Received-SPF: pass (google.com: domain of linux-kernel+bounces-30560-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Precedence: bulk
MIME-Version: 1.0
References: <20240118133504.2910955-1-shy828301@gmail.com> <ZamGsyJ5JTQ9GYNL@casper.infradead.org>
In-Reply-To: <ZamGsyJ5JTQ9GYNL@casper.infradead.org>
From: Yang Shi <shy828301@gmail.com>
Date: Thu, 18 Jan 2024 12:39:55 -0800
Message-ID: <CAHbLzkotGn9HifF9GQB5W5wMPZpxQviCFTStHcQCD5D37v8UVw@mail.gmail.com>
Subject: Re: [PATCH] mm: huge_memory: don't force huge page alignment on 32 bit
To: Matthew Wilcox <willy@infradead.org>
Cc: jirislaby@kernel.org, surenb@google.com, riel@surriel.com, cl@linux.com, 
	akpm@linux-foundation.org, yang@os.amperecomputing.com, linux-mm@kvack.org, 
	linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Thu, Jan 18, 2024 at 12:14=E2=80=AFPM Matthew Wilcox <willy@infradead.or=
g> wrote:
>
> On Thu, Jan 18, 2024 at 05:35:04AM -0800, Yang Shi wrote:
> > It doesn't make too much sense to force huge page alignment on 32 bit
> > system due to the constrained virtual address space.
> >
> > [1] https://lore.kernel.org/linux-mm/CAHbLzkqa1SCBA10yjWTtA2mKCsoK5+M1B=
thSDL8ROvUq2XxZMw@mail.gmail.com/T/#mf211643a0427f8d6495b5b53f8132f453d60ab=
95
> > [2] https://lore.kernel.org/linux-mm/CAHbLzkqa1SCBA10yjWTtA2mKCsoK5+M1B=
thSDL8ROvUq2XxZMw@mail.gmail.com/T/#me93dff2ccbd9902c3e395e1c022fb454e48ecb=
1d
>
> I feel sure there are shorter URLs for those messages ...

Oh, yeah, I just found.

[1] https://lore.kernel.org/linux-mm/d0a136a0-4a31-46bc-adf4-2db109a61672@k=
ernel.org/
[2] https://lore.kernel.org/linux-mm/CAJuCfpHXLdQy1a2B6xN2d7quTYwg2OoZseYPZ=
TRpU0eHHKD-sQ@mail.gmail.com/

>
> > @@ -811,6 +812,14 @@ static unsigned long __thp_get_unmapped_area(struc=
t file *filp,
> >       loff_t off_align =3D round_up(off, size);
> >       unsigned long len_pad, ret;
> >
> > +     /*
> > +      * It doesn't make too much sense to froce huge page alignment on
> > +      * 32 bit system or compat userspace due to the contrained virtua=
l
> > +      * address space and address entropy.
> > +      */
>
> I honestly wouldn't even comment this.  But if you must,
>
>         /* Using THP alignment is not as important as address randomisati=
on */

It is not only about address randomization. Removing the comment is fine to=
o.

>
> > +     if (IS_ENABLED(CONFIG_32BIT) || in_compat_syscall())
> > +             return 0;
> > +
> >       if (off_end <=3D off_align || (off_end - off_align) < size)
> >               return 0;
> >
> > --
> > 2.41.0
> >