Re: [RFC PATCH] text_poke/ftrace/x86: Allow text_poke() to be called in early boot

[Peter Zijlstra]

On Mon, Oct 24, 2022 at 05:11:13PM -0700, Linus Torvalds wrote:

> All of this comes from "poking_init()" being a steaming pile of bovine
> excrement, doing random odd things, and having that special
> "copy_init_mm()" helper that just makes things even worse. Nothing
> else uses that, and it shouldn't have called "dup_mm()" in the first
> place.

Agreed; dup_mm() makes no sense and it is easily removed, see my earlier
patch. Perhaps it can be simplified further to:

	__poking_mm = init_mm

omitting the mm_init() I retained, but I need to stare harder at all
that.

> I'm not even sure why "poking_mm" exists at all, and why it has
> created a whole new copy of "init_mm", and why this code isn't just
> using '&init_mm' like everything else that wants to just walk the
> kernel page tables.

Because it instantiates user-space page-tables in it, you really don't
want those in init_mm.

The whole (and sole) purpose of poking_mm is to contain the writable
aliases. Only the CPU that has the poking_mm active has access to them.