On Thu, Apr 21, 2022 at 9:50 AM Linus Torvalds
<torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
>
> where that "read(..) = 16" is the important part. It correctly figured
> out that it can only do 16 bytes (ok, 17, but we've always allowed the
> user accessor functions to block).
Bad choice of words - by "block" I meant "doing the accesses in
blocks", in this case 64-bit words.
Obviously the user accessors _also_ "block" in the sense of having to
wait for page faults and IO.
I think 'copy_{to,from}_user()' actually does go to the effort to try
to do byte-exact results, though.
In particular, see copy_user_handle_tail in arch/x86/lib/copy_user_64.S.
But I think that we long ago ended up deciding it really wasn't worth
doing it, and x86 ends up just going to unnecessary lengths for this
case.
Linus
