Page 1 of 1
feed callfunctionfast with dynamic params
Posted: Tue Mar 11, 2008 11:04 pm
by Lukaso
hi,
i work on some rpc procedures for pb to make a generalized protocol for some of my projects.
the problem that currently occured is that i have no idea how to feed callprocedure fast dynamicly with vars.
I look for something like this (pseudocode):
Code: Select all
*funcaddr = @bla()
appendParamString(*funcaddr, "string")
appendParamLong(*funcaddr, 234234)
appendParamByte(*funcaddr, 123)
callProcedure(*funcaddr)
I dont have alot of knowleadge about asm, but maybe it is possible to push the vars dynamicly on a stack.
maybe someone can help me
thanks, lukas
Posted: Tue Mar 11, 2008 11:41 pm
by Deeem2031
Something like this?
Code: Select all
Procedure bla(s.s,l.l,b.b,f.f,q.q,d.d)
Debug s
Debug l
Debug b
Debug f.f
Debug Hex(q)
Debug d
ProcedureReturn 2031
EndProcedure
Procedure callProcedure(adr,ll.l())
Protected tmp
!MOV Edx, dword[p.v_adr]
!MOV Eax, dword[p.v_adr+4]
!MOV Eax, dword[Eax]
!MOV Eax, dword[Eax+4]
!asm_callProcedure_loop:
!OR Eax, Eax
!JZ asm_callProcedure_loop_end
!PUSH dword[Eax+8]
!MOV Eax, dword[Eax+4]
!JMP asm_callProcedure_loop
!asm_callProcedure_loop_end:
!CALL Edx
ProcedureReturn
EndProcedure
Procedure callProcedure_AddString(ll.l(),*s)
AddElement(ll())
ll() = *s
EndProcedure
Procedure callProcedure_AddLong(ll.l(),l.l)
AddElement(ll())
ll() = l
EndProcedure
Procedure callProcedure_AddFloat(ll.l(),f.f)
AddElement(ll())
ll() = PeekL(@f)
EndProcedure
Procedure callProcedure_AddQuad(ll.l(),q.q)
AddElement(ll())
ll() = PeekL(@q)
AddElement(ll())
ll() = PeekL(@q+4)
EndProcedure
Procedure callProcedure_AddDouble(ll.l(),d.d)
AddElement(ll())
ll() = PeekL(@d)
AddElement(ll())
ll() = PeekL(@d+4)
EndProcedure
*funcaddr = @bla()
NewList ll()
callProcedure_AddString(ll(),@"string")
callProcedure_AddLong(ll(),23456)
callProcedure_AddLong(ll(),123)
callProcedure_AddFloat(ll(),123.456)
callProcedure_AddQuad(ll(),$12345678ABCDEF09)
callProcedure_AddDouble(ll(),123.45678)
Debug callProcedure(*funcaddr, ll())
Posted: Thu Mar 13, 2008 9:42 pm
by Lukaso
Wee, my hero
Thank you
Some bonus questions (not, assantual for me): what about quad/double/float ... and will it work under linux too?