| |
| |
| |
|
Page: 1 2
Comments:
<0> If one has the option of using AL instead of AX, should one do so?
<1> depends on which mode
<0> What do you mean?
<2> speaking of AL and AH being parts of AX, Is it possible to get two different halves of eax?
<0> Well, you can use AX directly, but not the other half.
<0> If I understood the documents I've read, that is.
<0> Yay! Success!
<0> Finally got this.
<0> Admittedly, I took a long break, and I got help from many people and documents, but I consider that part of the learning experience.
<0> I can now get and print characters without using interrupts :)
<3> any minute now lilo will send a global notice saying a netsplit has occured
<4> what does gdb use for stepping through ***embly code? (I'd like to have my own ***embly debugger), but I don't want it to be done through interfacing gdb via the ddd style
<5> i use stepi
<4> err stepi == commands in gdb right?
<5> yes, i meant the 'stepi' command, (maube I mis-understood your question :)
<6> I think he was asking how a debugger acutally steps one instruciton
<4> I don't want to use gdb as my debugger
<4> yes
<4> that's it
<5> ok
<6> theres probably going to be a healthy dose of "read the intel system docs"
<6> So I'll recommend it first.
<4> how do I manage
<4> to read the value of a register?
<4> it seems like to execute code to read the value of a register
<4> would change its balue
<4> *value*
<6> See "Debugging and performance monitoring" in IA-32 Intel Architecture software developer's manual, volume 3.
<5> there's a whole section about debugging in the intel manuals ; but it's not accessible to user ring progs I think. Under linux gdb uses a system call I think, something about ptrace...
<4> movl _start, %eax <-- moves the first instruction into %eax right?
<4> movl $_start, %eax <-- moves the address of the first isntruction into %eax ?
<4> (no, not hw questions; just working through a book on ***embly programming)
<6> you really really want to pick something easier then writing a deubgger.
<4> oh; I'm not going to write a debugger in ***embly
<4> it's just that writing a debugger will force me to understand ***embly really really well i feel
<6> then take a look at the gdb source code a bit too
<4> oh man
<4> gdb's code
<4> I head this nasty story
<6> is horrible.
<4> where after a few cvs revisions
<6> yeah binutils is horrible.
<4> gdb's code broke on the sparc arch
<4> and people had no idea how to debug it
<4> so then ran delta debugging on the set of cvs changes
<4> for a few days; and that condensed it down to the commit that screwed it up
<4> (something to do with a constant string being too big or something)
<4> ... I'd rather not look at gdb :-(
<6> if you want to write a debugger, it is the place to start
<6> we don't always like the answers I'm afraid =0
<7> thats true.
<4> man
<4> all these years
<4> and no one has written a _clean_ asm debugger?
<7> maybe the ones from borland where clean...
<7> pr, developing an asm debugger isn't clean.
<7> or*
<6> one of the things you learn is that alot of things become unclean the second you touch ***embly
<6> especially anything involving system programming
<4> hmm
<4> is the reason we can't use %esp (rather than wasintg %ebp) the existence of functions like alloca?
<4> which can modify the stack pointer by a varaible amount, with the var determined at run time rather than compile time
<4> so after that, we suddenly can't use %esp to calculate th eoffset of function parameters anymore
<6> the fact that the stack does not grow in functions is a "C" artifact
<6> not an ***embly restriction.
<4> oh man
<4> I'm enligntened
<4> damn
<4> wow, I can have a loop in asm that grows the stack
<4> weird
<7> a loop of subs ?
<4> yeah
<4> or pushl's
<8> irc://irc.rizon.net/md5
<9> =\
<10> heh
<6> heh?
<10> nm
<11> hehe
<12> hello there :)
<12> i want to dis***emble a small dos-exe-file, which was made in 1988 (just under 200 kb), but im a bit clueless a. how to do it b. and wich program i should use.
<12> someone has any hints? (i was told a coff dis***embler could maybe be the thng)
<1> debug.exe
<1> and 200kb is not what i call small
<1> borg also
<1> ida
<1> hiew
<1> there are a lot of tools
<12> small i meant regarding todays games
<1> oh it is a game
<6> those are debuggers, not dis***emblers?
<12> "IDA is a commercial program for dis***embling a wide array of different file types for different processors,"
<6> I stand corrected
<12> its just pasted from a site about dis***embling :)
<13> scuanor have you diss***embled programs before?
<12> edcba, yes, a game. one exe and one dat file alltogether some 250 kb and i want to habe a look into its workings
<12> archivist, no.
<14> debuggers usually can dis***emble too ;)
<14> but they are not dedicated dis***emblers of course
<6> well sure, but they arn't normally setup to do something like objdump
<12> oh no, please no debugger/dis***embler war smiles politely
<13> you generaly need more than one tool for the job
<12> archivist, is there any chance that i get to the sourcecode? or will i end up with raw machinecode?
<6> to get the src you need a dis***embler. Those are rare.
<13> yes and no, depends if the compiler has left its mark, usually a text string somewhwere, and optimised code can rarely be converted back to original source
<6> I've seen things that can get optimized code back to machine code, its not publicaly available though, you'd have to write your own.
<6> err optimized code back to source code
<6> the origonal source code? no.
<12> hm, i dont really think they did much optimizing in 1988. but... otherwise mabye they did it more in that time as today, as nowadays the machines are so powerful, that they can overcome rough coded games
<12> uhm.. are names allowed here?
<6> I think you'd be fine as long as your not giving out the game/offering it to anyone in the channel.
<12> you mean sending the game to someone? no, thats not what i would do.
<6> I am not an op, so my word isn't a guarentee.
<12> game im talking about is carrier command, a cultish strategiegame
<6> that game is almost as old as me =0
<12> ie=y
<12> but not as good as you :)
<6> =(
<12> looks like a gl*** of wine...
<12> archivist?
<13> yes?
<12> uhm, a moment please my pastebuffer has to be cleared :)
<12> ok.
<12> the mark you were takling about a while ago, could it be like this one? "Borland C++ - Copyright 1991 Borland Intl"
<12> #talking#
<13> yup thats a clue to the compiler
<12> hmpf, then that game has no mark
<13> any other text
<12> no
<12> only gamerelevant text
<12> ingame text i mean
<13> if the game was ***embler then only the writers name probably
<12> not even that
<13> may be obfuscated
<12> you mean uhm... encoded?
<13> or may be a graphics
<13> dis***embly is never easy
<12> the names i have, it was a small team of 3 students.
<14> why do you want to find the names anyway?
<12> god praise wikipedia :)
<12> oh. i understood it that way, that it could be good, if i know the mark of the compilerprogram
<12> huh? sorry archivist giggles i thought that from oggis_ line was from you
<12> oggis_ : "15:15:33] <12> archivist, is there any chance that i get to the sourcecode? or will i end up with raw machinecode?" / "15:17:22] <13> yes and no, depends if the compiler has left its mark, usually a text string somewhwere, and optimised code can rarely be converted back to original source"
<12> well thanks for help for now. :) maybe i get it to work or come back waves
<15> What's the syntax for .ifdef in as ?
<15> nvm I remembered
<16> http://www.esel-project.net/a/src/112627540048.jpg <-- Ich find den "Messerblock" so geil!
<15> How do you do #ifdef FOOBAR in AT&T asm ?
<17> Where would I find documentation on the calling conventions (typical register uses, etc) for gcc on Linux/x86_64?
<6> push arguments on stack right to left.
Return to
#asm
or
Go to some related
logs:
#php
#css
apache won't start no listening sockets available redhat
#gentoo
#perl
sansa secure clock
os.system sftp catch
#linux
deerme -site:deerme.org
#math
|
|
