Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <CAHk-=wh6MxaCA4pXpt1F5Bn2__6MxCq0Dr-rES4i=MOL9ibjpg@mail.gmail.com>
 <20221108194139.57604-1-torvalds@linux-foundation.org> <15A1A033-73DC-4ACA-87B0-3535E9236F7F@gmail.com>
In-Reply-To: <15A1A033-73DC-4ACA-87B0-3535E9236F7F@gmail.com>
From:   Linus Torvalds <torvalds@linux-foundation.org>
Date:   Tue, 8 Nov 2022 12:46:44 -0800
Message-ID: <CAHk-=wj10fOhx9-dkmRrttgXmhVihhwfTT=bdhqVQwC=sPUE2A@mail.gmail.com>
Subject: Re: [PATCH 1/4] mm: introduce 'encoded' page pointers with embedded
 extra bits
To:     Nadav Amit <nadav.amit@gmail.com>
Cc:     Hugh Dickins <hughd@google.com>,
        Johannes Weiner <hannes@cmpxchg.org>,
        Andrew Morton <akpm@linux-foundation.org>,
        kernel list <linux-kernel@vger.kernel.org>, linux-mm@kvack.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Tue, Nov 8, 2022 at 12:37 PM Nadav Amit <nadav.amit@gmail.com> wrote:
>
> > +static inline bool encoded_page_flags(struct encoded_page *page)
> > +{
> > +     return ENCODE_PAGE_BITS & (unsigned long)page;
> > +}
>
> I think this one wants to be some unsigned, as otherwise why have
> ENCODE_PAGE_BITS as 3ul ?

Right you are. That came from my old old version where this was just
"bool dirty".

Will fix.

Doesn't matter for the TLB flushing case, but I really did hope that
we could use this for mlock too, and that case needs both bits.

I did look at converting mlock (and it's why I wanted to make
release_pages() take that whole encoded thing in general, rather than
make some special case for it), but the mlock code uses that "struct
pagevec" abstraction that seems entirely pointless ("pvec->nr" becomes
"pagevec_count(pvec)", which really doesn't seem to be any clearer at
alll), but whatever.

               Linus