name should be "set" since in some architectures (x86) you can also use constants and such. "Copy" makes sense there but it's a bit less intuitive.

Code:

set eax, ebx
set eax, 5

etc

22 Oct 2017, 15:46

Tomasz Grysztar
Assembly Artist

Joined: 16 Jun 2003
Posts: 7235
Location: Kraków, Poland

I personally like the LD and ST that some architectures use. It would in fact fit x86 too, as there are separate opcodes that could be called LD (MOV reg,mem) and ST (MOV mem,reg), but Intel chose to use a common mnemonic for multiple instructions, perhaps to make it easier to remember.

22 Oct 2017, 20:40

_shura

Joined: 22 May 2015
Posts: 61

let instead of set. IMHO it sounds better Very Happy

.
But this is a question I also had for a long time. I guess, there is no reason.

22 Oct 2017, 21:42

Furs

Joined: 04 Mar 2016
Posts: 1400

I have nightmares with "let" due to "math speak"...

23 Oct 2017, 12:08

_shura

Joined: 22 May 2015
Posts: 61

let furs = not ( having nightmares ) Razz

23 Oct 2017, 18:07

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 16508
Location: M87*

In x86 MOV is a very common opcode to see in the source code. So why not remove it completely? Just put the operands without an opcode.

Code:

eax,ebx ;mov eax,ebx
add eax,ebx ;add eax,ebx
cx,5 ;mov cx,5

It could save some typing. It is not ambiguous. You could think of it as a blank opcode meaning MOV.

Code:

        eax,ebx ;blank opcode
add     eax,ebx ;explicit opcode
        cx,5    ;blank opcode

24 Oct 2017, 03:59

Furs

Joined: 04 Mar 2016
Posts: 1400

Nah, but:

Code:

eax = ebx
cx = 5

is good.

24 Oct 2017, 17:30

_shura

Joined: 22 May 2015
Posts: 61

well, you could define eax, ebx, ... as a macro and match the arguments like this:

Code:

irp name, al, ah, ax, eax, \
          cl, ch, cx, ecx, \
          dl, dh, dx, edx, \
          bl, bh, bx, ebx
  macro name arguments*
    match == operands, arguments
      mov name, operands
    else
      err `name, ': invalid operands'
    end match
  end macro
end irp

24 Oct 2017, 18:12

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 16508
Location: M87*

Don't forget about these also:

Code:

mov rsi,[mem]
mov [mem],rax
mov byte[mem],0xff

25 Oct 2017, 01:30

shoorick

Joined: 25 Feb 2005
Posts: 1605
Location: Ukraine

Code:

    eax ; mov eax,0

25 Oct 2017, 09:47

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 16508
Location: M87*

shoorick wrote:

Code:

    eax ; mov eax,0

Code:

macro a { xor eax,eax } ;works for both 32-bit and 64-bit modes

25 Oct 2017, 12:07

_shura

Joined: 22 May 2015
Posts: 61

memory as source-operand should work with this code, for memory as destination-operand you can do this:

Code:

irp name, byte, word, dword
  macro name arguments*
    match [ memory ] == operands, arguments
      mov name [ memory ], operands
    else
      err `name, ': invalid operands'
    end match
  end macro
end irp

but

Code:

[ eax ] = 1

is not possible afaik.
why should

Code:

eax

equal to

Code:

mov eax, 0

?
anyway, some extra bits:

Code:

irp name, byte, word, dword
  macro name arguments*
    match [ memory ] == operands, arguments
      mov name [ memory ], operands
    else match [ memory ] ++, arguments
      inc name [ memory ]
    else match [ memory ] --, arguments
      dec name [ memory ]
    else
      err `name, ': invalid operands'
    end match
  end macro
end irp
irp name, al, ah, ax, eax, \
          cl, ch, cx, ecx, \
          dl, dh, dx, edx, \
          bl, bh, bx, ebx
  macro name arguments*
    match == operands, arguments
      mov name, operands
    else match ++, arguments
      inc name
    else match --, arguments
      dec name
    else
      err `name, ': invalid operands'
    end match
  end macro
end irp

25 Oct 2017, 14:54

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 16508
Location: M87*

_shura wrote:

why should
Code:
eax
    
equal to
Code:
mov eax, 0
    
?

Yah, it should be xor eax,eax Twisted Evil

25 Oct 2017, 15:29

_shura

Joined: 22 May 2015
Posts: 61

I mean, if I read a single line »eax« in my code, it is not obvious what this mean. I looks like a function call:

Code:

call eax

but not like

Code:

xor eax, eax

and even a function call should look more like

Code:

eax()

which can be done with:

Code:

struc (symbol) ? definition&
  match =(=), definition
    call symbol
  else match =( arguments =), definition
    callFunction eax, arguments
  else
    err `symbol, ': what do ya want from me?'
  end match
end struc

but you have to implement »callFunction« yourself.

25 Oct 2017, 15:40

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 16508
Location: M87*

_shura wrote:

and even a function call should look more like
Code:
eax()
    

We don't write C code here. Laughing

25 Oct 2017, 16:26

_shura

Joined: 22 May 2015
Posts: 61

so eax = 1 should not allowed too Very Happy

25 Oct 2017, 18:00

Tomasz Grysztar
Assembly Artist

Joined: 16 Jun 2003
Posts: 7235
Location: Kraków, Poland

_shura wrote:

but
Code:
[ eax ] = 1
    
is not possible afaik.

You could do that with "macro ?" if you really wanted.

25 Oct 2017, 18:55

Goto page 1, 2 Next

< Last Thread | Next Thread >

Forum Rules:

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You can attach files in this forum
You can download files in this forum