linux-kernel - Re: [PATCH] mm/rmap: convert anon_vma.refcount from atomic_t to refcount

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] [day] [month] [year] [list]

Date:   Thu, 11 Mar 2021 13:57:31 -0800
From:   Andrew Morton <akpm@...ux-foundation.org>
To:     Yejune Deng <yejune.deng@...il.com>
Cc:     walken@...gle.com, hannes@...xchg.org, louhongxiang@...wei.com,
        linmiaohe@...wei.com, shakeelb@...gle.com,
        linux-kernel@...r.kernel.org, linux-mm@...ck.org
Subject: Re: [PATCH] mm/rmap: convert anon_vma.refcount from atomic_t to
 refcount_t

On Thu, 11 Mar 2021 13:56:15 +0800 Yejune Deng <yejune.deng@...il.com> wrote:

> refcount_t type should be used instead of atomic_t when the variable
> is used as a reference counter. This is because the implementation of
> refcount_t can prevent overflows and detect possible use-after-free.

The use of refcount_t comes at a cost:

q:/usr/src/25> size mm/rmap.o
   text    data     bss     dec     hex filename
  31114    2147      32   33293    820d mm/rmap.o
  31558    2147      32   33737    83c9 mm/rmap.o

That's a bunch more instructions to execute on some fairly hot
code paths.

I guess the debugging/checking features are nice, but this is pretty
mature code.  I'm quite unsure that this tradeoff is a favorable one.