linux-kernel - Re: [PATCH 2/2] module: fix building stats for 32-bit targets

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: <cd9d08ea-ffa5-4932-b296-7a58203ef701@app.fastmail.com>
Date:   Tue, 18 Apr 2023 00:49:29 +0200
From:   "Arnd Bergmann" <arnd@...db.de>
To:     "Luis Chamberlain" <mcgrof@...nel.org>,
        "Arnd Bergmann" <arnd@...nel.org>,
        "Linus Torvalds" <torvalds@...ux-foundation.org>
Cc:     "Aaron Tomlin" <atomlin@...hat.com>,
        "Christophe Leroy" <christophe.leroy@...roup.eu>,
        "Viktor Malik" <vmalik@...hat.com>,
        "Jason Baron" <jbaron@...mai.com>, "Song Liu" <song@...nel.org>,
        "Jim Cromie" <jim.cromie@...il.com>,
        "Kees Cook" <keescook@...omium.org>,
        "Masahiro Yamada" <masahiroy@...nel.org>,
        "Sami Tolvanen" <samitolvanen@...gle.com>,
        linux-modules@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/2] module: fix building stats for 32-bit targets

On Tue, Apr 18, 2023, at 00:15, Luis Chamberlain wrote:
> On Tue, Apr 18, 2023 at 12:02:47AM +0200, Arnd Bergmann wrote:
>> I have no idea if there is a risk of these variables actually
>> overflowing 'long' on 32-bit machines. If they provably can't, it
>> would be better to do the opposite patch.
>
> I had originally used atomic64_t and added a debugfs knob for it but
> Linus had advised against it because its not a stat we care too much
> on 32-bit and atomic64 is nasty on 32-bit [0].
>
> So I went with atomic_long and the cast becuase we're just reading.
> 
> Is there a way to fix this without doing the fully jump? If not oh well.

I've sent a v2 now that does it the other way round, which is
clearly much more efficient. Have only done minimal build testing
so far, but it passes the randconfigs that failed before.

     Arnd