| |
| |
| |
|
Page: 1 2
Comments:
<0> where can i find the syntax for gcc's ARM ***embler?
<1> hey all
<1> does anyone know a good resource for small asm programs sources?
<2> lukus001, I have one on my server, if you want a link
<1> yes please, everything help in my learning process ^^
<2> http://mikemiller.dyndns.org/rot13.asm
<2> please note that that program uses linux kernel interupts and will not work on a non-linux box
<1> or my 64bit linux box..? :)
<2> It should work
<1> probabily, but i wont risk running it anyway lol ! i just freshly installed this baby
<2> actually, on 64 bit it wouldn't work
<1> What is this program used for may i ask?
<2> that program is writen ***uming a pointer is 4 bytes wide (such as in ia32), and 64 bit machines use 8 byte pointers
<2> lukus001, the program does pseudo-encryption
<1> ah cool
<1> Thanks for the file
<2> your welcome, the source should be ***embled with fasm
<1> ill explore it's guts tomorrow
<1> fasm, ok ill make a note of that :)
<1> thanks again, bye :)
<0> can anyone recommend a fixed point division routine that doesn't require a much larger intermediate type?
<3> lol, rot13
<2> ned, what?
<4> ahoy hoy
<5> hello
<6> hello
<7> does someone know a good tool to convert simple asm to C ?
<6> hmm, not me :)
<8> hmm
<8> why don't you use asm directly?
<7> I want to convert the ***embler from the huffyuv project back to C
<8> hmm
<8> then it's better to rewrite it
<7> rewrite>?
<6> !pastebin
<6> no paste?
<9> http://modseven.de/pastebin.php
<8> that you write it new in C
<6> i have this problem:
<8> yeah
<8> what's the problem?
<6> http://modseven.de/pastebin.php?id=993
<6> i'm not very familiar with intel syntax
<8> learn it, it's quite easy
<6> i tried to do:
<6> movl 0xDEADBEEF, %eax
<6> movl $0xDEADBEEF, %eax
<6> but they both get outputted as: B8 EF BE AD DE mov eax,0DEADBEEFh
<6> anyone has any idea why this? is my intel syntax wrong?
<8> yes
<8> i know it
<8> its little endian
<8> i mean it's correct
<8> the lowest byte will be stored at first
<6> no
<6> not this
<6> if you look at disasm.exe output
<6> (the one from the link)
<6> you see that both instructions output same opcodes
<8> ok i will have a look
<9> that's a bit weird actually..
<6> so my intel syntax must be wrong, or it's something plain wrong with dissasm or msft compiler :)
<8> re
<8> so what is the $ ?
<8> the intel Syntax hasn't such thing
<8> you can just write
<6> no, please read the link
<6> http://modseven.de/pastebin.php?id=993
<6> pusha
<6> mov eax, [0DEADBEEFh]
<6> mov eax, 0DEADBEEFh
<8> you can't write mov 1233,eax
<6> that's what i've used
<8> only mov eax,123
<8> or mov [123],eax
<8> wait a moment i'm busy
<8> okay that the lines
<8> mov eax, [0DEADBEEFh]
<8> mov eax, 0DEADBEEFh
<8> get the same code is strange
<8> what ***embler are you using?
<6> i use inline ***embler
<8> oh
<8> in delphi?
<6> vc++
<6> http://modseven.de/pastebin.php?id=994
<6> here is the actual code
<8> hmm the inline ***emblers are not the best
<8> i think
<8> the syntax is a bit different
<8> in vc++ to store in asm something from a address
<8> not the brackets
<6> no?
<6> hmm
<8> i do not think
<6> it should report an error at least
<8> it would be the only explanation
<6> wow
<6> another strange thig:
<6> http://modseven.de/pastebin.php?id=995
<6> i've modified one instruction
<6> ah, that's wrong on my side
<6> i need mov dword ptr
<6> let's see if previous example works with dword ptr
<6> no
<6> doesn't compile at all
<6> errrem
<7> mov ebx,eax; mov cl,al; what's 'al' and 'cl' ?
<6> they are registers
<10> low 8 bits of eax and eax
<6> al is the first byte in ax, ax is first word in eax
<10> * and ecx
<6> cl is the first byte in cx, cx is first word in ecx
<7> in C it would be: int ebx = eax; char cl = (char) eax) & 0xff; - right?
<10> ecx = ecx&-0x100 | eax&0xFF
<10> since cl is a part of ecx
<8> re
<7> inline gcc ***embler is "mov src, dest", right?
<6> yes
<6> movl actually
<6> or movb or movw
<6> or movq
<7> how does "byte ptr [&decode_shift_table+ecx]" translate to gcc asm?
<6> movb %al, %ecx(decode_shift_table)
<6> this moves al to that location
<11> many times gas can "guess" the operand size though
<6> i don't like to rely on that
<6> that's why i don't use intel syntax in the first place :)
<7> and "[ebx+edx+1]" ?
<11> "rely"? you don't have to "rely" on it specifically, if there is any ambiguity you'll get an error
<6> %ebx(%edx,1,,) or something like that
<6> oggis_ yes but it's not readable
<11> that's your opinion
<6> yes but i use the code, so it's my opinion that matters :)
<11> my opinion is that intel syntax is more readable than at&t in every regard. at&t is nice in a way that it doesn't have ambiguity, which makes writing the ***embler easier, because it doesn't have ta have any deduction logic, but imo writing and reading the ***embly itself much harder
<11> well, it's not like at&t is impossible to hand-write, just more cumbersome and less enjoyable
<7> and gcc uses at&t ?
<11> it can use both
<11> by default, yes
<11> (this is of course only on intel)
<11> (platform)
<7> and yasm uses what?
<11> nasm, which is mostly like intel syntax
<11> by default
<11> and also at&t
<8> nasm uses intel syntax
<8> at 100 percent
<8> or say at 99 and two percent
<11> well, no
<8> 98 i mean
Return to
#asm
or
Go to some related
logs:
totem-exine
#dns
#linux
sed commands from command line :a;N;$!ba
#centos
Luminocity rpm
Driver error was 1/2002 mysql
ubuntu atiixp no sound
qemu fsimage
#qemu
|
|