Anon3 (anon3.delete@this.anon3.invalid) on August 17, 2020 5:34 pm wrote:

>
> The processor cannot know where the instruction boundaries are without decoding from
> the target of a branch and some x86 instructions are one byte thus it's a fundamental
> property of the ISA (without CFI) that you must be able to jump to any address.
>



On x86 not only it is permitted to jump into the middle of an instruction, but actually this was used very frequently in most commercial MS-DOS programs, for 2 reasons, some times it was used for code obfuscation, to desynchronize the less clever disassemblers, some times it was used because this reduced the code size, e.g. by using a 1-byte load instruction instead of a 2-byte jump (on 8088) or 3-byte jump (on 8080). The load was only 1 byte long because it loaded an immediate value which was then ignored, so this was equivalent to skipping over the next 2 bytes of code.

The first time I have seen this technique used in many Microsoft programs for CP/M on Intel 8080, e.g. in their BASIC interpreter. Later this style of programming was inherited by many of their MS-DOS programs.