Re: Simplify load_unaligned_zeropad() (was Re: [GIT PULL] Ceph updates for 5.20-rc1)

[Linus Torvalds]

On Sun, Aug 14, 2022 at 3:51 PM Kirill A. Shutemov <kirill@xxxxxxxxxxxxx> wrote:
>
> Do we gain enough benefit from the microoptimization to justify existing
> load_unaligned_zeropad()? The helper has rather confusing side-effects.

It's a *big* deal in pathname handling, yes. Doing things a byte at a
time is very noticeably slower.

If TDX has problems with it, then TDX needs to be fixed. And it's
simple enough - just make sure you have a guard page between any
kernel RAM mapping and whatever odd crazy page.

There is nothing subtle about this at all. You probably would want
that guard page *anyway*.

That "uaccepted memory" type needs to be just clearly fixed.

                Linus