[GIT PULL] keys: Fix key->sem vs mmap_sem issue when reading key

[Linus Torvalds]

On Sun, Apr 5, 2020 at 2:04 AM David Howells <dhowells at redhat.com> wrote:
>
> Should this be moved into core code, rather than being squirrelled away in
> security/keys/?

Yes. I do think that that __kvzfree() function makes sense in general
(the same way that kzfree does).

I just happen to despise the name, and think that the implementation
isn't great.

It also probably makes no sense to make it an inline function. It's
not like that function is done for performance reasons, and it might
only get worse if we then end up making it cause barriers or something
for CPU data leakage issues or whatever.

           Linus