mirror of
https://github.com/reactos/reactos.git
synced 2024-10-31 03:48:17 +00:00
e6cfb666ca
Fixing Li bug the register for dest was wrong calctions see file OpCodePPC.txt. fixing some meaing in ReadMe.txt svn path=/trunk/; revision=25511
76 lines
2.5 KiB
Plaintext
76 lines
2.5 KiB
Plaintext
CpuToIntel is a experment tools and is strict under havy devloping
|
|
|
|
|
|
The Idea
|
|
The idea is to converting binary files or win pe files
|
|
from one cpu to another cpu, But it does not exists
|
|
plan to port over diffent hardware architect like
|
|
how diffent hw comucate, example x86 DMA controller
|
|
to PPC like that stuff. It is only to convert the
|
|
binary or pe files to another cpu. it mean a user
|
|
mode apps will always be ported, but if it self
|
|
modify code it will not work. But it exists idea how
|
|
to deal with self modify code.
|
|
|
|
|
|
The idea to handling self modify code
|
|
The idea is to add a small emulator or adding
|
|
anaylysing process to dectect self modify code
|
|
and extract it. This is very hard part todo, some say
|
|
imposible, some other say almost imposble. and I say
|
|
it is posible todo but extream hard todo. for it is
|
|
very diffcul to dectect self modify code with a
|
|
analysing process.
|
|
|
|
|
|
Why the name are CpuToIntel
|
|
When I start write on it it was only ment to convert
|
|
from ARM, PPC, m68k to X86 but then I come think of
|
|
ReactOS PPC port that is going on. for or later we
|
|
will need something that doing convert from x86 to
|
|
PPC apps. It exists two way todo it. One is to use
|
|
dymatic translation a jit, like UAE or QEMU doing
|
|
converting. But it will lose of allot of speed if
|
|
it is a game or a havy apps to much. So the idea
|
|
is to convert the whole file in one sweep. will give
|
|
one other problem it will be a slow process todo it,
|
|
and hard dectect self modify program. so not all program
|
|
can be really convert with this process.
|
|
|
|
|
|
Who will it work
|
|
we take it step for step and I will describe the
|
|
binary translations how it works. The PE file
|
|
work simluare way.
|
|
|
|
step 1 : it will disambler the program frist
|
|
|
|
step 2 : translate everthing to a middle asm dialect,
|
|
it is own asm dialect it is not suite for a real
|
|
|
|
step 3 : (not implement) send it to ananalysing processs
|
|
to get any name or mark out which row is a new functions
|
|
|
|
step 3.5 (not implement) split the code into functions here
|
|
|
|
step 4 : Now it start the convert process.
|
|
|
|
step 4.5 (not implement) maybe a optimzer.
|
|
|
|
step 5 : now it is finish.
|
|
|
|
|
|
The arch that are plan
|
|
PPC to IA32, PPC (work in progress)
|
|
m68k to IA32, PPC (stubed)
|
|
ARM to IA32, PPC (stubed)
|
|
IA32 to IA32, PPC (work in progress)
|
|
|
|
|
|
The Winodws NT PPC and x85 diffrent
|
|
R1 The stack pointer equal with x86 esp
|
|
R3 The return reg equal with x86 eax
|
|
R4 The return reg equal with x86 edx
|
|
R31 The base pointer equal with x86 ebp
|
|
|
|
|