Using EAX in 16-Bit ASM

[Metallic-Red]

I've made a 16-Bit DOS style OS but in one section of code, I needed to store my data into a 32-Bit value, so I put it in EAX! Will this conflict or anything? The code works but i'm not sure what to expect later on...

[distantvoices]

I don't think so. You just need to take care to tell the processor that 32 bit values are to be expected/handled. Maybe you gonna need some prefixes (don't know them by heart. The Faq might contain some hints about that)

*shrugs* Others mileage might be different.

[fraserjgordon]

If using the AT&T syntax, specify the command with the "l" suffix, e.g. movl $1, %eax. The assembler will insert an "opsize" command before the opcode (0x67, i think) and this will tell the machine to use the 32-bit register (as long as it is not a 286 or older!)

The "l" suffix isn't strictly needed, but it helps and can make it easier to spot certain errors (e.g. movl $1, %ax).

One last thing - remember that if you use the eax register, you will trash any value stored in ax.

[AR]

The prefixes, in Intel/NASM Syntax, are o32, a32, o16 and a16. The "o" prefixes are for the operation, "a" is address size.
http://nasm.sourceforge.net/doc/html/nasmdoc9.html#section-9.3

NASM will automatically prefix the instructions for you though, dec eax and dec ax have the same opcode so NASM will prefix it to achieve the desired effect

[aladdin]

since the processor knows all its regs (in 16bit and 32bit modes) using EAX will not make any conflit, the only problem you can get, can happens with old CPUs

[bubach]

why not use a 32-bit variable?

[Metallic-Red]

Yeah I know that I can easily use a variable, but I was just asking about the use of EAX, etc.