Re: [PATCH v4 08/10] mmap locking API: add MMAP_LOCK_INITIALIZER

[Michel Lespinasse]

On Mon, Apr 20, 2020 at 12:28 PM Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote:
>
> On Tue, Apr 14, 2020 at 05:43:51PM -0700, Michel Lespinasse wrote:
> > @@ -90,7 +90,7 @@ static struct mm_struct tboot_mm = {
>                                             ^^^^^^^^
> >       .pgd            = swapper_pg_dir,
> >       .mm_users       = ATOMIC_INIT(2),
> >       .mm_count       = ATOMIC_INIT(1),
> > -     .mmap_sem       = __RWSEM_INITIALIZER(init_mm.mmap_sem),
> > +     .mmap_sem       = MMAP_LOCK_INITIALIZER(init_mm.mmap_sem),
>                                                 ^^^^^^^
>
> Shome mishtake, shirley?
>
> I don't see that this particular patch buys us much.  The name 'mmap_sem'
> is still used, and I appreciate we abstract away the type of the lock,
> but wouldn't this be better?
>
> -       .mmap_sem       = __RWSEM_INITIALIZER(init_mm.mmap_sem),
> +       MMAP_LOCK_INITIALIZER(tboot_mm),

Hmmm, that's significantly different from other initializers we have,
which may be a downside ? But other than that, it does seem completely
workable to me. Do you have a strong preference ?

-- 
Michel "Walken" Lespinasse
A program is never fully debugged until the last user dies.