Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20211026173822.502506-1-pasha.tatashin@soleen.com>
 <20211026173822.502506-2-pasha.tatashin@soleen.com> <YXhbq/6OIpIAr7Tx@casper.infradead.org>
In-Reply-To: <YXhbq/6OIpIAr7Tx@casper.infradead.org>
From:   Pasha Tatashin <pasha.tatashin@soleen.com>
Date:   Tue, 26 Oct 2021 17:34:18 -0400
Message-ID: <CA+CK2bCZO5J-NRqavWFKXt+yB1u17dFA9VrW48HLBepiQLJtcA@mail.gmail.com>
Subject: Re: [RFC 1/8] mm: add overflow and underflow checks for page->_refcount
To:     Matthew Wilcox <willy@infradead.org>
Cc:     LKML <linux-kernel@vger.kernel.org>, linux-mm <linux-mm@kvack.org>,
        linux-m68k@lists.linux-m68k.org,
        Anshuman Khandual <anshuman.khandual@arm.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        william.kucharski@oracle.com,
        Mike Kravetz <mike.kravetz@oracle.com>,
        Vlastimil Babka <vbabka@suse.cz>,
        Geert Uytterhoeven <geert@linux-m68k.org>,
        schmitzmic@gmail.com, Steven Rostedt <rostedt@goodmis.org>,
        Ingo Molnar <mingo@redhat.com>,
        Johannes Weiner <hannes@cmpxchg.org>,
        Roman Gushchin <guro@fb.com>, songmuchun@bytedance.com,
        weixugc@google.com, Greg Thelen <gthelen@google.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Tue, Oct 26, 2021 at 3:50 PM Matthew Wilcox <willy@infradead.org> wrote:
>
> On Tue, Oct 26, 2021 at 05:38:15PM +0000, Pasha Tatashin wrote:
> >  static inline void page_ref_add(struct page *page, int nr)
> >  {
> > -     atomic_add(nr, &page->_refcount);
> > +     int ret;
> > +
> > +     VM_BUG_ON(nr <= 0);
> > +     ret = atomic_add_return(nr, &page->_refcount);
> > +     VM_BUG_ON_PAGE(ret <= 0, page);
>
> This isn't right.  _refcount is allowed to overflow into the negatives.
> See page_ref_zero_or_close_to_overflow() and the conversations that led
> to it being added.

#define page_ref_zero_or_close_to_overflow(page) \
1204   ((unsigned int) page_ref_count(page) + 127u <= 127u)


Uh, right, I saw the macro but did not realize there was an (unsigned int) cast.

OK, I think we can move this macro inside:
include/linux/page_ref.h

modify it to something like this:
#define page_ref_zero_or_close_to_overflow(page) \
   ((unsigned int) page_ref_count(page) + v + 127u <= v + 127u)

The sub/dec can also be fixed to ensure that we do not underflow but
still working with the fact that we use all 32bits of _refcount.

Pasha