Is there any faster way (on Pentium and above, in protected mode) to get EIP than the standard:



I can't think of one, but maybe some guru out there has a better way. Thanks!

Cute; but this is a piece of 'template' code which will be moved into place; it needs to be position independent

I don't think there is any other way, at least no shorter way. This is the easiest and fastest way. If you CALL FAR remember to clean up the stack.
This is the shortest amount of code; I don't know if there is any faster way. I don't think there is any way to do it without CALL or unless you use f0dder's code.

Yes, it works fine -- but I am manually constructing the code in the space it needs to go, now. Much, much faster than what was happening before.

I know, I was just doing a bit of fun-poking

There are other ways, but this is the shortest and fastest (at least what I know of) - other methods are usually employed by evil people who want to circumvent heuristic antivirus scanners.

If you have speed problems, you need to look for them elsewhere. Construction + executing code has quite some penalty in itself, especially if you need to call the code short after you construct it. On modern processors, you only want to do code construction or self-modifying code if the modifications are few and the executions are many.

Hi, f0dder -

This is my Forth system, Reva, which interprets code into a compiled form for execution. So indeed, the modifications are few compared to the execution of the code.

Do you have to make PIC code anyway? Wouldn't it be possible to apply relocations?

Actually I am applying relocations, because otherwise I wouldn't be able to use the label method here. If I could have PIC code, I wouldn't have to do any tricks.