add nopi command it patchs full instructions

[therealdreg]

Description/Motivation/Screenshots

nopi command patchs full instructions, nop command usually breaks disasm

Also, the current way using nop command + calculating bytes by hand/commands when you want patch N instructions is a pain in the ass hehe

This command should be more useful & user friendly than nop command

Against which architecture was this tested ?

"Tested" indicates that the PR works and the unit test (see docs/testing.md) run passes without issue.

• x86-32
• x86-64
• ARM
• AARCH64
• MIPS
• POWERPC
• SPARC
• RISC-V

Checklist

• My PR was done against the dev branch, not main.
• My code follows the code style of this project.
• My change includes a change to the documentation, if required.
• If my change adds new code, adequate tests have been added.
• I have read and agree to the CONTRIBUTING document.

[hugsy]

I'm not a fan of having two commands that do the same thing, especially when the code is this similar. Why not simply add an option (like --nb-insn) to the nop command to specify the number of instructions instead? That'd seem cleaner and users will still have the option to create an alias to it.

[hugsy]

nop command usually breaks disasm

Thanks for pointing that out. There must have a been a mix-up somewhere, because I thought nop had an instruction granulariy, and use patch for a byte/word/dword/etc. granularity. That's probably what should change, which is another argument for making your nopi the default behavior of nop (i.e. instruction granularity) - and just leave a toggle to switch to a byte-granularity. To explain by example:

• nop $address 10 will nop out 10 instructions, cleanly
• whereas nop --bytes $address 10 will forcefully insert 10 nops instructions, regardless of what's there.

Can you update your PR to reflect that?

[therealdreg]

So...

nop -> patch $pc 1-full-instruction
nop --nb -> patch $pc with 1-nop

nop --ni 10 -> patch $pc 10-full-instruction
nop --nb 10 -> patch $pc with 10-nop

nop addr -> patch addr 1-full-instruction
nop addr --nb -> patch addr with 1-nop

nop addr --ni 10 -> patch addr 10-full-instructions
nop addr --nb 10 -> patch addr with 10-nop

do you like this way?

@hugsy I updated this comment, now it makes more sense

[therealdreg]

IMO, I dont like the "switch idea", have a nop & nopi commands its more fast and easy to use+remember :D (and also easy to mantain, the code+tests will be more simple and easy to follow). So, what do you want? tell me and I will do it :D

[hugsy]

My point is, there should not be two commands: I thought the current nop function we have was behaving your nopi. That's the behavior I wanted. Because if you only want to patch (and maybe break the disassembly) the command patch is here exactly for that.

So yes, let's have nop behave like your nopi i.e. with an instruction granularity and we simply keep a flag to toggle to a byte granularity instead (non-default behavior).

The syntax can be something like:

nop [0xAddress|Symbol] [NumNops] 

For instance (to pile on your examples):

nop -> patch $pc 1-full-instruction
nop --as-bytes -> patch $pc with 1-nop, breaking (maybe) disasm

nop $addr -> patch $addr with one NOP, preserving size, not breaking disasm
nop --as-bytes $addr-> patch $addr with one NOP, not preserving size, breaking disasm

nop 0xaddr 10 -> insert 10 nops, preserve size, not breaking disasm
nop --as-bytes $addr 10-> patch $addr with 10 NOP, not preserving size, breaking (maybe) disasm

WDYT ?

[therealdreg]

Yeah, gotcha

[therealdreg]

I think this should be the right thing: nop granularity + instruction granularity (for others uses ... use patch byte cmd)

[therealdreg]

done! @hugsy what do you think now?