AMD64\cl.exe Long Long uses 32 bit Register.

[Thema203]

Hi;


I am working on a linker program for Linux64. The development environment however

is running Windows 7 Professional 64 Bit. In other words, I am using a Windows Operating

System to develop the linker. Everything is being done in Windows, including testing and

when it's finished, I will do an overall test in Linux64. After getting errors with type-casting

I decided to create a function called 'ConvertStringToBinary' which takes a string of binary

numbers and converts it into a number. However, I am having some problems with that

function, where it changes 'tmpn4' to 0 after exceeding the maximum 32-bit number. I

guess that it is still using 32-bit registers when it multiplies 'tmpn4' by 16. The code is

below;


TRESULT ConvertStringToBinary(_IN char *instr, _IN int inn, _INOUT long long *outn){


//Variable initializations.

char *tmpstr = 0;

long long tmpn2 = 0;

long long tmpn3 = 0;

long long tmpn4 = 0;

long long tmpn5 = 0;

int tmpn = 0;

int ret = 0;

int i = 0;


//Argument checks.

if(instr == NULL){

printf("Error util0000050 - NULL pointer received in ConvertStringToBinary.\n");

return TPTRISNULL;

}

if(outn == NULL){

printf("Error util0000050 - NULL pointer received in ConvertStringToBinary.\n");

return TPTRISNULL;

}

if(inn == 0){

printf("Error util0000052 - Zero length received in ConvertStringToBinary.\n");

return TIDXISZERO;

}


//Initializations.

tmpstr = instr;

tmpn = inn;

tmpn4 = 16;



//Main logic.



//Convert the string to binary.

for(i = 0; i < tmpn; i++){

tmpn2 = (tmpstr & 0x0F);

if(i != 0){

tmpn2 = (tmpn2 * tmpn4);

tmpn4 = (tmpn4 * 16);

}

tmpn3 = (tmpstr & 0xF0);

tmpn3 = (tmpn3 >> 4);

tmpn3 = (tmpn3 * tmpn4);

tmpn4 = (tmpn4 * 16);

tmpn5 = (tmpn5 + tmpn2 + tmpn3);

}



//Returns

*outn = tmpn5;

return TSUCCESS;


}


Can anyone solve this. If they do then I will very happy.


Thema

Continue reading...