OSDev.org

This line doesn't work on nasm. It give two errors:

"warning: byte value exceeds bounds"

"'obj' output driver does not support one-byte relocations"

try mov dx, 0x1f6

"try mov dx, 0x1f6"

already did. No change.

This is the reason, but as i do not understand c maybe some one can convert.

case 020: case 021: case 022:
if (ins->oprs[c-020].offset < -256 || ins->oprs[c-020].offset > 255)
    errfunc (ERR_WARNING, "byte value exceeds bounds");
bytes[0] = ins->oprs[c-020].offset;
out (offset, segment, bytes, OUT_RAWDATA+1, NO_SEG, NO_SEG);
offset += 1;
break;

ASHLEY4.

Well, somehow the compiler mixex up the "dx" with the byte.
Does "mov dx, 1f0h" give an error, or is it only with this magic number?
What about "mov ax,1f6"?
Do you have any code settings(compiler hints) before this instruction?(default operand size-you know)
Anton.

no compiler hints, and 1f*h gets an error. Any ideas?

Did you try "mov dx, 01f6h" or "mov dx, 0x01f6"

The line works perfectly fine under NASM 0.98.38.

There are a number of Intel instructions that take byte sized offsets, such as a short Jcc (jump on condition code).  This code complains if you're trying to assemble one of those instructions but your offset is too big.

Note: it would never be called while assembling "mov dx,1f6h".  Since this assembles perfectly fine under NASM, I assume the poster's actual problem is with some other line of code.  Alas, not enough context was posted to determine where exactly the problem is...

Could you post the code around it(before and after)? It's hard to judge just by one instruction.
Anton