| |
| |
| |
|
Page: 1 2 3
Comments:
<0> hey dudes
<1> hi
<2> aaaay
<0> yo
<3> what's happening....dude?
<1> eiiiieeeeeeeeeiiiioooooooo!
<1> nothing. just terminal insanity stages.
<3> wth does that mean ..dude?
<1> this means i'm crazy guy
<1> y'know, madhouse breakout
<2> nisc?
<4> wat
<2> it mean anything?
<2> national inteligence security company maybe?
<4> yes
<4> exactly
<5> http://www.ics.uci.edu/~nisc/ how strange
<4> for more funny stuff visit acronymfinder.com
<6> does anybody have the IFS kit for XP?
<1> IFS?
<1> [why do i have a feeling i'll regret googling it...]
<6> installabel file system
<6> to intercept file I/O
<1> ah... one of those virtual-drive thingies?
<1> no, wait, i googled it out
<1> so, let me ask
<1> what the hell has it to do with asm?
<7> Is there an AT&T syntax directive to make sure .text is writeable ?
<7> Like for self-modifying code ?
<1> self-modifying code is evil.
<1> anyway,
<1> hm.. no
<1> there is no way to make .text writable
<7> are you sure ?
<1> the only solution is to make a new section and ***emble there.
<1> yup.
<1> .section .mytext, "wax"
<1> [your code]
<7> because in, for example, ELF executables you can set .text to writeable
<7> Taht much I know
<1> w is 'writable', a 'allocatable', x 'eXectuable'
<1> yes
<1> but gas doesn't like it.
<7> Aw :(
<1> besides, point is moot. ELF specification specifies standard type and flags for multiple sections, including .text
<1> it's SHT_PROGBITS, SHF_ALLOC|SHF_EXECINSTR
<7> OIC
<1> no SHF_WRITE.
<1> so, gas does exactly what it's told to do by the spec.
<1> [not that the spec matters much now... it's been extended so many times that you probably won't see any strict implementation]
<1> also, let me repeat this once more.
<1> self-modifying code is evil.
<7> How so ?
<1> except maybe linkage and signal trampolines generated at runtime... but those are made on stack
<1> let me enumerate:
<1> 1. non-portable
<1> 2. non-thread-safe if done in shared section
<7> markos_64, ASM in general isn't very portable
<1> 3. hard to pull off
<7> How about change "hard to pull off"
<7> To "challenging" ?
<1> 4. practically useless, except maybe for dynamic code compillation... which isn't done in pre-allocated data section anyway
<7> By no means would I write any real software in self-modifying code
<7> But it's interesting
<1> whatever suits your taste
<2> http://pastebin.ca/79155
<2> T6: shr (:= n != 0 -> C(31..0) <- 0#C<31..1>: n <- n-1; Shr));
<2> if the shr ckt has 1 gate, how many times is that gate used in T6
<2> I say n-1 times
<2> Goodbye 4th of July! Hello 5th of July!
<8> hello
<9> hello
<8> can u help me about asm?
<8> its about movsd and cdecl calling convention
<9> i'll try..
<9> okay
<9> what about 'em
<8> movsd copies block of memory right?
<9> movsd moves a doubleword address from DS:ESI to adresse ES:EDI
<9> yeah
<9> byte, word ou dword...
<9> D = dword
<9> there's also MOVSB and MOVSW
<8> what is DS:ESI there?
<8> h^^
<8> i'm pretty much new in asm
<9> DS and ES are place holders for segment selectors
<9> so id DS and ES are the same
<9> and let's say ESI = 0h and EDI = 32h
<9> MOVSD will move (actually copy) the double word at 0h to 32h
<9> since it's in the same segment
<9> but it lets you move data from segment to segment
<8> i see, so that is why for copying buffer you have to 'mov' you're buffer to esi
<8> then sub esp, ecx ; ecx = size of buffer
<9> yeah
<8> then mov esp, esp
<9> err
<8> sorrym mov edi,esp
<9> then movds
<8> should i make ecx dwords first?
<9> nah
<9> you have do make ecx containt the size, in bytes, of a double word
<8> shr ecx, 2 ?
<9> right divides...
<8> now, how can i call a function with 4-bytes size parameters the fasters way?
<8> i'm using Intel P4 prescott
<9> you're p***ing it on the stack ?
<9> like in the C calling convention... ?
<8> i guess but i'm not sure...
<9> well, if you're calling a C function, the params are on the stack
<8> Win32 dll
<9> let's check it out
<9> what function ?
<8> my custom dll, exported via ordinal
<9> okay, cdecl
<9> p*** them on the stack, the order they're pushed is right to left, and clean up the stack afterwards
<9> so you add esp, 4*number_of_args
<9> after the call
<9> there's technically no faster way
<9> depends on what you're p***ing...
<8> i'm p***ing whole dword data...
<8> i mean thats the only param
<9> yeah, 4 bytes...
<9> push 0DEADBEEFh
<9> call _dll_myfunc
<9> add esp, 4
<8> so its just 3 instructions?, seem better for my app
<9> well, yeah. only one arg
<8> if i need to get the return value i'll just pick in to eax
<9> yup
<9> if it's 32bits..
<3> !seen xark
<3> markos_64 no i mean what does Terminal Insanite Stages mean?
<8> thx pireau
<9> np
<10> AMD to purchase ATI ?
<9> wtf
<9> (chaos: ~/prog/asm/c_call)$ ld -o call_c call_c.o testing.o
<9> testing.o(.text+0x12): In function `testing':
<9> : undefined reference to `printf'
<9> (chaos: ~/prog/asm/c_call)$
<9> (chaos: ~/prog/asm/c_call)$ ld -o call_c call_c.o testing.o -lc
<9> (chaos: ~/prog/asm/c_call)$ ./call_c
<9> bash: ./call_c: No such file or directory
<9> (chaos: ~/prog/asm/c_call)$
<9> !
<9> when i link to libc the file is there but execve doesn't see it ?
<10> usr/lib/crt0.o
<9> what about it
<10> ld -dll-verbose -m486 /usr/lib/crt0.o -L/usr/lib/gcc-lib/i486-linux/2.x.x call_c.o testing.o -lgcc -lc
Return to
#asm
or
Go to some related
logs:
Can't locate RawIP.pm windows
pci-33w manual
jinzora error iconv Support Not Found!
#debian
#linux
#lgp
gentoo moonpatrol
why does ps2 mouse lag on gentoo
regular expretions php
#linux
|
|