fasm&fasmg

Index > Main > fasm&fasmg

Author

Thread

l4m2

Joined: 15 Jan 2015
Posts: 674

l4m2 14 Aug 2016, 16:08

In fasm there are three kinds of IF X=0:

Code:

match =0, x {
  ...
}

Code:

if x = 0
  ...
end if

Code:

cmp x, 0
jnz lab
...
lab:

In fasmg only two?

Code:

if x = 0 
  ...
end if

Code:

jnz lab, x
...
lab:

How does fasm&fasmg know x is linear function of a, b, c?(a, a+b, a+b+1 are, but a*a, a+d aren't)

14 Aug 2016, 16:08

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 20489
Location: In your JS exploiting you and your system

revolution 14 Aug 2016, 19:32

You are mixing all the stages of the process together and confusing yourself. fasm has preprocessor and assembler stages, whereas fasmg has combined those into the same stage. The generated code is executed at runtime.

To answer your second question: fasm doesn't know if anything is a linear function of anything else. It just tries to find values that produce a valid solution. There may be other solutions that are also valid and fasm can completely miss them.

14 Aug 2016, 19:32

l4m2

Joined: 15 Jan 2015
Posts: 674

l4m2 15 Aug 2016, 01:28

revolution wrote:

You are mixing all the stages of the process together and confusing yourself. fasm has preprocessor and assembler stages, whereas fasmg has combined those into the same stage. The generated code is executed at runtime.

When I see a topic, usually I can't know whether it's a processing code or an executing code, thus mixing them is acceptable.

revolution wrote:

fasm doesn't know if anything is a linear function of anything else.

I tried these cases:

Code:

lea eax, [3*ecx-2*ecx]      ;OK
lea eax, [ecx*ecx-ecx*ecx]  ;Error

which means that fasm requires all the summed elements to be linear

15 Aug 2016, 01:28

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 20489
Location: In your JS exploiting you and your system

revolution 15 Aug 2016, 01:51

l4m2 wrote:

I tried these cases:
Code:
lea eax, [3*ecx-2*ecx]      ;OK
lea eax, [ecx*ecx-ecx*ecx]  ;Error    
which means that fasm requires all the summed elements to be linear

Now you are talking about registers, rather than your previous example of variables. The x86 CPU addressing doesn't support multiplication of addresses, only shifts and adds.

BTW: Did you try [ecx*(ecx-ecx)*ecx]?

15 Aug 2016, 01:51

l4m2

Joined: 15 Jan 2015
Posts: 674

l4m2 15 Aug 2016, 02:06

revolution wrote:

Did you try [ecx*(ecx-ecx)*ecx]?

No. But maybe my "element" should be "monomial" Very Happy

15 Aug 2016, 02:06

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 20489
Location: In your JS exploiting you and your system

revolution 15 Aug 2016, 02:27

l4m2 wrote:

When I see a topic, usually I can't know whether it's a processing code or an executing code, thus mixing them is acceptable.

Only if you find it acceptable to confuse yourself. Wink

But seriously, if you don't understand at least the preprocessor vs assembler difference then you will have a hard time understanding complex macros.

15 Aug 2016, 02:27

< 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 cannot attach files in this forum
You can download files in this forum