linux-kernel - Re: [RFC 3/8] mm: Avoid using set_page_count() in set_page

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CA+CK2bD6x01PevPqshzYqkO3aokjP2jBbt_4e5H5U3DVEdcJ5Q@mail.gmail.com>
Date:   Tue, 26 Oct 2021 14:21:53 -0400
From:   Pasha Tatashin <pasha.tatashin@...een.com>
To:     John Hubbard <jhubbard@...dia.com>
Cc:     LKML <linux-kernel@...r.kernel.org>, linux-mm <linux-mm@...ck.org>,
        linux-m68k@...ts.linux-m68k.org,
        Anshuman Khandual <anshuman.khandual@....com>,
        Matthew Wilcox <willy@...radead.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        william.kucharski@...cle.com,
        Mike Kravetz <mike.kravetz@...cle.com>,
        Vlastimil Babka <vbabka@...e.cz>,
        Geert Uytterhoeven <geert@...ux-m68k.org>,
        schmitzmic@...il.com, Steven Rostedt <rostedt@...dmis.org>,
        Ingo Molnar <mingo@...hat.com>,
        Johannes Weiner <hannes@...xchg.org>,
        Roman Gushchin <guro@...com>, songmuchun@...edance.com,
        weixugc@...gle.com, Greg Thelen <gthelen@...gle.com>
Subject: Re: [RFC 3/8] mm: Avoid using set_page_count() in set_page_recounted()

Hi John,

Thank you for looking at this series.

> >   static inline void set_page_refcounted(struct page *page)
> >   {
> > +     int refcnt;
> > +
> >       VM_BUG_ON_PAGE(PageTail(page), page);
> >       VM_BUG_ON_PAGE(page_ref_count(page), page);
> > -     set_page_count(page, 1);
> > +     refcnt = page_ref_inc_return(page);
> > +     VM_BUG_ON_PAGE(refcnt != 1, page);


> I am acutely uncomfortable with this change, because it changes the
> meaning and behavior of the function to something completely different,
> while leaving the function name unchanged. Furthermore, in relies upon
> debug assertions, rather than a return value (for example) to verify
> that all is well.


It must return the same thing, if it does not we have a bug in our
kernel which may lead to memory corruptions and security holes.

So today we have this:
   VM_BUG_ON_PAGE(page_ref_count(page), page); -> check ref_count is 0
   < What if something modified here? Hmm..>
   set_page_count(page, 1); -> Yet we reset it to 1.

With my proposed change:
   VM_BUG_ON_PAGE(page_ref_count(page), page); -> check ref_count is 0
   refcnt = page_ref_inc_return(page);  -> ref_count better be 1.
   VM_BUG_ON_PAGE(refcnt != 1, page); -> Verify that it is 1.

>
> I understand where this patchset is going, but this intermediate step is
> not a good move.
>
> Also, for the overall series, if you want to change from
> "set_page_count()" to "inc_and_verify_val_equals_one()", then the way to
> do that is *not* to depend solely on VM_BUG*() to verify. Instead,
> return something like -EBUSY if incrementing the value results in a
> surprise, and let the caller decide how to handle it.

Actually, -EBUSY would be OK if the problems were because we failed to
modify refcount for some reason, but if we modified refcount and got
an unexpected value (i.e underflow/overflow) we better report it right
away instead of waiting for memory corruption to happen.

Thanks,
Pasha